On time-dependent symmetries and formal symmetries of evolution equations

合集下载

量子力学英文课件格里菲斯Charter10

量子力学英文课件格里菲斯Charter10

In molecular physics, this technique is known as the Born-Oppenheimer (玻恩-奥本海默)approximation.
In quantum mechanics, the essential content of the adiabatic approximation can be cast in the form of a theorem.
Here we assume that the spectrum is discrete and nondegenerate throughout the transition from Hi to Hf , so there is no ambiguity(歧义) about the
ordering of the states; these conditions can be relaxed, given a suitable procedure for “tracking” (跟踪)the eigenfunctions, but we’re not going to pursue that
A case in point is our discussion of the hydrogen molecule ion.
We began by assuming that the nuclei were at rest, a fixed distance R apart, and we solved for the motion of the electron.
and they are complete, so the general solution to the time-dependent Schrödinger equation

Beyond Undecidable

Beyond Undecidable

a rX iv:mat h /66713v1[mat h.GM ]28J un206BEYOND UNDECIDABLEPAOLA CATTABRIGA Abstract.The predicate complementary to the well-known G¨o del’s provabil-ity predicate is defined.From its recursiveness new consequences concerning the incompleteness argumentation are drawn and extended to new results of consistency,completeness and decidability with regard to Peano Arithmetic and the first order predicate calculus.Keywords:decision problem,provability predicate,G¨o del numbering.Introduction Of all the remarkable logical achievements of the twentieth century perhaps the most outstanding is the celebrated G¨o del incompleteness argumentation of 1931[1,2].In contrast to Hilbert’s program called for embodying classical mathematics in a formal system and proving that system consistent by finitary methods [4],G¨o del paper showed that not even the first step could be carried out fully,any formal system suitable for the arithmetic of integers was incomplete.The present article,in the most absolute respect for the extraordinary contribu-tion given by G¨o del to the logical inquiry,brings G¨o del’s achievement into question by the definition of the refutability predicate.As it is well-known self-reference plays a crucial role in G¨o del’s incompleteness argumentation and the methods of achiev-ing self-referential statements is the so-called “diagonalization”.The refutability predicate,defined by arithmetization as a number-theoretic statement,gives rise to new consequences properly regarding G¨o del’s incompleteness argumentation and the method of diagonalization.This article proposes a revision based on the logical investigation of the interactive links between provability and refutability predicates.Originally devised by G¨o del in order to arithmetize metamathematical notions,G¨o del numbering turns out to be the key of the problem in defining refutability with the same recursive status as provability.The inquiry comes up with a final solution for finitary methods and the related decision problem [3].The paper is organized as follows.Firstly,in the following of this section,we introduce diagonalization and the famous incompleteness argumentation of G¨o del.Section 1presents two new primitive recursive predicates for refutability and the enucleation of some of their consequences,which represent the first main result of this paper:G¨o del’s incompleteness argumentation is not a theorem in Peano Arithmetic.Section 2shows that any formula of Peano Arithmetic is proved if andonly if it is not refuted,and extends this result to the accomplishment of consistency and completeness for Peano Arithmetic and then to the achievement of decidability forfirst order predicate calculus.Basic Setup.We shall assume afirst order theory which adequately formalizes Peano Arithmetic(see for example the system S,with all the necessary assumptions, in[5]116-175).Let us call it PA.As is well known by means of the G¨o del numbering, each expression in PA can refer to itself.Numerals,as usual,are defined recursively,n+1is(n,and this name can be substituted back intoφ(v).This self-reference procedure is admitted by the so-called diagonalization lemma as follows.Diagonalization.For any formulaφwith only the variable v free there is a sen-tenceδsuch that⊢PAδ⇐⇒φ(n)= φ(m).We shall show thatδis the sentence we were looking for.To this purpose we notice that in PA they hold the following equivalences⊢δ⇐⇒β(m,β(v) ,β(δ )by definition.G¨o del’s Incompleteness.We present the version of the so-called G¨o del’sfirst incompleteness Theorem as it is given in([5]161-162),to which the reader can refers for the definition of the concepts which are involved.Letφ(v)be the formula∀x¬P f(x,v),hence by diagonalization lemma we attain⊢P Aδ⇐⇒∀x¬P f(x,r,r,δ ).By Biconditional Elimination,⊢P A∀x¬P f(x,r,δ ), Biconditional Elimination yields⊢P A¬∀x¬P f(x,δ ).On the other hand,since PA isω-consistent,PA is consistent.But,⊢P A¬δ.Hence,not⊢P Aδ;that is,there is no proofin PA ofδ.So Pf(n,q)is false for every natural number n and,therefore,⊢P A¬P f( δ )for every natural number n.(Remember that q.)Byω-consistency,not⊢P A∃x P f(x,1We shall not reproduce entirely this long list of definitions which is already well-known(see also[1]162-176).∃u u<x∃v v<x∃z z<x∃w w<x([x=2w∧Ax(w)]∨[Prf(u)∧Fml((u)w)∧x=u∗2v∧Gen((u)w,v)]∨[Prf(u)∧Fml((u)z)∧Fml((u)w)∧x=u∗2v∧MP((u)z,(u)w,v)]∨[Prf(u)∧x=u∗2v∧Ax(v)].Pf(x,v):x is the G¨o del number of a proof in PA of the formula with G¨o del number v:Prf(x)∧v=(x)lh(x) –1.By means of such definitions,we shall define two new predicates,Rf and Ref.Rf(x,v):x is the G¨o del number of a proof in PA of the negation of the formula with G¨o del number v:Pf(x,z)∧z=Neg(v).In other terms Rf(x,v)states x is the G¨o del number of a refutation in PA of the formula with G¨o del number v2.Rf is primitive recursive,as the relations obtained from primitive recursive rela-tions by means of propositional connectives are also primitive recursive([5]137). For its recursiveness Rf(x,v)is expressible in PA by a formula Rf(x,v).Ref(x):x is the G¨o del number of a refutation in PA:Prf(v)∧v=Neg(x)In other terms Ref(x)states x is the G¨o del number of a proof in PA of its negation.Ref(x)is primitive recursive,as the relations obtained from primitive re-cursive relations by means of propositional connectives are also primitive recursive. For its recursiveness Ref(x)is expressible in PA by a formula Ref(x).Lemma1.For any natural number n and for any formulaαnot both Rf(n, α ) and Pf(n, α ).Proof.Let us suppose to have both Rf(n, α )and Pf(n, α ).We should have thenPrf(n)∧ α =(n)lh(n) –1and Pf(n,z)∧z=Neg( α ),i.e.Prf(n)∧ α =(n)lh(n) –1and Prf(n)∧Neg( α )=(n)lh(n) –1.By the definition of Prf(x)this would mean to have∃u u<n∃v v<n∃z z<n∃w w<n([n=2w∧Ax(w)]∨[Prf(u)∧Fml((u)w)∧n=u∗2v∧Gen((u)w,v)]∨[Prf(u)∧Fml((u)z)∧Fml((u)w)∧n=u∗2v∧MP((u)z,(u)w,v)]∨[Prf(u)∧n=u∗2v∧Ax(v)]and bothα =(n)lh(n) –1and Neg( α )=(n)lh(n) –1and hence the four cases(1)[n=2 α ∧Ax( α )]and[n=2Neg( α )∧Ax(Neg( α ))](2)[Prf(u)∧Fml((u)w)∧n=u∗2 α ∧Gen((u)w, α )]and[Prf(u)∧Fml((u)w)∧n=u∗2Neg( α )∧Gen((u)w,Neg( α ))](3)[Prf(u)∧Fml((u)z)∧Fml((u)w)∧n=u∗2 α ∧MP((u)z,(u)w, α )]and[Prf(u)∧Fml((u)z)∧Fml((u)w)∧n=u∗2Neg( α )∧MP((u)z,(u)w,Neg( α ))](4)[Prf(u)∧n=u∗2 α ∧Ax( α )]and[Prf(u)∧n=u∗2Neg( α )∧Ax(Neg( α ))]which are all immediately impossible by the definitions of Ax(y),Gen(x,y)and MP(x,y,z)and thence by the definitions of the axioms of PA,the Generalization Rule and Modus Ponens,because no axiom belongs to PA together with its negation and the two inference rules preserve logical validity.We now recall the definition of characteristic function.If R is a relation of n arguments,then the characteristic function C R is defined as followsC R(x1,...,x n)= 0if R(x1,...,x n)is true,1if R(x1,...,x n)is false.Let us call the characteristic functions of Pf(x,v),Prf(x),Rf(x,v)and Ref(x) respectively C Pf,C Prf,C Rf,and C Ref.A relation R(x1,...,x n)is said to be primitive recursive(recursive)if and only if its characteristic function C R(x1,...,x n)is primitive recursive(recursive)([5] 137).As Pf(x,v),Prf(x),Rf(x,v)and Ref(x)are primitive recursive then also C Pf,C Prf,C Rf and C Ref are primitive recursive.Every recursive function is representable in PA([5]143),thence C Pf,C Prf,C Rf and C Ref,are representable in PA.We shall assume C P f,C P rf,C Rf and C Ref to represent respectively C Pf,C Prf,C Rf and C Ref in PA.Lemma2.For any formulaα,and n as the G¨o del number of a proof in PA ofα⊢P A C P f( α )=n,1Proof.One can easily see that the two conjuncts are true:as n is the G¨o del number of a proof in PA ofαC P f( α )=n,0∧C P f( α )=n,0is true.By Lemma(1)Pf(n, α )does not hold,therefore it is true that n is not the G¨o del number of a proof in PA of α. Lemma4.For any formulaα(i)not both⊢P A P f( α )⊢P A Rf( α ),(ii)for n as the G¨o del number of a refutation in PA ofα⊢P A Rf( α )⇐⇒¬P f( α ),(iii)for n as the G¨o del number of a proof in PA ofα⊢P A P f( α )⇐⇒¬Rf( α ).Proof.(i)Immediately by Lemma(1)and the definition of being expressible which holds for both Pf(x,v)and Rf(x,v)([5]130).(ii)Let us assume⊢P A Rf( α ),then Lemma(3)yields⊢P A C P f( α )=n,n,n,n,n,n,n,1.Hence by definition Rf( α )is false,consequently⊢P A¬Rf( α ).Con-versely let us assume⊢P A¬Rf( α )then Rf( α )is false and by Lemma(2) we attain⊢P A P f( α ).All preceding lemmas were carried out constructively,needlessly to assume con-sistency.We are now able to consider the consequences yielded by such lemmas to the G¨o del’s argumentation.(a′)Assume⊢P Aδ.Let r be the G¨o del number of a proof in PA ofδ.Then Pf(r,q).Hence,⊢P A P f(q),that is⊢P A P f( δ ).Hence by Lemma(4)(i)⊢P A Rf( δ )is not admitted,which means that r cannot be the G¨o delnumber of a refutation ofδ(indeed Lemma(2)yields⊢P A C Rf( δ )=r,r,r,0).(b′)Assume⊢P A¬δ.Let r be the G¨o del number of a proof in PA of¬δ.Then Rf(r,q).Hence⊢P A Rf(q)that is⊢P A Rf( δ ).Hence by Lemma(4)(i)⊢P A P f( δ )is not admitted.This means that r cannot be theG¨o del number of a proof ofδ(in fact,r is the G¨o del number of a refutation ofδ,Lemma(3)yields⊢P A C P f( δ )=r,r,r,0).We have thus shown that previous Lemmas prevent any accomplishment of(a) and(b)within G¨o del’s argumentation3.We have then established the following theorem.Theorem5.By the arithmetization of the refutability predicate G¨o del’s incom-pleteness does not hold as a theorem of PA.2.Consistency,Completeness and DecidabilityA recursive predicate defines a decidable set,by reason that its characteristic function is considered to be effectively computable([5]165,249).Let us call T P A the set of G¨o del numbers of theorems of PA and R P A the set of G¨o del numbers of refutations of PA.By the recursiveness of Pf(x,v),C Pf(x,v)=0if v∈T P A and C Pf(x,v)=1if v/∈T P A.By the recursiveness of Rf(x,v),C Rf(x,v)=0if v∈R P A and C Rf(x,v)= 1if v/∈R P A.We can than state the following theorem.Theorem6.T P A and R P A are decidable sets.r,It is furthermore well-known that if we have a computable function f(x1,...,x n) such thatf(x1,...,x n)= 0if<x1,...,x n>∈S1if<x1,...,x n>/∈S(where S is a set of natural number which turns out to be decidable just by this definition),then the function g(x1,...,x n)defined byg(x1,...,x n)= 1if f(x1,...,x n)=00if f(x1,...,x n)=1is effectively computable too.Accordingly the complement of S is decidable.One can easily see that for f(x1,...,x n)primitive recursive,g(x1,...,x n)is primitive recursive too.Consequently we haveC¬Prf(x)= 1if C Prf(x)=00if C Prf(x)=1C¬Ref(x)= 1if C Ref(x)=00if C Ref(x)=1C¬Pf(x,v)= 1if C Pf(x,v)=00if C Pf(x,v)=1C¬Rf(x,v)= 1if C Rf(x,v)=00if C Rf(x,v)=1where¬Prf,¬Pf,¬Ref and¬Rf are respectively complementary of Prf,Pf,Ref and Rf.Let us summarize,Prf,Pf,Ref and Rf are primitive recursive,then C Prf,C Pf, C Ref and C Rf are primitive recursive too.But C¬Prf(x)=1−C Prf(x),C¬Pf(x,v)= 1−C Prf(x,v),C¬Ref(x)=1−C Ref(x),and C¬Rf(x,v)=1−C Rf(x,v),thence¬Prf,¬Pf,¬Ref and¬Rf are primitive recursive too.We have then the following statements.Lemma7.For every xPrf(x)if and only if¬Ref(x).Proof.Let us assume Prf(x).C Prf(x)=0.Hence C Prf(Neg(x))=1,by the effective computability of C Prf.Prf(Neg(x))is false,then Ref(x)is false.Accordingly, C Ref(x)=1.Thus C¬Ref(x)=0and¬Ref(x).Conversely,let us assume¬Ref(x).Then C¬Ref(x)=0and C Ref(x)=1.If Ref(x)is false by its definition Prf(Neg(x))is false.Thus C Prf(Neg(x))=1and C¬Prf(Neg(x))=0.Consequently C¬Prf(x)=1,and C Prf(x)=0.Hence Prf(x).If we convent to formalize“a proof in PA ofθ”withθ1...θr⊢P Aθthen we have ⊢P A(θ1⇒(θ2⇒...(θr⇒θ)...))(Herbrand,1930).Indeed Lemma(7)could be read as follows:forθ1,...,θr,θformulas in PA Prf( (θ1⇒(θ2⇒...(θr⇒θ)...)) )if and only if¬Ref( (θ1⇒(θ2⇒...(θr⇒θ)...)) ).Furthermore,by the recursiveness of Prf(x),C Prf(x)=0if x∈T P A and C Prf(x)=1if x/∈T P A.By the recursiveness of Ref(x)C Ref(x)=0if x∈R P A and C Ref(x)=1if x/∈R P A.Lemma8.For every<x,v>Pf(x,v)if and only if¬Rf(x,v).Proof.Let us assume Pf(x,v).C Pf(x,v)=0.Hence C Pf(x,Neg(v))=1.Accord-ingly C¬Pf(x,Neg(v))=0.Thence C¬Rf(x,v)=0and¬Rf(x,v).Conversely,let us assume¬Rf(x,v).We have then¬Pf(x,Neg(v))and C¬Pf(x,Neg(v))=0.Therefore C¬Pf(x,v)=1,by the effective computability of C¬Pf.Accordingly C Pf(x,v)=0 and Pf(x,v).Indeed Lemma(8)could be read as follows:forθ1,...,θr,αformulas in PA Pf( (θ1⇒(θ2⇒...(θr⇒α)...) , α )if and only if¬Rf( (θ1⇒(θ2⇒...(θr⇒α)...) , α ).Lemma9.For m= α and n= (θ1⇒(θ2⇒...(θr⇒θ)...))(i)m∈T P A iffm/∈R P A,(ii)n∈T P A iffn/∈R P A.Proof.Immediately(i)by Lemma(8),(ii)by Lemma(7). Theorem10.PA is consistent;that is,there is no formulaαsuch that bothαand ¬αare theorems of PA.Proof.Let us assume m to be the G¨o del number of a proof of a formulaαof PA and n to be the G¨o del number of a proof of¬α.Then n,m∈T P A.But since n is the G¨o del number of a proof of¬αwe have also n∈R P A,accordingly n belongs to both T P A and R P A,which is impossible by Lemma(9). Theorem11.PA is complete;that is for any well formed formulaαof PA either ⊢P Aαor⊢P A¬α.Proof.Letαbe a well formed formula of PA,we can then yield by G¨o del numbering m= α .By Lemma(9)either m∈T P A or m∈R P A.Therefore either⊢P Aαor ⊢P A¬α.Let us call PF the fullfirst-order predicate calculus([5]172).Let T P F be then the set of G¨o del number of theorems of PF.Theorem12.T P F is decidable.Proof.By G¨o del Completeness Theorem,a formulaαof PA is provable in PA if and only ifαis logically valid,andαis provable in PF if and only ifαis logically valid.Hence⊢P Aαif and only if⊢P Fα.Accordingly,for n as the G¨o del number of a proof ofαin PA,n∈T P A iffn∈T P F.Hence,by theorem(6),T P F is decidable.Calling our attention to the diagonalization lemma we note that it holds for any formulaφwith only the variable v free.In other termsφcan be replaced by any formula with only one free variable.Let us suppose now that a sentenceδis a theorem of PA,i.e.⊢P Aδ.For n as the G¨o del number of a proof in PA ofδwe have⊢P A P f( δ ).But forφ(v)as∀x Rf(x,v)diagonalization lemma could have already yielded⊢PAδ⇐⇒∀x Rf(x,we have⊢PA∀x Rf(x,n,n,n,n,n,n,n,。

Time-dependent Density Functional Theory

Time-dependent Density Functional Theory

Time-dependent Density Functional Theory Miguel A.L.Marques and E.K.U.Gross1IntroductionTime-dependent density-functional theory(TDDFT)extends the basic ideas of ground-state density-functional theory(DFT)to the treatment of excita-tions or more general time-dependent phenomena.TDDFT can be viewed an alternative formulation of time-dependent quantum mechanics but,in contrast to the normal approach that relies on wave-functions and on the many-body Schr¨o dinger equation,its basic variable is the one-body electron density,n(r,t).The advantages are clear:The many-body wave-function,a function in a3N-dimensional space(where N is the number of electrons in the system),is a very complex mathematical object,while the density is a simple function that depends solely on3variables,x,y and z.The standard way to obtain n(r,t)is with the help of afictitious system of non-interacting electrons,the Kohn-Sham system.Thefinal equations are simple to tackle nu-merically,and are routinely solved for systems with a large number of atoms. These electrons feel an effective potential,the time-dependent Kohn-Sham potential.The exact form of this potential is unknown,and has therefore to be approximated.The scheme is perfectly general,and can be applied to essentially any time-dependent situation.Two regimes can however be observed:If the time-dependent potential is weak,it is sufficient to resort to linear-response theory to study the system.In this way it is possible to calculate,e.g.optical absorp-tion spectra.It turns out that,even with the simplest approximation to the Kohn-Sham potential,spectra calculated within this framework are in very good agreement with experimental results.On the other hand,if the time-dependent potential is strong,a full solution of the Kohn-Sham equations is required.A canonical example of this regime is the treatment of atoms or molecules in strong laserfields.In this case,TDDFT is able to describe non-linear phenomena like high-harmonic generation,or multi-photon ionization.Our purpose,while writing this chapter,is to provide a pedagogical in-troduction to TDDFT1.With that in mind,we present,in section2,a quite detailed proof of the Runge-Gross theorem[5],i.e.the time-dependent generalization of the Hohenberg-Kohn theorem[6],and the corresponding2Miguel A.L.Marques and E.K.U.GrossKohn-Sham construction[7].These constitute the mathematical foundations of TDDFT.Several approximate exchange-correlation(xc)functionals are then reviewed.In section3we are concerned with linear-response theory, and with its main ingredient,the xc kernel.The calculation of excitation en-ergies is treated in the following section.After giving a brief overlook of the competing density-functional methods to calculate excitations,we present some results obtained from the full solution of the Kohn-Sham scheme,and from linear-response theory.Section5is devoted to the problem of atoms and molecules in strong laserfields.Both high-harmonic generation and ion-ization are discussed.Finally,the last section is reserved to some concluding remarks.For simplicity,we will write all formulae for spin-saturated systems.Ob-viously,spin can be easily included in all expressions when necessary.Hartree atomic units(e=¯h=m=1)will be used throughout this chapter.2Time-dependent DFT2.1PreliminariesA system of N electrons with coordinates r∂tΨ(r,t)Ψ(r,t)|2, is interpreted as the probability offinding the electrons in positions r)+ˆW(r,t).(2) Thefirst term is the kinetic energy of the electronsˆT(r2Ni=1∇2i,(3)whileˆW accounts for the Coulomb repulsion between the electronsˆW(r2N i,j=1i=j1,t).The Hamiltonian(2)is completely general and describesTime-dependent Density Functional Theory3 a wealth of physical and chemical situations,including atoms,molecules,and solids,in arbitrary time-dependent electric or magneticfields,scattering ex-periments,etc.In most of the situations dealt with in this article we will be concerned with the interaction between a laser and matter.In that case,we can write the time-dependent potential as the sum of the nuclear potential and a laserfield,ˆV TD=ˆU en+ˆV laser.The termˆU en accounts for the Coulomb attraction between the electrons and the nuclei,ˆU en (r|r i−Rν(t)|,(5)where Zνand Rνdenote the charge and position of the nucleusν,and N n stands for the total number of nuclei in the system.Note that by allowing the Rνto depend on time we can treat situations where the nuclei move along a classical path.This may be useful when studying,e.g.,scattering experiments, chemical reactions,etc.The laserfield,ˆV laser,reads,in the length gauge,ˆVlaser(r4Miguel A.L.Marques and E.K.U.Grossthe calculation of a“simple”two electron system(the Helium atom)in a laser field takes several months in a modern computer[8](see also the work on theH+2[9]molecule and the H++3molecule[10]).The effort to solve Eq.(1)growsexponentially with the number of particles,therefore rapid developments re-garding the exact solution of the Schr¨o dinger equation are not expected.In these circumstances,the natural approach of the theorist is to trans-form and approximate the basic equations to a manageable level that still retains the qualitative and(hopefully)quantitative information about the system.Several techniques have been developed throughout the years in the quantum chemistry and physics world.One such technique is TDDFT.Its goal,like always in density-functional theories,is to replace the solution of the complicated many-body Schr¨o dinger equation by the solution of the much simpler one-body Kohn-Sham equations,thereby relieving the computational burden.Thefirst step of any DFT is the proof of a Hohenberg-Kohn type theo-rem[6].In its traditional form,this theorem demonstrates that there exists a one-to-one correspondence between the external potential and the(one-body)density.With the external potential it is always possible(in princi-ple)to solve the many-body Schr¨o dinger equation to obtain the many-body wave-function.From the wave-function we can trivially obtain the density. The second implication,i.e.that the knowledge of the density is sufficient to obtain the external potential,is much harder to prove.In their seminal paper,Hohenberg and Kohn used the variational principle to obtain a proof by reductio ad absurdum.Unfortunately,their method cannot be easily gen-eralized to arbitrary DFTs.The Hohenberg-Kohn theorem is a very strong statement:From the density,a simple property of the quantum mechani-cal system,it is possible to obtain the external potential and therefore the many-body wave-function.The wave-function,by its turn,determines every observable of the system.This implies that every observable can be written as a functional of the density.Unfortunately,it is very hard to obtain the density of an interacting sys-tem.To circumvent this problem,Kohn and Sham introduced an auxiliary system of non-interacting particles[7].The dynamics of these particles are governed by a potential chosen such that the density of the Kohn-Sham system equals the density of the interacting system.This potential is local (multiplicative)in real space,but it has a highly non-local functional de-pendence on the density.In non-mathematical terms this means that the potential at the point r can depend on the density of all other points(e.g. through gradients,or through integral operators like the Hartree potential). As we are now dealing with non-interacting particles,the Kohn-Sham equa-tions are quite simple to solve numerically.However,the complexities of the many-body system are still present in the so-called exchange-correlation(xc) functional that needs to be approximated in any application of the theory.Time-dependent Density Functional Theory5 2.2The Runge-Gross theoremIn this section,we will present a detailed proof of the Runge-Gross theorem[5], the time-dependent extension of the ordinary Hohenberg-Kohn theorem[6]. There are several“technical”differences between a time-dependent and a static quantum-mechanical problem that one should keep in mind while try-ing to prove the Runge-Gross theorem.In static quantum mechanics,the ground-state of the system can be determined through the minimization of the total energy functionalE[Φ]= Φ|ˆH|Φ .(7) In time-dependent systems,there is no variational principle on the basis of the total energy for it is not a conserved quantity.There exists,however,a quantity analogous to the energy,the quantum mechanical actionA[Φ]= t1t0dt Φ(t)|i∂2If the two potentials differ solely by a time-dependent function,they will produce wave-functions which are equal up to a purely time-dependent phase.This phase6Miguel A.L.Marques and E.K.U.Grossdensity,n(r,t),i.e.v(r,t)=v (r,t)+c(t)⇒ρ(r,t)=ρ (r,t).(9) This statement immediately implies the one-to-one correspondence between the potential and the density.In the following we will utilize primes to dis-tinguish the quantities of the systems with external potentials v and v .Due to technical reasons that will become evident during the course of the proof, we will have to restrict ourselves to external potentials that are Taylor ex-pandable with respect to the time coordinate around the initial time t0v(r,t)=∞k=0c k(r)(t−t0)k,(10)with the expansion coefficientsc k(r)=1dt kv(r,t) t=t0.(11)We furthermore define the functionu k(r)=∂k2i ∇ˆψ†(r) ˆψ(r)−ˆψ†(r) ∇ˆψ(r) .(15) We now use the quantum-mechanical equation of motion,which is valid for any operator,ˆO(t),i d∂tˆO(t)+ ˆO(t),ˆH(t) |Ψ(t) ,(16)Time-dependent Density Functional Theory7 to write the equation of motion for the current density in the primed and unprimed systemsdij (r,t)= Ψ (t)| ˆj(r),ˆH (t) |Ψ (t) .(18)dtAs we start from afixed initial many-body state,at t0the wave-functions, the densities,and the current densities have to be equal in the primed and unprimed systems|Ψ(t0) =|Ψ (t0) ≡|Ψ0 (19)n(r,t0)=n (r,t0)≡n0(r)(20)j(r,t0)=j (r,t0)≡j0(r).(21) If we now take the difference between the equations of motion(17)and(18) we obtain,when t=t0,didt k+1 j(r,t)−j (r,t) t=t0=n0(r)∇u k(r).(23) The right-hand side of Eq.(23)differs from zero,which again implies that j(r,t)=j (r,t)for t>t0.In a second step we prove that j=j implies n=n .To achieve that purpose we will make use of the continuity equation∂[n(r,t)−n (r,t)]=−∇· j(r,t)−j (r,t) .(25)∂t8Miguel A.L.Marques and E.K.U.GrossAs before,we would like an expression involving the k th time derivative of the external potential.We therefore take the(k+1)st time-derivative of the previous equation to obtain(at t=t0)∂k+2∂t k+1 j(r,t)−j (r,t) t=t0=−∇·[n0(r)∇u k(r)].(26) In the last step we made use of Eq.(23).By the hypothesis(13)we have u k(r)=const.hence it is clear that if∇·[n0(r)∇u k(r)]=0,(27) then n=n ,from which follows the Runge-Gross theorem.To show that Eq.(27)is indeed fulfilled,we will use the versatile technique of demonstra-tion by reductio ad absurdum.Let us assume that∇·[n0(r)∇u k(r)]=0with u k(r)=const.,and look at the integrald3r n0(r)[∇u k(r)]2=− d3r u k(r)∇·[n0(r)∇u k(r)](28)+ S n0(r)u k(r)∇u k(r)·d S.This equality was obtained with the help of Green’s theorem.Thefirst term on the right-hand side is zero by assumption,while the second term vanishes if the density and the function u k(r)decay in a“reasonable”manner when r→∞.This situation is always true forfinite systems.We further notice that the integrand n0(r)[∇u k(r)]2is always positive.These diverse conditions can only be satisfied if either the density n0or∇u k(r)vanish identically. Thefirst possibility is obviously ruled out,while the second contradicts our initial assumption that u k(r)is not a constant.This concludes the proof of the Runge-Gross theorem.2.3Time-dependent Kohn-Sham equationsAs mentioned in section2.1,the Runge-Gross theorem asserts that all observ-ables can be calculated with the knowledge of the one-body density.Nothing is however stated on how to calculate that valuable quantity.To circumvent the cumbersome task of solving the interacting Schr¨o dinger equation,Kohn and Sham had the idea of utilizing an auxiliary system of non-interacting (Kohn-Sham)electrons,subject to an external local potential,v KS[7].This potential is unique,by virtue of the Runge-Gross theorem applied to the non-interacting system,and is chosen such that the density of the Kohn-Sham electrons is the same as the density of the original interacting system.In the time-dependent case,these Kohn-Sham electrons obey the time-dependentTime-dependent Density Functional Theory9 Schr¨o dinger equationi ∂2+v KS(r,t) ϕi(r,t).(29)The density of the interacting system can be obtained from the time-dependentKohn-Sham orbitalsn(r,t)=occi|ϕi(r,t)|2.(30)Eq.(29),having the form of a one-particle equation,is fairly easy to solve numerically.We stress,however,that the Kohn-Sham equation is not a mean-field approximation:If we knew the exact Kohn-Sham potential,v KS,we would obtain from Eq.(29)the exact Kohn-Sham orbitals,and from these the correct density of the system.The Kohn-Sham potential is conventionally separated in the following wayv KS(r,t)=v ext(r,t)+v Hartree(r,t)+v xc(r,t).(31) Thefirst term is again the external potential.The Hartree potential accounts for the classical electrostatic interaction between the electronsv Hartree(r,t)= d3r n(r,t)δn(r,τ) n(r,t),(33) whereτstands for the Keldish pseudo-time.Inevitably,the exact expression of v xc as a functional of the density is unknown.At this point we are obliged to perform an approximation.It is im-portant to stress that this is the only fundamental approximation in TDDFT. In contrast to stationary-state DFT,where very good xc functionals exist, approximations to v xc(r,t)are still in their infancy.Thefirst and simplest of these is the adiabatic local density approximation(ALDA),reminiscent of the ubiquitous LDA.More recently,several other functionals were proposed,from which we mention the time-dependent exact-exchange(EXX)functional[13], and the attempt by Dobson,B¨u nner,and Gross[14]to construct an xc func-tional with memory.In the following section we will introduce the above mentioned functionals.10Miguel A.L.Marques and E.K.U.Gross2.4xc functionalsAdiabatic approximations There is a very simple procedure that allows the use of the plethora of existing xc functionals for ground-state DFT in the time-dependent theory.Let us assume that˜v xc[n]is an approximation to the ground-state xc density functional.We can write an adiabatic time-dependent xc potential asv adiabatic xc (r,t)=˜v xc[n](r)|n=n(t).(34)I.e.we employ the same functional form but evaluated at each time with the density n(r,t).The functional thus constructed is obviously local in time. This is,of course,a quite dramatic approximation.The functional˜v xc[n]is a ground-state property,so we expect the adiabatic approximation to work only in cases where the temporal dependence is small,i.e.,when our time-dependent system is locally close to equilibrium.Certainly this is not the case if we are studying the interaction of strong laser pulses with matter.By inserting the LDA functional in Eq.(34)we obtain the so-called adi-abatic local density approximation(ALDA)v ALDA xc (r,t)=v HEGxc(n) n=n(r,t).(35)The ALDA assumes that the xc potential at the point r,and time t is equal to the xc potential of a(static)homogeneous-electron gas(HEG)of density n(r,t).Naturally,the ALDA retains all problems already present in the LDA. Of these,we would like to emphasize the erroneous asymptotic behavior of the LDA xc potential:For neutralfinite systems,the exact xc potential decays as−1/r,whereas the LDA xc potential falls offexponentially.Note that most of the generalized-gradient approximations(GGAs),or even the newest meta-GGAs have asymptotic behaviors similar to the LDA.This problem gains particular relevance when calculating ionization yields(the ionization potential calculated with the ALDA is always too small),or in situations where the electrons are pushed to regions far away from the nuclei(e.g.,by a strong laser)and feel the incorrect tail of the potential.Despite this problem,the ALDA yields remarkably good excitation ener-gies(see sections4.2and4.3)and is probably the most used xc functional in TDDFT.Time-dependent optimized effective potential Unfortunately,when one is trying to write v xc as explicit functionals of the density,one encoun-ters some difficulties.As an alternative,the so-called orbital-dependent xc functionals were introduced several years ago.These functionals are written explicitly in terms of the Kohn-Sham orbitals,albeit remaining implicit den-sity functionals by virtue of the Runge-Gross theorem.A typical member of this family is the exact-exchange(EXX)functional.The EXX action is ob-tained by expanding A xc in powers of e2(where e is the electronic charge),Time-dependent Density Functional Theory11 and retaining the lowest order term,the exchange term.It is given by the Fock integralA EXX x=−1|r−r |.(36) From such an action functional,one seeks to determine the local Kohn-Sham potential through a series of chain rules for functional derivatives.The pro-cedure is called the optimized effective potential(OEP)or the optimized potential method(OPM)for historical reasons[15,16].The derivation of the time-dependent version of the OEP equations is very similar to the ground-state case.Due to space limitations we will not present the derivation in this article.The interested reader is advised to consult the original paper[13],one of the more recent publications[17,18],or the chapter by E.Engel contained in this volume.Thefinal form of the OEP equation that determines the EXX potential is0=occj t1−∞dt d3r [v x(r ,t )−u x j(r ,t )](37)×ϕj(r,t)ϕ∗j(r ,t )G R(r t,r t )+c.c.The kernel,G R,is defined byi G R(r t,r t )=∞k=1ϕ∗k(r,t)ϕk(r ,t )θ(t−t ),(38)and can be identified with the retarded Green’s function of the system.More-over,the expression for u x is essentially the functional derivative of the xc action in relation to the Kohn-Sham wave-functionsu x j(r,t)=1δϕj(r,t).(39)Note that the xc potential is still a local potential,albeit being obtained through the solution of an extremely non-local and non-linear integral equa-tion.In reality,the solution of Eq.(37)poses a very difficult numerical prob-lem.Fortunately,by performing an approximationfirst proposed by Krieger, Li,and Iafrate(KLI)it is possible to simplify the whole procedure,and ob-tain an semi-analytic solution of Eq.(37)[19].The KLI approximation turns out to be a very good approximation to the EXX potential.Note that both the EXX and the KLI potential have the correct−1/r asymptotic behavior for neutralfinite systems.A functional with memory There is a very common procedure for the construction of approximate xc functionals in ordinary DFT.It starts with12Miguel A.L.Marques and E.K.U.Grossthe derivation of exact properties of v xc,deemed important by physical ar-guments.Then an analytical expression for the functional is proposed,such that it satisfies those rigorous constraints.We will use this recipe to generate a time-dependent xc potential which is non-local in time,i.e.that includes the“memory”from previous times[14].A very important condition comes from Galilean invariance.Let us look at a system from the point of view of a moving reference frame whose origin is given by x(t).The density seen from this moving frame is simply the density of the reference frame,but shifted by x(t)n (r,t)=n(r−x(t),t).(40) Galilean invariance then implies[20]v xc[n ](r,t)=v xc[n](r−x(t),t).(41) It is obvious that potentials that are both local in space and in time,like the ALDA,trivially fulfill this requirement.However,when one tries to deduce an xc potential which is non-local in time,onefinds condition(41)quite difficult to satisfy.Another rigorous constraint follows from Ehrenfest’s theorem which re-lates the acceleration to the gradient of the external potentiald2dt2 d3r r n(r,t)=− d3r n(r,t)∇v ext(r).(43)In the same way we can write Ehrenfest’s theorem for the Kohn-Sham system d2Time-dependent Density Functional Theory13 i.e.the total xc force of the system is zero.This condition reflects Newton’s third law:The xc effects are only due to internal forces,the Coulomb inter-action among the electrons,and should not give rise to any net force on the system.A functional that takes into account these exact constraints can be con-structed[14].The condition(46)is simply ensured by the expression1F xc(r,t)=R(t |r,t)=j(R,t )∂tdt Πxc(n(R,t ),t−t ).(50)n(r,t)∇Finally,an expression for v xc can be obtained by direct integration of F xc (see[14]for details).2.5Numerical considerationsAs mentioned before,the solution of the time-dependent Kohn-Sham equa-tions is an initial value problem.At t=t0the system is in some initial state described by the Kohn-Sham orbitalsϕi(r,t0).In most cases the initial state will be the ground state of the system(i.e.,ϕi(r,t0)will be the solution of the ground-state Kohn-Sham equations).The main task of the computational physicist is then to propagate this initial state until somefinal time,t f.14Miguel A.L.Marques and E.K.U.GrossThe time-dependent Kohn-Sham equations can be rewritten in the inte-gral formϕi(r,t f)=ˆU(t f,t0)ϕi(r,t0),(51) where the time-evolution operator,ˆU,is defined byˆU(t ,t)=ˆT exp −i t t dτˆH KS(τ) .(52)Note thatˆH KS is explicitly time-dependent due to the Hartree and xc po-tentials.It is therefore important to retain the time-ordering propagator,ˆT, in the definition of the operatorˆU.The exponential in expression(52)is clearly too complex to be applied directly,and needs to be approximated in some suitable manner.To reduce the error in the propagation from t0 to t f,this large interval is usually split into smaller sub-intervals of length ∆t.The wave-functions are then propagated from t0→t0+∆t,then from t0+∆t→t0+2∆t and so on.The simplest approximation to(56)is a direct expansion of the exponen-tial in a power series of∆tˆU(t+∆t,t)≈kl=0 −iˆH(t+∆t/2)∆t lTime-dependent Density Functional Theory15 i)Obtain an estimate of the Kohn-Sham wave-functions at time t+∆t by propagating from time t using a“low quality”formula forˆU(t+∆t,t).The expression(53)expanded to third or forth order is well suited for this purpose. ii)With these wave-functions construct an approximation toˆH(t+∆t)and toˆU(t+∆t/2,t+∆t).iii)Apply Eq.(55).This procedure leads to a very stable propagation.The split-operator method In afirst step we neglect the time-ordering in Eq.(52),and approximate the integral in the exponent by a trapezoidal rule ˆU(t+∆t,t)≈exp −iˆH KS(t)∆t =exp −i(ˆT+ˆV KS)∆t .(56) We note that the operators exp −iˆV KS∆t and exp −iˆT∆t are diagonal respectively in real and Fourier spaces,and therefore trivial to apply in those spaces.It is possible to decompose the exponential(56)into a form involving only these two operators.The two lowest order decompositions are exp −i(ˆT+ˆV KS)∆t =exp −iˆT∆t exp −iˆV KS∆t +O(∆t2)(57) andexp −i(ˆT+ˆV KS)∆t =exp −iˆT∆t2+O(∆t3).(58) For example,to apply the splitting(58)toϕ(r,t)we start by Fourier trans-forming the wave-function to Fourier space.We then apply exp −iˆT∆t16Miguel A.L.Marques and E.K.U.GrossInstead perturbation theory may prove sufficient to determine the behavior of the system.We will focus on the linear change of the density,that allows us to calculate,e.g.,the optical absorption spectrum.Let us assume that for t<t0the time-dependent potential v TD is zero–i.e.the system is subject only to the nuclear potential,v(0)–and furthermore that the system is in its ground-state with ground-state density n(0).At t0 we turn on the perturbation,v(1),so that the total external potential now consists of v ext=v(0)+v(1).Clearly v(1)will induce a change in the density. If the perturbing potential is sufficiently well-behaved(like almost always in physics),we can expand the density in a perturbative seriesn(r,t)=n(0)(r)+n(1)(r,t)+n(2)(r,t)+ (60)where n(1)is the component of n(r,t)that depends linearly on v(1),n(2) depends quadratically,etc.As the perturbation is weak,we will only be con-cerned with the linear term,n(1).In frequency space it readsn(1)(r,ω)= d3r χ(r,r ,ω)v(1)(r ,ω).(61)The quantityχis the linear density-density response function of the sys-tem.In other branches of physics it has other names,e.g.,in the context of many-body perturbation theory it is called the reducible polarization func-tion.Unfortunately,the evaluation ofχthrough perturbation theory is a very demanding task.We can,however,make use of TDDFT to simplify this process.We recall that in the time-dependent Kohn-Sham framework,the density of the interacting system of electrons is obtained from afictitious system of non-interacting electrons.Clearly,we can also calculate the linear change of density using the Kohn-Sham systemn(1)(r,ω)= d3r χKS(r,r ,ω)v(1)KS(r ,ω).(62)Note that the response function that enters Eq.(62),χKS,is the density re-sponse function of a system of non-interacting electrons and is,consequently, much easier to calculate than the full interactingχ.In terms of the unper-turbed stationary Kohn-Sham orbitals it readsχKS(r,r ,ω)=limη→0+∞jk(f k−f j)ϕj(r)ϕ∗j(r )ϕk(r )ϕ∗k(r)Time-dependent Density Functional Theory17 v(1)KS.This latter quantity can be calculated explicitly from the definition of the Kohn-Sham potentialv(1)KS(r,t)=v(1)(r,t)+v(1)Hartree(r,t)+v(1)xc(r,t).(64) The variation of the external potential is simply v(1),while the change in the Hartree potential isv(1)(r,t)= d3r n(1)(r ,t)Hartreen(1)(r ,t ).(66)δn(r ,t )It is useful to introduce the exchange-correlation kernel,f xc,defined byδv xc(r,t)f xc(r t,r t )=n(1)(r ,ω).|x−r |+f xc(x,r ,ω)From Eq.(61)and Eq.(68)trivially follows the relationχ(r,r ,ω)=χKS(r,r ,ω)+(69) d3x d3x χ(r,x,ω) 118Miguel A.L.Marques and E.K.U.Gross3.2The xc kernelAs we have seen in the previous section,the main ingredient in linear response theory is the xc kernel.f xc,as expected,is a very complex quantity that includes–or,in other words,hides–all non-trivial many-body effects.Many approximate xc kernels have been proposed in the literature over the past years.The most ancient,and certainly the simplest is the ALDA kernelf ALDA xc (r t,r t )=δ(r−r )δ(t−t )f HEGxc(n) n=n(r,t),(70)wheref HEG xc (n)=dn(r,t)[u x k(r,t)+c.c].(72)Using the definition(67)and after some algebra we arrive at thefinal form of the PGG kernelf PGG x (r t,r t )=−δ(t−t )1|r−r || occ kϕk(r)ϕ∗k(r )|2dn2 n HEG xc(n) ≡f0(n),(74)Time-dependent Density Functional Theory19where HEGxc ,the xc energy per particle of the homogeneous electron gas,isknown exactly from Monte-Carlo calculations[23].Also the infinite frequency limit can be written as a simple expressionlim q→0f HEGxc(q,ω=∞)=−43dn2/3 +6n1dnHEG xc(n)πf HEGxc(q,ω)πf HEGxc(q,ω)−f HEGxc(q,∞)15ω3/2.(79)The real part can be obtained with the help of the Kramers-Kronig relationslim ω→∞ f HEGxc(q=0,ω)=f∞(n)+23π(1+β(n)ω2)520Miguel A.L.Marques and E.K.U.GrossFig.1.Real and imaginary part of the parametrization for f HEGxc.Figure reproduced from Ref.[25].The coefficientsαandβare functions of the density,and can be determined uniquely by the zero and high frequency limits.A simple calculation yieldsα(n)=−A[f∞(n)−f0(n)]53,(83) where A,B>0and independent of n.Once again,by applying the Kramers-Kronig relations we can obtain the corresponding real part of f HEGxcf HEGxc (q=0,ω)=f∞+2√π√√2Π 1−r√2Π 1+r√1+βω2and E andΠare the elliptic integrals of second andthird kind.In Fig.1we plot the real and imaginary part of f HEGxc for twodifferent densities(r s=2and r s=4,where r s is the Wigner-Seitz radius, 1/n=4πr3s/3).The ALDA corresponds to approximating these curves by their zero frequency value.For very low frequencies,the ALDA is naturally a good approximation,but at higher frequencies it completely fails to reproducethe behavior of f HEGxc .To understand how the ALDA can yield such good excitation energies, albeit exhibiting such a mediocre frequency dependence,we will look at a spe-cific example,the process of photo-absorption by an atom.At low excitation frequencies,we expect the ALDA to work.As we increase the laser frequency, we start exciting deeper levels,promoting electrons from the inner shells of the atom to unoccupied states.The atomic density increases monotonically as we approach the nucleus.f xc corresponding to that larger density(lower。

The chords' problem

The chords' problem

The chords’problemAlain Daurat∗,Yan G´e rard†&Maurice Nivat‡January26,2000AbstractThe chords’problem is a variant of an old problem of computa-tional geometry:given a set of points of n,one can easily build themultiset of the distances between the points of the set but the con-verse construction is known for a longtime as difficult.The problemthat we are going to investigate is also a converse construction withthe difference that it no more the one of the distances multisets butthe one of the chords’multisets.In dimension1,the old distances’problem and the chords’problem coincide with each other whereasin other dimensions,the chords’multisets contain more informationson the sets than their distances’multisets.This paper provides,indimension1,two different algorithms to reconstruct the set of pointsaccording to their chords’multiset.Thefirst one is given for its effec-tiveness in spite of an uncertain complexity whereas the second one isthefirst polynomial algorithm solving the chords’problem.At least,we will explain how to transform a chords’problem in dimension ninto an equivalent chords’problem in dimension1.∗daurat@llaic.u-clermont1.fr,Laboratoire de Logique et d’Informatique de Cler-mont1(LLAIC1),IUT,D´e partement d’Informatique,Ensemble Universitaire des C´e zeaux, B.P.86,F-63172Aubi`e re Cedex†gerard@llaic.u-clermont1.fr,Laboratoire de Logique et d’Informatique de Cler-mont1(LLAIC1),IUT,D´e partement d’Informatique,Ensemble Universitaire des C´e zeaux, B.P.86,F-63172Aubi`e re Cedex‡Maurice.Nivat@liafa.jussieu.fr,Laboratoire d’Informatique,Algorithmique, Fondements et Applications.(LIAFA),Universit´e Denis Diderot,Case7014-2,place Jussieu F-75251Paris Cedex0512 Keywords:chords’multiset,symmetrical polynomials,Minkowski’s sum1IntroductionThe chords’problem is close to the old problem of computational geometry which among others has aroused the curiosity of Paul Erd¨o s and which consists in recog-nizing the distances’multisets.The variant which interests us does not necessitate any distance and could be formulated in any group.We have not the ambition to solve the question in a so larger framework and then,we will content ourself with working withfinite subsets of n.A rigorous formulation of the chords’problem will necessitate the introduction of some classical tools used in particular in the domain of convex geometry.After the complete presentation of the chords’prob-lem and of its relations with the polynomials,we provide two algorithms to solve it in dimension1.These are both new versions of algorithms previously presented in [1]and whose we have improved the complexity.After the resolution of the chords’problem in dimension1,we will explain how a chords’problems in dimension n can be projected in dimension1so that it could be solved by one or the other of the previous algorithms.This kind of reduction is quite unusual because generally speaking,the problems in dimension n are more as complicated as in dimension1. It is however not the case for the chords’problem and we will see by using the same tools that it is neither the one of the factorization of polynomials in[X i]1≤i≤n. 2The chordsBefore giving a sense to the notion of”chord”,let us introduce a few classical definitions of mathematics.We begin with the term multiset.Definition2.1.−A multiset is a set where some elements can occur many times.Example.−M={−4,−4,−4,3,6,6,9}is a multiset of.Definition2.2.−The multiplicity of an element in a multiset is its number of occurences.Example.−The multiplicity of−4in M is3because−4occurs3times in this multiset.The multiplicity of−2in M is0because−2does not occur in M and the multiplicity of9is1.Now,we are going to introduce several operators which act on the sets or the multisets but before doing it,let us recall that we will only work withfinite sets3 andfinite multisets.It is also essential to notice that these operators can be defined in the two frameworks,the one of the sets and the one of the multisets.In both of them,the definitions are the same and then,each time,we will give only one which will be valuable in both frameworks.Thefirst operator that we can define is the symmetry:Definition2.3.−Let A be a set or a multiset,its symmetric−A is defined by−A={−a|a∈A}.Example.−The symmetric set of S={−3,5,9}is−S={3,−5,−9}and the symmetric multiset of M is−M={4,4,4,−3,−6,−6,−9}.Now,let us introduce the Minkowski’s sum:Definition2.4.−Let A and B be two sets or two multisets of n,the Minkowski’s sum of A and B is defined by A+B={a+b|a∈A and b∈B}.Remark.−In the framework of the sets,the Minkowski’s sum of two sets provides a set while in the framework of the multisets,the Minkowski’s sum of two multisets provides a multiset.It is important to see that one can also sum two sets in the framework of the multisets.Example.−In the framework of the sets,{1,3,5}+{7,9}={8,10,12,14}whereas in the framework of the multisets,we have{1,3,5}+{7,9}={8,10,10,12,12,14} and{1,1,2}+{−3,−3}={−2,−2,−2,−2,−1,−1}.Properties.−The Minkowski’s sum is associative and commutative.−The symmetry and the union are distributive with the Minkowski’s sum:we have(A∪B)+C=(A+C)∪(B+C)and−(A+B)=(−A)+(−B).−We can also notice that a translation of a set or a multiset is nothing else than its Minkowski’s sum with a singleset.The operator that we are going to use most often in this paper is the following:Definition2.5.−Let A and B be two set or multisets of n,an element of the form b−a where a∈A and b∈B is called a chord from A to B.Then,the set or multiset of the chords from A to B is defined byC(A,B)={b−a|a∈A and b∈B}.4 Examples.−We have in the framework of the sets C({0,4},{−3,2,6})={−3,2,6,−7,2}and in the framework of the multisets C({1,1,2},{1,4})={0,0,3,3,−1,2}. Remark.−The difference between b and a has a geometric sens:the chords’set or multiset C(A,B)is the set or the multiset of the vectors going from A to B.In this paper,it is not exactly the operator C(.,.)which interests us but the one which takes a set A(and no multiset)and provides the multiset C(A,A)where C(.,.)is the operator acting in the framework of the multisets.In order to reduce the number of notations,we will call it the chord’s multiset of A and denote it C(A).Example.−let A be the subset{−1,0,2,4}of,its chord’s multiset C(A)is {−1+1,−1−0,−1−2,−1−4,0+1,0−0,0−2,0−4,2+1,2−0,2−2,2−4,4+1,4−0,4−2,4−4}={−5,−4,−3,−2,−2,−1,0,0,0,0,1,2,2,3,4,5}.The operator C(.)takes a set and provides a multiset.Given a multiset,the chord’s problem consists is knowing wether it is an image of C(.)and if it is,in computing one of its antecedant.In other words,is it a chords’multiset?and if it is one,of which set of points?Properties.−Translating a set or taking its symmetric does not modify its chords’multiset.−The chords’multisets are symmetric with respect to the origin and the square of the multiplicity of0is always equal to the sum of all the multiplicities.−With the previous notations of the symmetry and of Minkowski’s sum,the chords’multiset can be defined by C(A)=A+(−A)in the framework of the multisets.Remarks.−The chords’multiset C(A)is the multiset of the vectors going from A to A.−The chords’problem has also a sense withfinite susbsets of n but in this framework,it can always be reduced to a problem in p(by using an isomorphism between the additive group generated by the chords and p).−By using convolution products,the chords’multiset and Minkowski’s sum can also be defined for infinite subsets of n.In this framework,these are some classical tools of convex geometry:the chords’multiset is also known as the difference body, the chords’problem is called the covariogram problem and for the convex sets apart from symmetry and translations,the operator which provides the difference5 body is injective.This let us think that apart from symmetry and translations, the operator C(.)could be injective for the convex sets of n.In fact,several counter-examples prove that it is false.3A new pyramid algorithmIn this part,we provide an efficient algorithm in order to solve the chord’s problem in dimension n=1.Then let B be a multiset of which verifies the two simple properties of the chords’multisets(it is symmetric with respect to the origin and the square of the multiplicity of0is equal to the sum of all the multiplicities). The question is to know whether or not it is a chords’multiset and if it is,then we would like to know a set of points solution.Two sets translated one from the other have the same chords’multiset and then we can assume that the smallest positive element of each solution is0.By denoting l the largest element of B,we can even assume that each solution is a subset of {0,...,l}.Now,let A be any subset of{0,...,l}.For all i∈{0,...,l},we can code the existence of the point i in the set A with a variable a i equal to1if i belongs to A and to0otherwise.The multiplicity of the chord i(where i∈{0,...,l})and of the chord−i in the multiset C(A)is equal to the sum a0a i+a1a i+1+......+ a j a i+j+......+a l−i a l.By denoting b i the multiplicity of i and−i in the symmetric multiset B,we have C(A)=B if and only ifa0a l=b la0a l−1+a1a l=b l−1a0a l−2+a1a l−1+a2a l=b l−2 ............................................a0a i+a1a i+1+......+a j a i+j+......+a l−i a l=b i .......................................................................................a0a1+a1a2+...................+a j a j+1+....................+a l−1a l=b1a0a0+a1a1+.................................+a j.a j+..........................+a l.a l=b0Then we can transform the chords’problems in dimension1in systems of equa-tions,where(b0,b1,···,b j,···,b l)is the given and where(a0,a1,···,a i,···a l),sup-posed in{0,1}l+1,is unknown.A classical pyramidal algorithm can be used to solve these systems:you solve the first equation,then the second one,then the third.....Sometimes,you have more than one solution for solving an equation.In that case you need to make a choice, which will sometimes lead to a general solution and sometimes not(then you have6 to make a different choice and start again).Zhang proved that this algorithm is exponential[2].Our algorithm follows the same main principle,but with some improvements.We use an idea of Elias Tahhan Bittar which consists in postponing some choices until we have better exploited the system.Performing this algorithm,we transform the initial system.The general form of the equations during the process isk∈K a k+ (m,m )∈M a m.a m =dwhere K is a multiset of indices(an index can not occur more than two times),M a subset of{0,···,l}2and d a possibly negative integer.Among all the equations of the system,we distinguishe three special types:•I.Case d<0or d>card(K)+card(M).Those equations have no solution, and as soon as we detect such one,we can be sure that the choices we have done were bad.•II.(a)case d>card(K)+card(M).This equation imposes a k=1,a m=1,a m =1for all k∈K and(m,m )∈M.(b)case d=0.This equation imposes a k=0for all k∈K.•III.Case card(K)=2,card(M)=0,d=1.The form of these equations is a k+a k =1.This last type of equation is quite simple and some other equations where the sum a k+a k appears can become simplified.A new pyramid algorithm can be the following:As in the classical algorithm,we solve thefirst equation,then the second one,then the third....and sometimes,we have to make some choices.The new idea is that between the solvation of two following equations,we use a recursive process:•We read all the system.•If we meet an equation of the form I,then the system has no solution.•If we meet an equation of the form II,then we have some new informations about the solutions,we substitute in the system the new values that we know and we obtain a new system.•If we meet an equation of the form III,then we try to simplify the other equations.7•If we have transformed the system then we restart the process.So we obtain a new pyramid algorithm,which differs from the old classical one by a more complete analysis of the equations.The examples of Zhang,which prove that the classical pyramid algorithm is exponential do not run anymore.Is this new algorithm still exponential?Probably(even if it is not proved for the moment),but we can observe in practical that it is efficient.4Relations between multisets and polynomials Now,we are going to consider the relations between the multisets and the poly-nomials in order to interpret the chords’problem in terms of polynomials.The fundamental relation which interests us concerns the multisets of n and the poly-nomials of n variables with coefficients in.One can even say that these poly-nomials and these multisets are a lonely thing,seen in two different ways.Letfor instance P(X)= d i=1a i X i be any polynomial of one variable with all its coefficients a i in.We can also write it P(X)= i∈P X i where P is the mul-tiset,where for all i∈,the multiplicity of i is exactly a i.We have,in this way,a natural one to one map between the polynomials P(X)with coefficients in and thefinite multisets P of.There exists also the same correspondance between the polynomials P(X i)1≤i≤n of n variables with coefficients in and the finite multisets P of n.We can also notice that the sets of n correspond to the polynomials of n variables whose coefficients belong to{0,1}.By going from one way of thinking to the other,the addition of the polynomials corresponds to the union of the multisets and the product of the polynomials corresponds,in the framework of multisets,to the Minkowski’s sum.Let us explain shortly this last point:LetP(X1,X2,···X n)= (α1,α2,···αn)∈P Xα11.Xα22.···.Xαn nandQ(X1,X2,···X n)= (β1,β2,···βn)∈Q Xβ11.Xβ22.···.Xβn nbe two polynomials of n variables X1,X2,...X n with coefficients in and re-spectively associated with the multisets P and Q of n.The product of the polynomials P(X1,X2,···X n).Q(X1,X2,···X n)is equal to(α1,α2,···αn)∈P and(β1,β2,···βn)∈Q Xα1+β11.Xα2+β22.···.Xαn+βnn8 namelyXγ11.Xγ22.···.Xγn n.(γ1,γ2,···γn)∈P+QRemark.−The product of the two polynomials corresponds to the Minkowski’s sum of the two associated multisets because these are both some convolutions’products.Now we are going to use this correspondance in order to give a formulation of the chords’problem in dimension1in terms of polynomials:Let B be a multiset of which is symmetric with respect to the origin and for which the square of the multiplicity of0is equal to the sum of all the multiplicities.The chord’s problem is the following:Is B a chords’multiset?and if it is,of which set of points?As we have already explained for the pyramid algorithm,by denoting l the largest element of B,we can assume that each solution is included in{0,...,l}.The points0and l necessarily belong to each solution because otherwise,l could not be one of their chord.We remind that C(A)=A+(−A)and then,we can say that the solutions are the subsets A of{0,...,l},containing0and l and verifyingA+(−A)=B.(1) Now,we are going to translate the sets and the multisets of this equality so that they would be all included in.We have noticed that the Minkowski’s sum of a set with a singleset is nothing else than a translation.Then by adding{l}to each side of the previous equality,we obtain the equivalent equality A+(−A)+{l}=B+{l}. The Minkowski’s sum is commutative and associative and it follows thatA+({l}+(−A))=B+{l}.(2) The sets A,{l}+(−A)and the multiset B+{l}are all included in and then one can associate them respectively with the polynomials A(X),SA(X)which have no other coefficients than some0and1,and with the polynomial B(X).By using the correspondance that we have described previously,the equality(2)is equivalent toA(X).SA(X)=B(X).(3) We have chosen the notation SA(X)for the polynomial associated with{l}+(−A) because by construction,this polynomial is the symmetric of the polynomial A(X). Let us remind to the reader with two exemples how is defined this symmetry S of polynomials:if for instance,P(X)=2+3X+5X3then SP(X)=5+3X2+2X39 and if P(X)=4X+5X2then SP(X)=5X+4X2.The symmetry of the polynomials has several properties,for instance towards irreducibility,but we will recall and use them later.At this step,we have showed the following theorem that one can attribute to Rosenblatt-Seymour:Theorem4.1.−The multiset B of whose smallest element is denoted−l is a chords’multiset if and only if there exists a polynomial A(X)whose coefficients belong to{0,1}and which verifiesA(X).SA(X)=B(X)(3) where B(X)is the polynomial associated with the multiset B+{l}of and where S denotes the symmetry of polynomials.Everything is not told in the theorem and it is better to add some details:the multiset B can be supposed symmetric because otherwise,there can not exist any solution and it has the consequence that the polynomial B(X)is also symmetric. The degree of B(X)is2l,the degree of A(X)and of its symmetric is l.The coefficients of degree0and l in A(X)are1and it follows that the coefficients of degree0and2l in B(X)should also be1because otherwise there can neither exist any solution.The theorem4.1provides a formulation of the chords’problem in terms of poly-nomials and now,we are going to make the most of it in order to provide a new algorithm to solve the chords’problem in dimension1.5A new version of the Rosenblatt-Seymour algo-rithmThe chords’problem can be formulated in terms of polynomials:given a sym-metric multiset B,one constructs the symmetric polynomial B(X)of degree2l with a constant coefficient different from0and the problem is to compute the set of the polynomials A(X)whose coefficients belong to{0,1}and which verify A(X).SA(X)=B(X).According to the theorem4.1,the polynomials A(X)solu-tions correspond to the sets of points A⊂{0,...,l}whose chords’multiset is B. Notations.−S still represents the symmetry of the polynomials.We introduce the set E of the polynomials P(X)of[X]which verify P(X).SP(X)=B(X).We denote also E{0,1}the subset of E containing the polynomials whose coefficients belong to{0,1}and E{0,−1}the subset of E containing the polynomials whose coefficients belong to{0,1}.10 Remark.−We can notice that if a polynomial P(X)belongs to E,then it is also the case for−P(X).It follows that P(X)∈E{0,1}if and only if−P(X)∈E{0,−1}.With these notations,the chords’problem consits in computing E{0,1}and ac-cording to the previous remarks,one can consider that it is equivalent to compute E{0,1}∪E{0,−1}.In fact,we have introduced the sets E and E{0,1}∪E{0,−1}for two main reasons:-Thefirst one is that E is not difficult to compute,and that will be the subject of the part5.1.-The second reason is that according to the lemma5.1coming in part5.2,the set E{0,1}∪E{0,−1}that we consider as the solutions of the chords’problem,is either equal to E or wether empty.It is then possible to know this subset of E without having to examine all the polynomials of E.-As conclusion,we will see in part5.3,that it provides an algorithm to compute E{0,1}∪E{0,−1}and then solve the chords’problem.5.1Computation of EThe computation of E uses the properties of the symmetry of the polynomials. Properties of the symmetry of the polynomials.−The main property is that the symmetry is distributive towards the polynomials’product:for all polynomials P and Q in[X],we have SP.SQ=S(P.Q).The distributivity of the symme-try is also true in the other rings of polynomials but we do not need it there. Now,let Q(X)be any polynomial of[X].A consequence of the distributivity is that a polynomial Q is irreducible if and only if its symmetric SQ is irreducible. It follows thatΠi∈I Qαi i is the factorization in the lowest terms of Q in[X]if and only ifΠi∈I SQαi i is the one of SQ.One can deduce from it that if Q is symmetric(Q=SQ)then its factorization in its lowest terms is of the form Q=(Πi∈I Qαi i.SQαi i).(Πj∈J Qβj j)where the indices i∈I are used for the non-symmetric irreducible factors and the indices j∈J for the symmetric factors. This last property provides a characterization of the polynomials in E: Theorem5.2.−We denote B=(Πi∈I Bαi i.SBαi i).(Πj∈J Bβj j)the factorization of the symmetric polynomial B in its lowest terms in[X],where the indices i∈I are used for the non-symmetric irreducible factors and the indices j∈J for the symmetric factors.If there exists an odd exponentβj then E is empty andotherwise E is composed of the polynomials P=(Πi∈I Bγi i.SBδi i).(Πj∈J Bβj/2j )withsome exponentsγi∈andδi∈verifyingγi+δi=αi.11 This theorem gives a description of E according to the factorization of B(X)in its lowest terms.5.2Relations between E and its subset E{O,1}∪E{O,−1}This part contains a lemma that we have already mentionned by saying that E{0,1}∪E{0,−1}is either equal to E or wether empty.This lemma comes from [1]where it is given under a different form and used differently.On the other hand,the proof that we provide is exactly the one which is given by Skiena, Smith and Lemke in[1].Lemma5.1.−Only two cases are possible:E0,1=∅or E=E0,1∪E0,−1. Proof.Let us remind that the degree of the polynomial B(X)is2l and then that all the polynomials P in E are of degree l.The constant coefficient of the polynomials in E can not be null because they verify P(0).SP(0)=B(0)which is by construction different from0.We suppose that the set E0,1is not empty and we take anyone of its elements A(X)= l i=1a i X i.For all i∈{0,...,l},the coefficient a i is either1wether0.In other words,all the coefficients verify a2i=a i and we will use it in the following. Now,we take any element P(X)= l i=1p i X i of E and we are going to prove that it belongs to E0,1∪E0,−1.As a0and p0are not null,we have SA(X)= l i=1a l−i X i and SP(X)= l i=1p l−i X i. Since A(X)and P(X)belong to E,we have A(X).SA(X)=P(X).SP(X)=B(X) and it implies two equalities:for x=1,we obtain( l i=0a i)2=( l i=0p i)2or namelyl i=0a i=±l i=0p i.(4) and by equality of the two coefficients of degree n,we obtainl i=0(a2i)=l i=0(p2i).(5)First case.−We suppose l i=0a i=+ l i=0p i.(4+) The coefficients a i are only some0and some1and then we have l i=0a i= l i=0(a2i).By combining this last equality with(4+)and(5),we obtain that l i=0p i= l i=0(p2i)namely li=0(p2i−p i)=0.(6+)12 As for all i the numbers p i are integer,the values p2i−p i are all positive and according to(6+),their sum is null.It implies that p2i−p i=0for all i∈[0,l]. The two solutions of the equation x2−x=0are0and1,and we conclude that all the coefficients p i are in{0,1}or in other words that P(X)belongs to E0,1. Second case.−We suppose l i=0a i=− l i=0p i.(4−) Then the coefficients p i=−p i verify(4+)and(5).According to thefirst case,it follows that for all i in[0,l]we have p i∈{0,1}and that proves that P(X)belongs to E0,−1.2 5.3Description of an algorithm computing E{0,1}∪E{0,−1}Given a symmetric multiset B,we construct the symmetric polynomial B(X)of degree2l with a constant coefficient different from0and which is associated with B+{l}.Now,we are going to give an algorithm in polynomial time in order to compute E{0,1}∪E{0,−1}.The polynomials A(X)of this set correspond to the sets of points A included in{0,...,l}and whose chords’multiset is B.Remark.−In the algorithm,we can assume that the constant coefficient of B(X) is1because otherwise we know that there does not exist any solution.This will allow us to have a better theoritical complexity.First step.−We factorize the symmetric polynomial B(X)in its lowest terms in [X]by using the LLL-algorithm.In the general case,this factorization takes a pseudo-polynomial time but as we have assumed that the constant coefficient is1, it takes a polynomial time[3].According to the part5.1,the factorization of the symmetric polynomial B in[X]is of the form B=(Πi∈I Bαi i.SBαi i).(Πj∈J Bβj j), where the indices i∈I are used for the non-symmetric irreducible factors and the indices j∈J for the symmetric factors.Second step.−If there exists an odd exponentβj then E{0,1}∪E{0,−1}is empty. There does not exist any solution and we stop.Last step.−The exponentsβj are all even and then,we compute the polynomialP=(Πi∈I Bαi i).(Πj∈J Bβj/2j )which belongs to E.If all the coefficients of P are in{0,1}or if they are all in{0,−1}then E{0,1}∪E{0,−1}is not empty and according to the lemma5.1,it implies E=E{0,1}∪E{0,−1}.Otherwise the lemma ensures that E{0,1}∪E{0,−1}is empty and then that there does not exist any solution. At the end of the algorithm,wether we know that there are not any solution either we know a polynomial of E{0,1}∪E{0,−1}and in that case,we can obtain13all the solutions with the formula of the elements of E given in theorem 5.1.The compexity of this algorithm is polynomial.It is proved in [1]that the number of solutions of a chords’problem is less than a polynomial boundary.This property is important to increase the complexity of the algorithm which consists in computing all the elements of E before seeing if one of them belongs to E 0,1,because it shows that this classical version of the Rosenblatt-Seymour algorithm is pseudo-polynomial (the new version that we have exposed is polynomial).Now that we have an efficient algorithm and a polynomial one in order to solve the chords’problem in dimension 1,let us investigate the problem in any dimension.6Transformation of the chords’problem by a mor-phismAlthough we have not given any example in another dimension than 1,the def-initions of the operators acting on the sets or multisets were in any dimension.Example .−Let A be the set {(0,3);(1,1);(2,1);(3,1)}drawn fig 1.s s s s --01234Figure 1.−The subset A of 2.The chords’multiset of A is C (A )={(0,3)−(0,3);(1,1)−(0,3);(2,1)−(0,3);(3,1)−(0,3);(0,3)−(1,1);(1,1)−(1,1);(2,1)−(1,1)...........}and we obtain C (A )={(−3,2);(−2,2);(−1,2);(−2,0);(−1,0);(−1,0);(0,0);(0,0);(0,0);(0,0);(1,0);(1,0);(2,0);(1,−2);(2,−2);(3,−2)}(fig 2)14s 4s 2s 2s 1s 1s 1s 1s 1s 1s 1s 1-----4-3-2-101234Figure 2.−The chords’multiset C (A )with the multiplicity of each one of its points.The general aim of this part is to develop a technical in order to transform a chords’problem in dimension n into an equivalent chords’problem in dimension1.This technical exceeds the single framework of the chords and we are going to present it for any operator K defined by symmetries and Minkowski’s sums.6.1Introduction of a general operator KLet K be an operator acting on the sets or multisets and defined by combining some symmetries with some Minkowski’s sums.As with the Minkowski’s sum or the symmetry,we do not precise if we are in the framework of the sets or in the one of the multisets because in each one of both,the properties are the same.The operator K could be the Minkowski’s sum itself,the symmetry itself,the chords’operator or almost any other operator of this kind.We just suppose that each set or multiset given as argument appears exactly one time in the expression of K .It leaves out for instance the operators DONALD (A,B,D,E )=A −D −E or C (A )=A −A .Example .−The operator K could be the one which takes four sets or multisets A ,B ,D ,E ,and associates with them K (A,B,D,E )=A +B −D +E .From now until the end,we suppose that K is an operator of this kind and we denote k the number of its arguments.Remarks .−We can also define the action of the operator K on the points of n .For the example which is above,we have K ({a },{b },{d },{e })={a +b −d +e }where a ,b ,d and e are supposed in n and then it is natural to define the action of K on the points by K (a,b,d,e )=a +b −d +e .−For any indice i ∈[1,k ],we have K (A 1,A 2,···,A i ∪A i ,···A k )=K (A 1,A 2,···。

$N=2$ and $N=4$ SUSY Yang-Mills action on $M^4times (Z_2oplus Z_2)$ non-commutative geometr

$N=2$ and $N=4$ SUSY Yang-Mills action on $M^4times (Z_2oplus Z_2)$ non-commutative geometr

a rXiv:h ep-th/9411158v122N ov1994ASITP-94-61November,1994.N =2and N =4SUSY Yang-Mills action on M 4×(Z 2⊕Z 2)non-commutative geometry Bin Chen 1Hong-Bo Teng 2Ke Wu 3Institute of Theoretical Physics,Academia Sinica,P.O.Box 2735,Beijing 100080,China.Abstract We show that the N =2and N =4SUSY Yang-Mills action can be reformulated in the sense of non-commutative geometry on M 4×(Z 2⊕Z 2)in a rather simple way.In this way thescalars or pseudoscalars are viewed as gauge fields along two directions in the space of one-forms on Z 2⊕Z 2.1IntroductionSince A.Connes introduced his non-commutative geometry into particle physics[1,2],a successful geometrical interpretation of the Higgs mechanism and Yukawa coupling has been acquired.In this interpretation,the Higgsfields are regarded as gaugefields on the discrete gauge group. The bosonic part of the action is just the pure YM action containing the gaugefields on both continuous and discrete gauge group,and the Yukawa coupling is viewed as a kind of gauge interactions of fermions.At the same time,applying non-commutative geometry to SUSY theories has encountered many difficulties.A natural way is to introduce a non-commutative space which is the product of the superspace and a set of discrete points,similar to those which have been done in non-SUSY theories.However,it proved that such an extension of superspace is rather difficult to accomplish.In[3],A.H.Chamseddine gave an alternative way in which the supersymmetry theories in their component forms were considered.He showed that N=2and N=4SUSY Yang-Mills actions could be derived as action functionals for non-commutative spaces,and in some cases the coupling of N=1and N=2super Yang-Mills to N=1and N=2matter could also be reformulated.This paper is thefirst one in which the non-commutative geometry is successfully applied to SUSY theories.Intrigued by A.H.Chamseddine’s work,we show the N=2and N=4SUSY Yang-Mills actions can be derived with non-commutative geometry on M4×(Z2⊕Z2)in a rather simple way.We use the differential calculus on discrete group developed by A.Sitarz[4].With a specific case of differential algebra on Z2⊕Z2,we take the scalar components of the super Yang-Mills field as gaugefields on discrete symmetry group.In N=2case,the scalar and pseudoscalar fields are set to two directions of the space of one-forms.In case of N=4,the complex scalar fields which belong to6of SU(4)are regarded as components of the connection one-form.First in section2we show the specific case of the differential algebra on Z2⊕Z2.Then in section3we derive the N=2and N=4super Yang-Mills Lagrangian with this algebra.Finally we end with conclusions.2Differential Calculus on Z2⊕Z2In this section,we briefly describe the differential calculus on Z2⊕Z2and introduce a specific case of this algebra.For a detailed description of differential calculus on discrete group,we refer to[4].Let’s write the four elements of Z2⊕Z2as(e1,e2),(r1,e2),(e1,r2),(r1,r2).And the group multiplication is(g1,g2)(h1,h2)=(g1h1,g2h2).(1)Let A be the algebra of complex valued functions on Z2⊕Z2.The derivative on A is defined as∂g f=f−R g f g∈Z2⊕Z2,f∈A(2)with R g f(h)=f(hg).We will write∂i and R i for convenience where i=1,2,3refers to (r1,e2),(e1,r2),(r1,r2)respectively.{∂i,i=1,2,3}forms the basis of F,the space of left invariant vectorfields over A.One canfind that the following relations hold∂1∂2=∂1+∂2−∂3∂2∂3=∂2+∂3−∂1(3)∂3∂1=∂3+∂1−∂2∂1∂1=2∂1∂2∂2=2∂2(4)∂3∂3=2∂3∂1∂2=∂2∂1∂1∂3=∂3∂1(5)∂2∂3=∂3∂2.The Haar integral on Z2⊕Z2is defined asZ2⊕Z2f=1Next we introduce the space of one-forms,Ω1,which is the dual space of F.Haven chosen the basis of F,we automatically have the dual basis ofΩ1,{χi,i=1,2,3}.It is defined withχi(∂j)=δi j i,j=1,2,3.(7) The definition of higher forms is natural,Ωn is taken to be tensor product of n copies ofΩ1,,Ωn=Ω1⊗···⊗Ω1n copiesandΩ0=A.AndΩ= ⊕Ωn is the tensor algebra on Z2⊕Z2.As usual,the external derivative satisfies the graded Leibniz rule and is nilpotentd(ab)=d(a)b+(−1)dega a(db)a,b∈Ω(8)d2=0and for f∈Ω0=Ad f=∂1fχ1+∂2fχ2+∂3fχ3.(9)From(8),(9),onefindsχi f=(R i f)χi i=1,2,3f∈A(10)anddχ1=−χ1⊗χ2−χ1⊗χ3+χ2⊗χ3−2χ1⊗χ1−χ2⊗χ1−χ3⊗χ1+χ3⊗χ2dχ2=−χ2⊗χ3−χ2⊗χ1+χ3⊗χ1−2χ2⊗χ2−χ3⊗χ2−χ1⊗χ2+χ1⊗χ3(11)dχ3=−χ3⊗χ1−χ3⊗χ2+χ1⊗χ2−2χ3⊗χ3−χ1⊗χ3−χ2⊗χ3+χ2⊗χ1We also need the involution¯on our differential algebra,which agrees with the complex conjugation on A and(graded)commutes with d,i.e.d(da.For one-forms,then we getdχi=0i=1,2,3.(14) It is a specific case of the above algebra.Obviously the postulation induces an antisymmetric tensor algebraΛwhich is a subalgebra ofΩ.In this subalgebra,we denote the product of forms with∧.And we will use this subalgebra to derive the N=2and N=4SUSY Yang-Mills action in next section.3N=2and N=4super Yang-Mills actionLet’sfirst consider the N=2case.The N=2super Yang-Mills Lagrangian is given byL=T r(−12DµSDµS+12[S,P]2)(15)whereψis the Dirac4−spinor which is the combination of two2−spinors,S is the scalar and P is the pseudoscalar.Allfields are valued in the adjoint representation of the gauge group,i.e.,f=f a T a f=Aµ,S,P,ψ(16)where T a is the generator of gauge group.AndFµν=∂µAν−∂νAµ+i[Aµ,Aν]DµS=∂µS+i[Aµ,S]DµP=∂µP+i[Aµ,P]Dµψ=∂µψ+i[Aµ,ψ].(17)From a simple observation,we see that it is possible to choose the generalized connection one-form on M4×(Z2⊕Z2)A(x,g)=iAµ(x,g)dxµ−iS(x,g)χ1−iγ5P(x,g)χ2g∈Z2⊕Z2.(18)And we setAµ(x,g)=Aµ(x);S(x,g)=S(x);P(x,g)=P(x);g∈Z2⊕Z2.(19)And later in fermionic sector we will also letψ(x,g)=ψ(x).Since all quantities take the same values on different points of Z2⊕Z2,we will ignore the variable g.Please note in this assignment the scalar and pseudoscalar are assigned to two directions of the connection one-form,and we have manifestly set the component inχ3direction of A to zero.We will see it is sufficient for our construction.With(13),(14),(18),(19)the curvature two form is easily foundF=dA(x)+A(x)∧A(x)=i(∂µAν+iAµAν)dxµ∧dxν−i(∂µS+i[Aµ,S])dxµ∧χ1−iγ5(∂µP+i[Aµ,P])dxµ∧χ2−γ5[S,P]χ1∧χ2=iF(h)>(21) we get exactly the bosonic sector of the LagrangianL B=T r(1dxρ∧dxσ>+DµS(DνS)∗<dxµ∧χ1,dxν∧χ2>+[S,P][S,P]∗<χ1∧χ2,4FµνFµν+12DµP DµP+1dxρ∧dxσ>=<dxµ∧dxν,dxσ∧dxρ>=gµσgνρ<dxµ∧χ1,2gµν<dxµ∧χ2,2gµν<χ1∧χ2,2.(23)As to the fermionic sector of the Lagrangian,we haveL F= Z2⊕Z2T r(i¯ψ[γµDµ+D1+D2+D3,ψ])=T r(i¯ψγµDµψ+¯ψ[S+γ5P,ψ])(24)where[D1,ψ]=∂1ψ−i[SR1,ψ]=−i[S,ψ][D2,ψ]=∂2ψ−i[γ5P R2,ψ]=−i[γ5P,ψ][D3,ψ]=∂3ψ=0.(25) So the N=2super Yang-Mills Lagrangian is reformulated.We next consider N=4case,the Lagrangian isL=T r[−12DµM ij DµM ij+iλi[M ij,λj]+i¯λi[M ij,¯λj]+12ǫijkl M kl=M ij,and transforms as6of SU(4), andλi is the2−spinor transforming as4of SU(4)whereas¯λi transforming as¯4.So L is SU(4) invariant.In order to write the spinor in four component form,we introduce Majorana spinorsλi= λαi¯λ˙αi ;¯λi= λαi,¯λi˙α (27) andφij= M ij M ijI ij= δj iδi j (28)α=1,2and˙α=˙1,˙2are spinor indices.So the Lagrangian can be written asL=T r[−12¯λiγµDµλi+18[φij,φkl][φ†ij,φ†kl]].(29)here the trace also takes on the2×2matrix in whichfields in two2−component form are arranged into4−component form.We set the connection one-form on M4×(Z2⊕Z2)A(x,g)=iAµ(x,g)I ij dxµ+φij(x,g)χ1+φij(x,g)χ2g∈Z2⊕Z2.(30)Again we will let allfields take the same values on different points of Z2⊕Z2,so we will ignore the variable g,and as in the N=2case,we let the component inχ3direction be zero.Then we haveF=dA(x)+A(x)∧A(x)=i2 Z2⊕Z2T r<F(h),2T r(1dxρ∧dxσ>+Dµφij(Dνφij)∗<dxµ∧χ1,dxν∧χ2>+[φij,φkl][φij,φkl]∗<χ1∧χ2,8FµνFµνI2+18[φij,φkl][φ†ij,φ†kl])(32)here again in order to give appropriate normalization constants,we have chosen the metric <dxµ∧dxν,dxν∧χ1>=<dxµ∧χ1,−χ1∧dxν>=1dxν∧χ2>=<dxµ∧χ2,−χ2∧dxν>=1χ1∧χ2>=<χ1∧χ2,χ2∧χ1>=12 Z2⊕Z2T r(i¯λi[γµ(Dµ)I ij+(D1)ij+(D2)ij+(D3)ij,λj])=T r(i4ConclusionsWe show that the N=2and N=4super Yang-Mills Lagrangian can be reformulated on M4×(Z2⊕Z2)non-commutative geometry.We use the differential calculus on discrete group developed by A.Sitarz and regard the scalar or pseudoscalarfields as connection one-forms on the discrete symmetry group.This reformulation is different from that of A.H.Chamseddine and seems much simpler and clearer.References[1]A.Connes:Non-commutative Geometry English translation ofGeometrie Non-commutative,IHES Paris,Interedition.[2]A.Connes and J.Lott:Nucl.Phys.(Proc.Suppl.)B18,44(1990).[3]A.H.Chamseddine:Phys.Lett.B332,(1994)349-357.[4]A.Sitarz:To appear in J.Geo.Phys.[5]A.H Chamseddine,G Felder and J.Frohlich:Phys.Lett.296B(1993)109.[6]R.Coquereaux,G.Esposito-Farese and G.V aillant:Nucl.Phys.B353689(1991).。

Threshold Perspectives on Meson Production

Threshold Perspectives on Meson Production

a rXiv:n ucl-e x /31111v112N ov23Threshold Perspectives on Meson Production M.Wolke 1The Svedberg Laboratory,Uppsala University,Box 533,75121Uppsala,Sweden ∗Abstract.Studies of meson production in nucleon–nucleon collisions at threshold are characterised by few degrees of freedom in a configuration of well defined initial and final states with a transition governed by short range dynamics.Effects from low–energy scattering in the exit channel are inherent to the data and probe the interaction in baryon–meson and meson–meson systems otherwise difficult to access.From dedicated experiments at the present generation of cooler rings precise data are becoming available on differential and eventually spin observables allowing detailed comparisons between complementary final states.To discuss physics implications of generic and specific properties,recent experimental results on meson production in proton–proton scattering obtained at CELSIUS and COSY serve as a guideline.INTRODUCTION High precision data from the present generation of cooler rings,IUCF,CELSIUS,and COSY ,have contributed significantly over the last decade to our present knowledge and understanding of threshold meson production (for a recent review see [1]).Due to the high momentum transfers required to create a meson or mesonic system in production experiments close to threshold the short range part of the interaction is probed.In nucleon–nucleon scattering,for mesons in the mass range up to 1GeV /c 2distances from 0.53fm (π0)down to less than 0.2fm (φ)are involved.At such short distances it is a priori not clear,whether the relevant degrees of freedom are still baryons and mesons,or rather quarks and gluons.As there is no well defined boundary,one goalof the threshold production approach is to explore the limits in momentum transfer for a consistent description using hadronic meson exchange models.Within this framework,questions concerning both the underlying meson exchange contributions and especially the role of intermediate baryon resonances have to be answered.Another aspect which enriches the field of study arises from the low relative centre–of–mass velocities of the ejectiles:Effects of low energy scattering are inherent to the observables due to strong final state interactions (FSI)within the baryon–baryon,baryon–meson,and meson–meson subsystems.In case of short–lived particles,low energy scattering potentials are otherwise difficult or impossible to study directly.DYNAMICS OF THE TWO PION SYSTEMInγandπinduced double pion production on the nucleon the excitation of the N∗(1440) P11resonance followed by its decay to the Nσchannel,i.e.N∗(1440)→p(ππ)I=l=0, is found to contribute non–negligibly close to threshold[2,3,4].Nucleon–nucleon scattering should provide complementary information,eventually on theππdecay mode of the N∗(1440),which plays an important part in understanding the basic structure of the second excited state of the nucleon[5,6,7].Exclusive CELSIUS data from the PROMICE/W ASA setup on the reactions pp→ppπ+π−,pp→ppπ0π0and pp→pnπ+π0[8,9,10]are well described by model cal-culations[11]:For theπ+π−andπ0π0channels,the reaction preferentially proceeds close to threshold via heavy meson exchange and excitation of the N∗(1440)Roper resonance,with a subsequent pure s–wave decay to the Nσchannel1.While nonres-onant contributions are expected to be small,resonant processes with Roper excita-tion and decay via an intermediate∆(pp→pN∗→p∆π→ppππ)and∆∆excitation (pp→∆∆→pπpπ)are strongly momentum dependent and vanish directly at thresh-old.Double∆excitation,which is expected to dominate at higher excess energies be-yond Q=250MeV[11]involves higher angular momenta and consequently strongly anisotropic proton and pion angular distributions.On the other hand,the Roper decay amplitude via an intermediate∆depends predominantly on a term symmetric in the pion momenta(eq.(1)),leading to the p(π+π−)I=l=0channel and an interference with the direct Nσdecay.Experimentally,for the reaction pp→ppπ+π−at excess energies of Q=64.4MeV and Q=75MeV angular distributions give evidence for only s–waves in thefinal state,in line with a dominating pp→pN∗→pp(π+π−)I=l=0process,with the initial inelastic pp collision governed by heavy meson(σ,ρ)exchange.Roper excitation is disclosed in the pπ+π−invariant mass distribution(Fig.1a),where the data are shifted towards higher invariant masses compared to phase space in agreement with resonance excitation in the low energy tail of the N∗(1440).Compared with Monte Carlo simulations including both heavy meson exchange for N∗excitation,and pp S–wave final state interaction,but only the direct decay N∗→p(π+π−)I=l=0(dotted lines),the production process involves additional dynamics,which is apparent from discrepancies especially in observables depending on theπmomentum correlation k1· k2,i.e.π+π−invariant mass Mππ(Fig.1b)and opening angleδππ=1For the pnπ+π0final state,this reaction mechanism is trivially forbidden by isospin conservation.An underestimation of the total cross section data[9]by the model predictions[11]might be explained by the neglect of effects from the pnfinal state interaction in the calculation[12].M π+π- [MeV/c 2]d σ/d M π+π- [n b /(M e V /c 2)]-π+πδcos d σ/d c o s δπ+π- [n b ]a)M p π+π- [MeV/c 2]d σ/d M p π+π- [n b /(M e V /c 2)]b)c)FIGURE 1.Differential cross sections for the reaction pp →pp π+π−at an excess energy of Q =75MeV.Experimental data (solid circles)for invariant mass distributions of the (p π+π−)–(a)and(π+π−)–subsystems (b),and the π+π−opening angle (c)are compared to pure phase space (shaded areas)and Monte Carlo simulations for direct decays N ∗→N σ(dotted lines),decays via an intermediate ∆resonance N ∗→∆π→N σ(dashed lines)and an interference of the two decay routes (solid lines)according to eq.(1).Figures are taken from [10].allows to determine the ratio of partial decay widths R (M N ∗)=ΓN ∗→∆π→N ππ/ΓN ∗→N σat average masses <M N ∗>corresponding to excess energies Q =64.4MeV and Q =75MeV relative to the π+π−threshold.The numerical results,R (1264)=0.034±0.004and R (1272)=0.054±0.006,exhibit the clear dominance of the direct decay to the N σchannel in the low energy region of the Roper resonance.On the other hand they indicate the strong energy dependence of the ratio from the momentum dependence in the decay branch via an intermediate ∆,which will surpass the direct decay at higher energies [10].A model dependent extrapolation based on the validity of ansatz (1)leads to R (1440)=3.9±0.3at the nominal resonance pole in good agreement with the PDG value of 4±2[13].Within the experimental programme to determine the energy dependence of the N ∗→N ππdecay exclusive data (for details see [14])have been taken simultaneously at the CELSIUS/W ASA facility on both the pp π+π−and pp π0π0final states.In case of the π+π−system the preliminary results at an excess energy of Q =75MeV are in good agreement with the relative strength of the decay routes adjusted to an extrapolated ratio R (1440)=3.However,at slightly higher excess energy (Q =127MeV)the data might be equally well described by a value R (1440)=1,which is noticeably favoured at both excess energies by the data on π0π0production,indicating distinct underlying dynam-ics in π0π0and π+π−production.One difference becomes obvious from the isospin decomposition of the total cross section [9]:An isospin I =1amplitude in the ππsys-tem,and accordingly a p–wave admixture,is forbidden by symmetry to contribute to the neutral pion system in contrast to the charged complement.A p–wave component was neglected so far in the analysis,since the unpolarized angular distributions show no devi-ation from isotropy.However,there is evidence for small,but non–negligible analysing powers from a first exclusive measurement of π+π−production with a polarized beam at the COSY–TOF facility [14,15],suggesting higher partial waves especially in the ππsystem.At higher energies,i.e.Q =208MeV and Q =286MeV with respect to the π+π−threshold,preliminary data for both π+π−and π0π0from CELSIUS/W ASA rather fol-low phasespace than expectations based on a dominating pp →pN ∗→pp σreaction mechanism[14].At these energies,the ∆∆excitation process should influence observ-ables significantly,and,thus,a phase space behaviour becomes even more surprising,unless the ∆∆system is excited in a correlated way.THE PROTON–PROTON–ETA FINAL STATEAs a general trait in meson production in nucleon–nucleon scattering,the primary pro-duction amplitude,i.e.the underlying dynamics can be regarded as energy independent in the vicinity of threshold [16,17,18].Consequently,for s–wave production processes,the energy dependence of the total cross section is essentially given by a phase space behaviour modified by the influence of final state interactions.In Fig.2total cross sec-tion data obtained in proton–proton scattering are shown for the pseudoscalar isosinglet mesons ηand η′[19].In both cases,the energy dependence of the total cross sec-σ[n b ]10101010excess energy [MeV]110102FIGURE 2.Total cross section data for η(squares [20,21,22,23,24,25])and η′(circles [24,26,27,28,29])production in proton–proton scattering versus excess energy Q [19].In comparison,the energy dependences from a pure phase space behaviour (dotted lines,normalized arbitrarily),from phase space modified by the 1S 0proton–proton FSI including Coulomb interaction (solid lines),and from additionally including the proton–ηinteraction phenomenologically (dashed line),are shown.Meson exchange calculations for ηproduction including a P–wave component in the proton–proton system [30]are depicted by the dashed–dotted line,while the dashed–double–dotted line corresponds to the arbitrarily normalized energy dependence from a full three–body treatment of the pp ηfinal state [31](see also [32]).tion deviates significantly from phase space expectations.Including the on–shell 1S 0proton–proton FSI enhances the cross section close to threshold by more than an order of magnitude,in good agreement with data in case of η′.As expected from kinemat-ical considerations[1]the cross section for ηproduction deviates from phase space including the pp FSI at excess energies Q ≥40MeV,where the 1S 0final state is no longer dominant compared to higher partial waves.Deviations at low excess energies seem to be well accounted for by an attractive proton–ηFSI (dashed line),treated phe-nomenologically as an incoherent pairwise interaction [1,17,33].In comparison to the proton–η′(Fig.2)and proton–π0systems only the p ηinteraction is strong enough to become apparent in the energy dependence of the total cross section [34].In differen-tial observables,effects should be more pronounced in the phase space region of low proton–ηinvariant masses.However,to discern effects of proton–ηscattering from the influence of proton–proton FSI,which is stronger by two orders of magnitude,requires high statistics measurements,which have only become available recently [19,35,36]:Close to threshold,the distribution of the invariant mass of the proton–proton subsys-s pp [GeV 2/c 4]d σ/d s p p [µb /(Ge V 2/c 4)]FIGURE 3.Invariant mass squared of the (pp )–subsystem in the reaction pp →pp ηat excess energies of Q =15.5MeV (COSY–11,solid circles [19],Q =15MeV (COSY–TOF,open circles [36]and Q =16MeV (PROMICE/W ASA,open triangles [35]).The dotted and dashed lines follow a pure phase space behaviour and its modification by the phenomenological treatment of the three–body FSI as an incoherent pairwise interaction,respectively.The latter was normalized at small invariant mass values.Effects from including a P–wave admixture in the pp system are depicted by the dashed–dotted line [30],while the dashed–double–dotted line corresponds to a pure s–wave final state with a full three–body treatment [32].tem is characteristically shifted towards low invariant masses compared to phase space (dotted line in Fig.3).This low–energy enhancement is well reproduced by modifying phase space with the 1S 0pp on–shell interaction.A second enhancement at higher pp invariant masses,i.e.low energy in the p ηsystem,is not accounted for even when including additionally the proton–ηinteraction incoherently (dashed line).However,in-cluding a P–wave admixture in the pp system by considering a 1S 0→3P 0s transition in addition to the 3P 0→1S 0s threshold amplitude,excellent agreement with the ex-perimental invariant mass distribution is obtained (dashed–dotted line [30]).In return,with the P–wave strength adjusted to fit the invariant mass data,the approach fails to reproduce the energy dependence of the total cross section (Fig.2)below excess ener-gies of Q =40MeV.Preliminary calculations considering only s–waves in the final statebut using a rigorous three–body treatment of the ppηfinal state actually decrease the cross section at large values of the pp invariant mass(dashed–double–dotted lines[32]) compared to an incoherent two–body calculation within the same framework.However, close to threshold the energy dependence of the total cross section is enhanced com-pared to the phenomenological incoherent treatment and the data(Fig.2).Although part of this enhancement has to be attributed to the neglect of Coulomb repulsion in the pp system,consequently overestimating the pp invariant mass at low values,qualitatively the full three–body treatment has opposite effects compared to a P–wave admixture in the proton–proton system in view of both the total cross section as well as the pp in-variant mass distribution.In the approximate description of the total cross section by the phenomenological s–wave approach with an incoherent FSI treatment these two effects seem to cancel casually.Close to threshold,resonance excitation of the S11(1535)and subsequent decay to the pηfinal state is generally2believed to be the dominantηproduction mechanism [17,30,38,39,40,41,42,43].In this context,the issue of the actual excitation mechanism of the S11(1535)remains to be addressed.Theηangular distribution is sensitive to the underlying dynamics:A dominantρexchange favoured in[41]results in an inverted curvature of theηangular distribution compared toπandηexchanges which are inferred to give the largest contribution to resonance excitation in[42].In the latter approach the interference of the pseudoscalar exchanges in the resonance current with non–resonant nucleonic and mesonic exchange currents turns the curvature to the same angular dependence as expected forρexchange.Presently,due to the statistical errors of the available unpolarised data at an excess energy of Q≈40MeV[35,36]it is not possible to differentiate between a dominantρorπ,ηexchange,as discussed in [36].Data recently taken at the CELSIUS/W ASA facility with statistics increased by an order of magnitude compared to the available data might provide an answer in the near future[44].Spin observables,like theηanalyzing power,should even disentangle a dominantρmeson exchange and the interference ofπandηexchanges in resonance excitation with small nucleonic and mesonic currents[42],which result in identical predictions for the unpolarisedηangular distribution.First data[45]seem to favour the vector dominance model,butfinal conclusions both on the underlying reaction dynamics and the admixture of higher partial waves[30]have to await the analysis of data taken with higher statistics for the energy dependence of theηanalysing power[46].ASSOCIATED STRANGENESS PRODUCTIONIn elementary hadronic interactions with no strange valence quark in the initial state the associated strangeness production provides a powerful tool to study reaction dynamicsby introducing a“tracer”to hadronic matter.Thus,quark model concepts might even-tually be related to mesonic or baryonic degrees of freedom,with the onset of quark degrees of freedom expected for kinematical situations with large enough transverse momentum transfer.First exclusive close–to–threshold data onΛandΣ0production[47,48]obtained at the COSY–11facility showed at equal excess energies below Q=13MeV a cross section ratio ofRΛ/Σ0(Q≤13MeV)=σ(pp→pK+Λ)3For further complementary theoretical approaches see references in[1,58,59].Σ0p FSI being much weaker compared to the Λp system.However,the interpretation implies dominant S–wave production and reaction dynamics that can be regarded as energy independent.Within the present level of statistics,contributions from higher partial waves can be neither ruled out nor confirmed at higher excess energies for Σ0production.The energy dependence of the production ratio RΛ/Σ0is shown in Fig.4in comparisonwith theoretical calculations obtained within the approach of [50]assuming a destructive interference of πand K exchange and employing different choices of the microscopic hyperon nucleon model to describe the interaction in the final state [61].The resultexcess energy [MeV]4060200σ (p p → p K +Λ)σ (p p→p K +Σ)010203040FIGURE 4.Λ/Σ0production ratio in proton–proton scattering as a function of the excess energy.Data are from [48](shaded area)and [60].Calculations [61]within the Jülich meson exchange model imply a destructive interference of K and πexchange using the microscopic Nijmegen NSC89(dashed line [62])and the new Jülich model (solid line [63])for the Y N final state interaction.crucially depends on the details —especially the off–shell properties —of the hyperon–nucleon interaction employed.At the present stage both the good agreement found in[50]with the threshold enhancement (2)and for the Nijmegen model (dashed line in Fig.4)with the energy dependence of the cross section ratio should rather be regarded as accidental 4.Calculations using the new Jülich model (solid line in Fig.4)do not reproduce the tendency of the experimental data.It is suggested in [61]that neglecting the energy dependence of the elementary amplitudes and higher partial waves might no longer be justified beyond excess energies of Q =20MeV.However,once the reaction mechanism for close–to–threshold hyperon production is understood,exclusive data should provide a strong constraint on the details of hyperon–nucleon interaction models.PRESENT AND FUTUREIntermediate baryon resonances emerge as a common feature in the dynamics of the exemplary cases for threshold meson production in nucleon–nucleon scattering dis-cussed in this article.However,this does not hold in general for meson production in the1GeV/c2mass range(for a discussion onη′production see[64]).Moreover,the extent to which resonances are evident in the observables or actually govern the reaction mechanism depends on the specific channels,which differ in view of the level of present experimental and theoretical understanding.The N∗(1440)resonance dominatesπ+π−production at threshold,and exclusive data allow to extract resonance decay properties in the low–energy tail of the Roper. Dynamical differences between the different isospin configurations of theππsystem and the behaviour at higher energies remains to be understood withfirst experimental clues appearing.With three strongly interacting particles in thefinal state,a consistent description ofηproduction close to threshold requires an accurate three–body approach taking into account the possible influence of higher partial waves.High statistics differential cross sections and polarization observables coming up should straighten out both the excitation mechanism of the N∗(1535)and the admixture of higher partial waves.At present,the available experimental data on the elementary strangeness produc-tion channels give evidence for both an important role of resonances coupling to the hyperon–kaon channels and on a dominant non–resonant kaon exchange mechanism. Experiments on different isospin configurations,high statistics and spin transfer mea-surements close to threshold should disentangle the situation in future.From the cornerstone of total cross section measurements,it is apparent from the above examples to what extent our knowledge is presently enlarged by differential observables and what will be the impact of polarization experiments in future to get new perspectives in threshold meson production.ACKNOWLEDGMENTSThe author gratefully acknowledges the pleasure to work with the CELSIUS/W ASA and COSY–11collaborations,and,in particular,thanks M.Bashkanov,H.Clement,R. Meier,P.Moskal and W.Oelert for helpful discussions.This work has been supported by The Swedish Foundation for International Cooperation in Research and Higher Education(STINT Kontrakt Dnr02/192).REFERENCES1.Moskal,P.,Wolke,M.,Khoukaz,A.,and Oelert,W.,Prog.Part.Nucl.Phys.,49,1–90(2002).2.Oset,E.,and Vicente-Vacas,M.J.,Nucl.Phys.,A446,584–612(1985).3.Bernard,V.,Kaiser,N.,and Meißner,U.-G.,Nucl.Phys.,B457,147–174(1995).4.Gómez Tejedor,J.A.,and Oset,E.,Nucl.Phys.,A600,413–435(1996).5.Morsch,H.P.,and Zupranski,P.,Phys.Rev.,C61,024002(2000).6.Krehl,O.,Hanhart,C.,Krewald,S.,and Speth,J.,Phys.Rev.,C62,025207(2000).7.Hernández,E.,Oset,E.,and Vicente Vacas,M.J.,Phys.Rev.,C66,065201(2002).8.Brodowski,W.,et al.,Phys.Rev.Lett.,88,192301(2002).9.Johanson,J.,et al.,Nucl.Phys.,A712,75–94(2002).10.Pätzold,J.,et al.,Phys.Rev.,C67,052202(2003).11.Alvarez-Ruso,L.,Oset,E.,and Hernández,E.,Nucl.Phys.,A633,519–546(1998).12.Alvarez-Ruso,L.(2002),private communications.13.Hagiwara,K.,et al.,Phys.Rev.,D66,010001(2002).14.Bashkanov,M.,et al.,proceedings of this conference(2003).15.Clement,H.,et al.,Annual Report2002,Forschungszentrum Jülich(2003),Jül–4052.16.Moalem,A.,Gedalin,E.,Razdolskaja,L.,and Shorer,Z.,Nucl.Phys.,A600,445–460(1996).17.Bernard,V.,Kaiser,N.,and Meißner,U.-G.,Eur.Phys.J.,A4,259–275(1999).18.Gedalin,E.,Moalem,A.,and Razdolskaja,L.,Nucl.Phys.,A650,471–482(1999).19.Moskal,P.,et al.,e–Print Archive:nucl–ex/0307005(2003),Phys.Rev.C,in print.20.Bergdolt,A.M.,et al.,Phys.Rev.,D48,2969–2973(1993).21.Chiavassa,E.,et al.,Phys.Lett.,B322,270–274(1994).22.Calén,H.,et al.,Phys.Lett.,B366,39–43(1996).23.Calén,H.,et al.,Phys.Rev.Lett.,79,2642–2645(1997).24.Hibou,F.,et al.,Phys.Lett.,B438,41–46(1998).25.Smyrski,J.,Wüstner,P.,et al.,Phys.Lett.,B474,182–187(2000).26.Moskal,P.,et al.,Phys.Rev.Lett.,80,3202–3205(1998).27.Moskal,P.,et al.,Phys.Lett.,B474,416–422(2000).28.Balestra,F.,et al.,Phys.Lett.,B491,29–35(2000).29.Khoukaz,A.,et al.,Annual report2000/01,Institute of Nuclear Physics,University of Münster(2001),URL http://www.uni-muenster.de/Physik/KP/anrep.30.Nakayama,K.,Haidenbauer,J.,Hanhart,C.,and Speth,J.,Phys.Rev.,C68,045201(2003).31.Fix,A.(2003),private communications.32.Fix,A.,and Arenhövel,H.,Nucl.Phys.,A697,277–302(2002).33.Schuberth,U.,Phd thesis,Uppsala University(1995).34.Moskal,P.,et al.,Phys.Lett.,B482,356–362(2000),and references therein.35.Calén,H.,et al.,Phys.Lett.,B458,190–196(1999).36.Abdel-Bary,M.,et al.,Eur.Phys.J.,A16,127–137(2003).37.Peña,M.T.,Garcilazo,H.,and Riska,D.O.,Nucl.Phys.,A683,322–338(2001).38.Batini´c,M.,Švarc,A.,and Lee,T.-S.H.,Phys.Scripta,56,321–324(1997).39.Santra,A.B.,and Jain,B.K.,Nucl.Phys.,A634,309–324(1998).40.Gedalin,E.,Moalem,A.,and Razdolskaja,L.,Nucl.Phys.,A634,368–392(1998).41.Fäldt,G.,and Wilkin,C.,Phys.Scripta,64,427–438(2001).42.Nakayama,K.,Speth,J.,and Lee,T.-S.H.,Phys.Rev.,C65,045210(2002).43.Baru,V.,et al.,Phys.Rev.,C67,024002(2003).44.Zlomanczuk,J.(2003),private communications.45.Winter,P.,et al.,Phys.Lett.,B544,251–258(2002).46.Czy˙z ykiewicz,R.,et al.,proceedings of this conference(2003).47.Balewski,J.T.,et al.,Phys.Lett.,B420,211–216(1998).48.Sewerin,S.,Schepers,G.,et al.,Phys.Rev.Lett.,83,682–685(1999).49.Baldini,A.,Flaminio,V.,Moorhead,W.G.,and Morrison,D.R.O.,Total Cross-Sections forReactions of High–Energy Particles,vol.12B of Landolt–Börnstein:New Series.Group1,edited by H.Schopper,Springer,Heidelberg,Germany,1988,ISBN3–540–18412–0.50.Gasparian,A.,et al.,Phys.Lett.,B480,273–279(2000).51.Siebert,R.,et al.,Nucl.Phys.,A567,819–843(1994).52.Tan,T.H.,Phys.Rev.Lett.,23,395–398(1969).53.Balestra,F.,et al.,Phys.Rev.Lett.,83,1534–1537(1999).54.Maggiora,M.,πN Newslett.,16,273–279(2002).get,J.M.,Phys.Lett.,B259,24–28(1991).56.Ro˙z ek,T.,and Grzonka, D.,COSY Proposal117,IKP,FZ Jülich,Germany(2002),URLhttp://www.fz-juelich.de/ikp/en/publications.shtml.57.Shyam,R.,Penner,G.,and Mosel,U.,Phys.Rev.,C63,022202(2001).58.Kowina,P.,et al.,proceedings of this conference(2003).59.Wagner,M.,et al.,proceedings of this conference(2003).60.Kowina,P.,Wolke,M.,et al.,e–Print Archive:nucl–ex/0302014(2003).61.Gasparyan,A.,Symposium on Threshold Meson Production in pp and pd Interaction,Forschungs-zentrum Jülich,Germany,2002,vol.11of Matter and Material,pp.205–211.62.Maessen,P.M.M.,Rijken,T.A.,and de Swart,J.J.,Phys.Rev.,C40,2226–2245(1989).63.Haidenbauer,J.,Melnitchouk,W.,and Speth,J.,AIP Conf.Proc.,603,421–424(2001).64.Moskal,P.,et al.,proceedings of this conference(2003).。

学术英语(社科)Unit3三单元原文及翻译

学术英语(社科)Unit3三单元原文及翻译

UNIT 3 Unemployment1. Losing a job can be the most distressing economic event in a person’s life. Most people rely on their labor earnings to maintain their standard of living, and many people get from their work not only income but also a sense of personal accomplishment.A job loss means a lower living standard in the present, anxiety about the future, and reducedself-esteem. It is not surprising, therefore, that politicians campaigning for office often speak about how their proposed policies will help create jobs.失去一份工作可能是最痛苦的经济事件在一个人的生活。

大多数人们依靠自己的劳动收入来维持他们的生活标准,许多人会从他们的工作得到的不仅是收入,还有自己的成就感。

一个失去工作意味着现在要定一个更低的生活标准,焦虑未来,并丧失自尊心。

这并不奇怪,因此,政治家竞选办公室经常谈论他们所提出的政策将帮助创造就业机会。

4.In judging how serious the problem of unemployment is, one question to consider is whether unemployment is typically a short-term or long-term condition. If unemployment is short-term, one might conclude that it is not a big problem. Workers may require a few weeks between jobs to find the openings that best suit their tastes and skills. Yet if unemployment is long-term, one might conclude that it is a serious problem. Workers unemployed for many months are more likely to suffer economic and psychological hardship.判断失业问题有多么严重时,其中一个问题就是要考虑是否失业通常是一个短期或长期的条件。

时间依赖问题

时间依赖问题

Ωn (t) ,
(5)
Ω1 (t) =
0 ′ 2.) Ω2 = −1/2[Ω1, A] so that t t1
A(t1 )dt1 ,
(6)
Ω2 (t) = 1/2
0 0
[A(t1 ), A(t2 )]dt2 dt1 ,Leabharlann (7)and so on.
3
The procedure can be written as Magnus expansion generator: ′ Ω2 = −1/2[Ω1, A] so that
Multi-product expansion, Suzuki’s method and the Magnus integrator for solving time-dependent problems
J¨ urgen Geiser and Siu A. Chin
geiser@mathematik.hu-berlin.de
4
The fourth order method is given in [5], see also the coefficients in the appendix. The convergence results are given in [17]. In the next we present the algorithms for a Hamiltonian application. Application to a Hamiltonian The algorithm is given for a Hamiltonian as: H = T (p, t) + V (q, t) For example : H = p2 /2 + V (q, t), where T (p, t) = p2 /2 and
  1. 1、下载文档前请自行甄别文档内容的完整性,平台不提供额外的编辑、内容补充、找答案等附加服务。
  2. 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
  3. 3、如文档侵犯您的权益,请联系客服反馈,我们会尽快为您处理(人工客服工作时间:9:00-18:30)。

a rXiv:solv-int/9922v131Ja n1999On time-dependent symmetries and formal symmetries of evolution equations Artur Sergyeyev Institute of Mathematics of NAS of Ukraine,3Tereshchenkivs’ka Str.,252004Kyiv,Ukraine E-mail:arthurser@imath.kiev.ua,arthur@apmat.freenet.kiev.ua Abstract We present the explicit formulae,describing the structure of symme-tries and formal symmetries of any scalar (1+1)-dimensional evolution ing these results,the formulae for the leading terms of com-mutators of two symmetries and two formal symmetries are found.The generalization of these results to the case of system of evolution equations is also discussed.1Introduction It is well known that provided scalar (1+1)-dimensional evolution equation pos-sesses the infinite-dimensional commutative Lie algebra of time-independent gen-eralized (Lie –B¨a cklund)symmetries,it is either linearizable or integrable via inverse scattering transform (see e.g.[1],[2],[3]for the survey of known results and [4]for the generalization to (2+1)dimensions).The existence of such alge-bra is usually proved by exhibiting the recursion operator [3]or mastersymmetry [5].But in order to possess the latter,the equation in question must have some (possibly nonlocal)time-dependent symmetries.This fact is one of the main reasons of growing interest to the study of whole algebra of time-dependent symmetries of evolution equations [6],[7].However,to the best of author’s knowledge,there exist almost no results,describing the structure of this algebra for the generic scalar evolution equation (even in 1+1dimensions)without any initial conjectures about the specific properties of this equation,like the existence of formal symmetry,Lax pair,mastersymmetry,etc.1Although Vinogradov et al.[10]had outlined the generalscheme of study of local and nonlocal symmetries of evolution equations,it remained unrealized in its general form until now.In this paper we put the part of this scheme into life and go even ly,in Theorems 1and 2we describe the general structure of local sym-metries and formal symmetries of scalar evolution equation,and in Theorems 3and 5we present the formulae for the leading terms of the commutators of two symmetries (of sufficiently high order)and of two formal symmetries.We present also the generalization of these results to the case of systems of evolution equations.2Basic definitions and known factsWe consider the scalar1+1-dimensional evolution equation∂u/∂t=F(x,t,u,u1,...,u n),n≥2,∂F/∂u n=0,(1)where u l=∂l u/∂x l,l=0,1,2,...,u0≡u,and the symmetries of this equation, i.e.the right hand sides G of evolution equations∂u/∂τ=G(x,t,u,u1,...,u k),(2)compatible with equation(1).The greatest number k such that∂G/∂u k=0is called the order of symmetry and is denoted as k=ord G.If G is independent of u,u1,...,we assume that ord G=0.Let S(k)be the space of symmetries of order not higher than k of(1)and S= ∞j=0S(k).S is Lie algebra with respect to the so-called Lie bracket[1],[3]{h,r}=r∗(h)−h∗(r),where for any sufficiently smooth function f of x,t,u,u1,...,u s we have intro-duced the notationf∗=si=0∂h/∂u i D i,∇f=∞ i=0D i(f)∂/∂u i,where D=∂/∂x+∞i=0u i+1∂/∂u i.G is symmetry of Eq.(1)if and only if[3]∂G/∂t=−{F,G}.(3) In many examples Eq.(1)is quasilinear,so letn0= max(1−j,0),if∂F/∂u n−i=φi(x,t),i=0,...,j,2otherwise.(4)It is known[3]that for any G∈S,ord G=k≥n0,we have∂G/∂u k=c k(t)Φk/n,(5) where c k(t)is a function of t andΦ=∂F/∂u n.It is also well known[1]that for any sufficiently smooth functions P,Q of x,t,u,u1,u2,...the relation R={P,Q}impliesR∗=∇P(Q∗)−∇Q(P∗)+[Q∗,P∗],(6)where∇P(Q∗)≡∞i,j=0D j(P)∂2QIn particular,Eq.(3)yields∂G∗/∂t≡(∂G/∂t)∗=∇G(F∗)−∇F(G∗)+[F∗,G∗].(7) Equating the coefficients at D s,s=0,1,2,...on right and left hand sides of Eq.(7),we obtain∂2G∂u m∂u l −k r=0D r(F)∂2G∂u iD i+j−l ∂G∂u jD i+j−l ∂Fp!(q−p)!and we assume that1/p!=0for negative integer p.Let us also remind some facts concerning the formal series in powers of D (see e.g.[2],[4]for more information),i.e.the expressions of the formH=mj=−∞h j(x,t,u,u1,...)D j.(9)The greatest integer m such that h m=0is called the degree of formal series H and is denoted by deg H.For any formal series H of degree m there exists unique(up to the multiplication by m-th root of unity)formal series H1/m of degree1such that(H1/m)m=H.Now we can define the fractional powers of H as H l/m=(H1/m)l for all integer l.The key result here is that[H p/m,H q/m]=0(10) for all integer p and q.The formal symmetry of Eq.(1)of rank l is the formal series R,satisfying the relation[2]deg(∂R/∂t+∇F(R)−[F∗,R])≤deg F∗+deg R−l.(11) The commutator of two formal symmetries of ranks l and m obviously is again a formal symmetry of rank not lower than min(l,m),and thus the set F S r of all formal symmetries of given equation(1)of rank not lower than r is a Lie algebra.Like for the case of symmetries,we shall denote by F S(k)r the set of formal symmetries of degree not higher than k and of rank not lower than r. Note that if G is symmetry of order k,then by virtue of Eq.(7)G∗is the formal symmetry of degree k and rank k+n−deg∇G(F∗).33Explicit form of symmetries and formal sym-metriesIn this section we shall consider a symmetry G of order k≥n0.The successive solving of Eq.(8)for l=k+n−1,...,n0+n−1yields Eq.(5)and∂G/∂u i=c i(t)Φi/n+k p=i+1[p−in2c k(t)D−1(Φ−1−1/n˙Φ)F(k−n+1)/n∗+(1/n)˙c k(t)D−1(Φ−1/n)F(k−n+1)/n∗,(14) where d i(t)are some functions of t(in fact they are linear combinations ofc max(k−n+1,n0)(t),...,c k(t))and N is some formal series,deg N<max(k−n+1,n0).Likewise,for n0≤k≤n+n0−2Eq.(14)remains true,if two last terms on its right hand side are rejected.Dot here and below stands for the partial derivative with respect to t.The analysis of Eq.(11),similar to the above analysis of Eq.(7),yields Theorem2For any formal symmetry R of(1)of degree k and of rank r>nR=˜R+k j=k−n+1d j(t)F j/n∗+kTheorem 3Let P,Q ∈S/S (n +n 0−2),ord P =p ,ord Q =q .By virtue of Eq.(5)∂P/∂u p =c p (t )Φp/n ,∂Q/∂u q =d q (t )Φq/n .Then ord {P,Q }≤p +q −n and{P,Q }=1n u p +q −n q ˙c p (t )d q (t )−pc p (t )˙d q (t ) +˜R,(16)where ord ˜R<p +q −n .Corollary 1For all integer p ≥n +n 0−1the spaces S (p )are invariant under the adjoint action of S (n ),i.e.the Lie bracket of any symmetry from S (p )with any symmetry from S (n )again belongs to S (p ).Theorem 4For all p =0,...,n S (p )are Lie subalgebras in S .Note that for p =0,1the result of Theorem 4is well known,while for p =2,...,n it is essentially new.Theorem 3shows that Lie algebra S has Virasoro type structure (if we forget about low order symmetries and consider just the leading terms of symmetries).In particular,one may easily establish the existence of Virasoro (or hereditary)algebra [5],[6]of time-independent symmetries and mastersymmetries for many integrable equations (1),using Eq.(16).More generally,Theorem 3is very use-ful in the proof of existence of infinite number of symmetries for given evolution equation,starting from few initially found ones and analyzing their commuta-tors,as described in [10].Note that for particular cases of KdV and Burgers equations it was proved in [10].Likewise,for formal symmetries we have proved the following results:Theorem 5Let P ,Q be formal symmetries of Eq.(1),deg P =p ,deg Q =q ,and the ranks of P and Q are greater than n .By virtue of Theorem 2P =c p (t )F p/n ∗+˜P and Q =d q (t )F q/n ∗+˜Q,deg ˜P <p ,deg ˜Q <q .Then deg[P ,Q]≤p +q −n and[P ,Q]=−1n ∗ q ˙c p (t )d q (t )−pc p (t )˙d q (t ) +˜R ,(17)where deg ˜R<p +q −n .Corollary 2For all integer p the spaces F S (p )r are invariant under the adjointaction of F S (n )r ,provided r >n .Corollary 3For all integer p ≤n ,r >n F S (p )r are Lie subalgebras in F S r .55Some possible generalizationsOur results may be easily generalized to the case of systems of evolution equa-tions of the form(1),when u,u1,u2,...,F,G,P,Q become m-component vec-tors,while c p(t),d q(t)become m×m matrices,which should commute with the m×m matrixΦ=∂F/∂u n,etc.(see[2]for more information).Namely,the second part of Theorem1,concerning the symmetries of order k,n0≤k≤n+ n0−2,Theorem3,Theorem4for n0≤1,Corollary1,Corollary3for p<n hold true,provided all the eigenvalues ofΦare distinct.The part of Theorem1,con-cerning the symmetries of order k>n+n0−2,Theorems2,4,5,Corollaries2and 3hold true if,in addition to the above,detΦ=0.For the case,when detΦ=0, we have proved instead of thefirst part of Theorem1the following result: Theorem6If detΦ=0and all the eigenvalues ofΦare distinct,for any symmetry G of system(1)of order k>n+n0−2G∗=N+kj=k−n+2d j(t)F j/n∗,(18)where d j(t)are some matrices,commuting withΦ,and N is some formal series (with matrix coefficients),deg N<k−n+2.In its turn,Theorem2is replaced by the followingTheorem7If detΦ=0and all the eigenvalues ofΦare distinct,for any formal symmetry R of system(1)of degree k≥0and of rank r>nR=˜R+kj=max(k−n+2,0)d j(t)F j/n∗,(19)where d j(t)are some matrices,commuting withΦ,and˜R is some formal series (with matrix coefficients),deg˜R<max(k−n+2,0).As afinal remark,let us note that Theorems1–7and Corollaries1–3may be also extended(under some extra conditions)to the symmetries and formal symmetries,involving nonlocal variables.We shall discuss this in more detail in separate paper.AcknowledgementsI am sincerely grateful to Profs.B.Fuchssteiner and V.V.Sokolov for stimu-lating discussions.References[1]V.V.Sokolov,Russian Math.Surveys43,,no.5,165(1988).6[2]A V Mikhailov,A B Shabat and V V Sokolov in What is integrability?,ed.V E Zakharov(Springer,N.Y.,1991).[3]P Olver,Applications of Lie Groups to Differential Equations(Springer,N.Y.,1986).[4]A.V.Mikhailov,R.I.Yamilov,J.Phys.A31,6707(1998).[5]B.Fuchssteiner,Progr.Theor.Phys.70,1508(1983).[6]B.Fuchssteiner,J.Math.Phys.34,5140(1993).[7]W.X.Ma,P.K.Bullough,P.J.Caudrey and W.I.Fushchych,J.Phys.A30,5141(1997).[8]B.Flach,Lett.Math.Phys.17,321(1989).[9]Magadeev B A1994St.Petersburg Math.J.5,,no.2,345(.)[10]A M Vinogradov,I S Krasil’shchik,V V Lychagin,Introduction toGeometry of Nonlinear Differential Equations(Nauka,Moscow,1986).7。

相关文档
最新文档