decompositions of
杜宾斯基的APOS理论

APOS: A Constructivist Theory of Learningin Undergraduate Mathematics Education ResearchEd Dubinsky, Georgia State University, USAandMichael A. McDonald, Occidental College, USAThe work reported in this paper is based on the principle that research in mathematics education is strengthened in several ways when based on a theoretical perspective. Development of a theory or model in mathematics education should be, in our view, part of an attempt to understand how mathematics can be learned and what an educational program can do to help in this learning. We do not think that a theory of learning is a statement of truth and although it may or may not be an approximation to what is really happening when an individual tries to learn one or another concept in mathematics, this is not our focus. Rather we concentrate on how a theory of learning mathematics can help us understand the learning process by providing explanations of phenomena that we can observe in students who are trying to construct their understandings of mathematical concepts and by suggesting directions for pedagogy that can help in this learning process.Models and theories in mathematics education can•support prediction,•have explanatory power,•be applicable to a broad range of phenomena,•help organize one’s thinking about complex, interrelated phenomena,•serve as a tool for analyzing data, and•provide a language for communication of ideas about learning that go beyond superficial descriptions.We would like to offer these six features, the first three of which are given by Alan Schoenfeld in “Toward a theory of teaching-in-context,” Issues in Education, both as ways in which a theory can contribute to research and as criteria for evaluating a theory.In this paper, we describe one such perspective, APOS Theory, in the context of undergraduate mathematics education. We explain the extent to which it has the above characteristics, discuss the role that this theory plays in a research and curriculum development program and how such a program can contribute to the development of the theory, describe briefly how working with this particular theory has provided a vehicle for building a community of researchers in undergraduate mathematics education, and indicate the use of APOS Theory in specific research studies, both by researchers who are developing it as well as others not connected with its development. We provide, in connection with this paper, an annotated bibliography of research reports which involve this theory.APOS TheoryThe theory we present begins with the hypothesis that mathematical knowledge consists in an individual’s tendency to deal with perceived mathematical problem situations by constructing mental actions, processes, and objects and organizing them in schemas to make sense of the situations and solve the problems. In reference to these mental constructions we call it APOS Theory. The ideas arise from our attempts to extend to the level of collegiate mathematics learning the work of J. Piaget on reflective abstraction in children’s learning. APOS Theory is discussed in detail in Asiala, et. al. (1996). We will argue that this theoretical perspective possesses, at least to some extent, the characteristics listed above and, moreover, has been very useful in attempting to understand students’learning of a broad range of topics in calculus, abstract algebra, statistics, discrete mathematics, and other areas of undergraduate mathematics. Here is a brief summary of the essential components of the theory.An action is a transformation of objects perceived by the individual as essentially external and as requiring, either explicitly or from memory, step-by-step instructions on how to perform the operation. For example, an individual with an action conception of left coset would be restricted to working with a concrete group such as Z20 and he or she could construct subgroups, such asH={0,4,8,12,16} by forming the multiples of 4. Then the individual could write the left coset of 5 as the set 5+H={1,5,9,13,17} consisting of the elements of Z20which have remainders of 1 when divided by 4.When an action is repeated and the individual reflects upon it, he or she can make an internal mental construction called a process which the individual can think of as performing the same kind of action, but no longer with the need of external stimuli. An individual can think of performing a process without actually doing it, and therefore can think about reversing it and composing it with other processes. An individual cannot use the action conception of left coset described above very effectively for groups such as S4, the group of permutations of four objects and the subgroup H corresponding to the 8 rigid motions of a square, and not at all for groups S n for large values of n. In such cases, the individual must think of the left coset of a permutation p as the set of all products ph, where h is an element of H. Thinking about forming this set is a process conception of coset.An object is constructed from a process when the individual becomes aware of the process as a totality and realizes that transformations can act on it. For example, an individual understands cosets as objects when he or she can think about the number of cosets of a particular subgroup, can imagine comparing two cosets for equality or for their cardinalities, or can apply a binary operation to the set of all cosets of a subgroup.Finally, a schema for a certain mathematical concept is an individual’s collection of actions, processes, objects, and other schemas which are linked by some general principles to form a framework in the individual’s mind that may be brought to bear upon a problem situation involving that concept. This framework must be coherent in the sense that it gives, explicitly or implicitly, means of determining which phenomena are in the scope of the schema and which are not. Because this theory considers that all mathematical entities can be represented in terms of actions, processes, objects, and schemas, the idea of schema is very similar to the concept image which Tall and Vinner introduce in“Concept image and concept definition in mathematics with particular reference to limits and continuity,” Educational Studies in Mathematics, 12, 151-169 (1981). Our requirement of coherence, however, distinguishes the two notions.The four components, action, process, object, and schema have been presented here in a hierarchical, ordered list. This is a useful way of talking about these constructions and, in some sense, each conception in the list must be constructed before the next step is possible. In reality, however, when an individual is developing her or his understanding of a concept, the constructions are notactually made in such a linear manner. With an action conception of function, for example, an individual may be limited to thinking about formulas involving letters which can be manipulated or replaced by numbers and with which calculations can be done. We think of this notion as preceding a process conception, in which a function is thought of as an input-output machine. What actually happens, however, is that an individual will begin by being restricted to certain specific kinds of formulas, reflect on calculations and start thinking about a process, go back to an action interpretation, perhaps with more sophisticated formulas, further develop a process conception and so on. In other words, the construction of these various conceptions of a particular mathematical idea is more of a dialectic than a linear sequence.APOS Theory can be used directly in the analysis of data by a researcher. In very fine grained analyses, the researcher can compare the success or failure of students on a mathematical task with the specific mental constructions they may or may not have made. If there appear two students who agree in their performance up to a very specific mathematical point and then one student can take a further step while the other cannot, the researcher tries to explain the difference by pointing to mental constructions of actions, processes, objects and/or schemas that the former student appears to have made but the other has not. The theory then makes testable predictions that if a particular collection of actions, processes, objects and schemas are constructed in a certain manner by a student, then this individual will likely be successful using certain mathematical concepts and in certain problem situations. Detailed descriptions, referred to as genetic decompositions, of schemas in terms of these mental constructions are a way of organizing hypotheses about how learning mathematical concepts can take place. These descriptions also provide a language for talking about such hypotheses.Development of APOS TheoryAPOS Theory arose out of an attempt to understand the mechanism of reflective abstraction, introduced by Piaget to describe the development of logical thinking in children, and extend this idea to more advanced mathematical concepts (Dubinsky, 1991a). This work has been carried on by a small group of researchers called a Research in Undergraduate Mathematics Education Community (RUMEC) who have been collaborating on specific research projects using APOS Theory within abroader research and curriculum development framework. The framework consists of essentially three components: a theoretical analysis of a certain mathematical concept, the development and implementation of instructional treatments (using several non-standard pedagogical strategies such as cooperative learning and constructing mathematical concepts on a computer) based on this theoretical analysis, and the collection and analysis of data to test and refine both the initial theoretical analysis and the instruction. This cycle is repeated as often as necessary to understand the epistemology of the concept and to obtain effective pedagogical strategies for helping students learn it.The theoretical analysis is based initially on the general APOS theory and the researcher’s understanding of the mathematical concept in question. After one or more repetitions of the cycle and revisions, it is also based on the fine-grained analyses described above of data obtained from students who are trying to learn or who have learned the concept. The theoretical analysis proposes, in the form of a genetic decomposition, a set of mental constructions that a student might make in order to understand the mathematical concept being studied. Thus, in the case of the concept of cosets as described above, the analysis proposes that the student should work with very explicit examples to construct an action conception of coset; then he or she can interiorize these actions to form processes in which a (left) coset gH of an element g of a group G is imagined as being formed by the process of iterating through the elements h of H, forming the products gh, and collecting them in a set called gH; and finally, as a result of applying actions and processes to examples of cosets, the student encapsulates the process of coset formation to think of cosets as objects. For a more detailed description of the application of this approach to cosets and related concepts, see Asiala, Dubinsky, et. al. (1997).Pedagogy is then designed to help the students make these mental constructions and relate them to the mathematical concept of coset. In our work, we have used cooperative learning and implementing mathematical concepts on the computer in a programming language which supports many mathematical constructs in a syntax very similar to standard mathematical notation. Thus students, working in groups, will express simple examples of cosets on the computer as follows.Z20 := {0..19};op := |(x,y) -> x+y (mod 20)|;H := {0,4,8,12,16};5H := {1,5,9,13,17};To interiorize the actions represented by this computer code, the students will construct more complicated examples of cosets, such as those appearing in groups of symmetries.Sn := {[a,b,c,d] : a,b,c,d in {1,2,3,4} | #{a,b,c,d} = 4};op := |(p,q) -> [p(q(i)) : i in [1..4]]|;H := {[1,2,3,4], [2,1,3,4], [3,4,1,2], [4,3,2,1]};p := [4,3,2,1];pH := {p .op q : q in H};The last step, to encapsulate this process conception of cosets to think of them as objects, can be very difficult for many students. Computer activities to help them may include forming the set of all cosets of a subgroup, counting them, and picking two cosets to compare their cardinalities and find their intersections. These actions are done with code such as the following.SnModH := {{p .op q : q in H} : p in Sn};#SnModH;L := arb(SnModH); K := arb(SnModH); #L = #K; L inter K;Finally, the students write a computer program that converts the binary operation op from an operation on elements of the group to subsets of the group. This structure allows them to construct a binary operation (coset product) on the set of all cosets of a subgroup and begin to investigate quotient groups.It is important to note that in this pedagogical approach, almost all of the programs are written by the students. One hypothesis that the research investigates is that, whether completely successful or not, the task of writing appropriate code leads students to make the mental constructions of actions, processes, objects, and schemas proposed by the theory. The computer work is accompanied by classroom discussions that give the students an opportunity to reflect on what they have done in the computer lab and relate them to mathematical concepts and their properties and relationships. Once the concepts are in place in their minds, the students are assigned (in class, homework and examinations) many standard exercises and problems related to cosets.After the students have been through such an instructional treatment, quantitative and qualitative instruments are designed to determine the mental concepts they may have constructed and the mathematics they may have learned. The theoretical analysis points to questions researchers may ask in the process of data analysis and the results of this data analysis indicates both the extent to which the instruction has been effective and possible revisions in the genetic decomposition.This way of doing research and curriculum development simultaneously emphasizes both theory and applications to teaching practice.Refining the theoryAs noted above, the theory helps us analyze data and our attempt to use the theory to explain the data can lead to changes in the theory. These changes can be of two kinds. Usually, the genetic decomposition in the original theoretical analysis is revised and refined as a result of the data. In rare cases, it may be necessary to enhance the overall theory. An important example of such a revision is the incorporation of the triad concept of Piaget and Garcia (1989) which is leading to a better understanding of the construction of schemas. This enhancement to the theory was introduced in Clark, et. al. (1997) where they report on students’ understanding of the chain rule, and is being further elaborated upon in three current studies: sequences of numbers (Mathews, et. al., in preparation); the chain rule and its relation to composition of functions (Cottrill, 1999); and the relations between the graph of a function and properties of its first and second derivatives (Baker, et. al., submitted). In each of these studies, the understanding of schemas as described above was not adequate to provide a satisfactory explanation of the data and the introduction of the triad helped to elaborate a deeper understanding of schemas and provide better explanations of the data.The triad mechanism consists in three stages, referred to as Intra, Inter, and Trans, in the development of the connections an individual can make between particular constructs within the schema, as well as the coherence of these connections. The Intra stage of schema development is characterized by a focus on individual actions, processes, and objects in isolation from other cognitive items of a similar nature. For example, in the function concept, an individual at the Intra level, would tend to focus on a single function and the various activities that he or she could perform with it. TheInter stage is characterized by the construction of relationships and transformations among these cognitive entities. At this stage, an individual may begin to group items together and even call them by the same name. In the case of functions, the individual might think about adding functions, composing them, etc. and even begin to think of all of these individual operations as instances of the same sort of activity: transformation of functions. Finally, at the Trans stage the individual constructs an implicit or explicit underlying structure through which the relationships developed in the Inter stage are understood and which gives the schema a coherence by which the individual can decide what is in the scope of the schema and what is not. For example, an individual at the Trans stage for the function concept could construct various systems of transformations of functions such as rings of functions, infinite dimensional vector spaces of functions, together with the operations included in such mathematical structures.Applying the APOS TheoryIncluded with this paper is an annotated bibliography of research related to APOS Theory, its ongoing development and its use in specific research studies. This research concerns mathematical concepts such as: functions; various topics in abstract algebra including binary operations, groups, subgroups, cosets, normality and quotient groups; topics in discrete mathematics such as mathematical induction, permutations, symmetries, existential and universal quantifiers; topics in calculus including limits, the chain rule, graphical understanding of the derivative and infinite sequences of numbers; topics in statistics such as mean, standard deviation and the central limit theorem; elementary number theory topics such as place value in base n numbers, divisibility, multiples and conversion of numbers from one base to another; and fractions. In most of this work, the context for the studies are collegiate level mathematics topics and undergraduate students. In the case of the number theory studies, the researchers examine the understanding of pre-college mathematics concepts by college students preparing to be teachers. Finally, some studies such as that of fractions, show that the APOS Theory, developed for “advanced” mathematical thinking, is also a useful tool in studying students’understanding of more basic mathematical concepts.The totality of this body of work, much of it done by RUMEC members involved in developing the theory, but an increasing amount done by individual researchers having no connection with RUMEC or the construction of the theory, suggests that APOS Theory is a tool that can be used objectively to explain student difficulties with a broad range of mathematical concepts and to suggest ways that students can learn these concepts. APOS Theory can point us towards pedagogical strategies that lead to marked improvement in student learning of complex or abstract mathematical concepts and students’ use of these concepts to prove theorems, provide examples, and solve problems. Data supporting this assertion can be found in the papers listed in the bibliography.Using the APOS Theory to develop a community of researchersAt this stage in the development of research in undergraduate mathematics education, there is neither a sufficiently large number of researchers nor enough graduate school programs to train new researchers. Other approaches, such as experienced and novice researchers working together in teams on specific research problems, need to be employed at least on a temporary basis. RUMEC is one example of a research community that has utilized this approach in training new researchers.In addition, a specific theory can be used to unify and focus the work of such groups. The initial group of researchers in RUMEC, about 30 total, made a decision to focus their research work around the APOS Theory. This was not for the purpose of establishing dogma or creating a closed research community, but rather it was a decision based on current interests and needs of the group of researchers.RUMEC was formed by a combination of established and beginning researchers in mathematics education. Thus one important role of RUMEC was the mentoring of these new researchers. Having a single theoretical perspective in which the work of RUMEC was initially grounded was beneficial for those just beginning in this area. At the meetings of RUMEC, discussions could focus not only on the details of the individual projects as they developed, but also on the general theory underlying all of the work. In addition, the group’s general interest in this theory and frequent discussions about it in the context of active research projects has led to growth in the theory itself. This was the case, for example, in the development of the triad as a tool for understanding schemas.As the work of this group matures, individuals are beginning to use other theoretical perspectives and other modes of doing research.SummaryIn this paper, we have mentioned six ways in which a theory can contribute to research and we suggest that this list can be used as criteria for evaluating a theory. We have described how one such perspective, APOS Theory is being used, in an organized way, by members of RUMEC and others to conduct research and develop curriculum. We have shown how observing students’ success in making or not making mental constructions proposed by the theory and using such observations to analyze data can organize our thinking about learning mathematical concepts, provide explanations of student difficulties and predict success or failure in understanding a mathematical concept. There is a wide range of mathematical concepts to which APOS Theory can and has been applied and this theory is used as a language for communication of ideas about learning. We have also seen how the theory is grounded in data, and has been used as a vehicle for building a community of researchers. Yet its use is not restricted to members of that community. Finally, we provide an annotated bibliography which presents further details about this theory and its use in research in undergraduate mathematics education.An Annotated Bibliography of workswhich develop or utilize APOS TheoryI. Arnon. Teaching fractions in elementary school using the software “Fractions as Equivalence Classes” of the Centre for Educational Technology, The Ninth Annual Conference for Computers in Education, The Israeli Organization for Computers in Education, Book of Abstracts, Tel-Aviv, Israel, p. 48, 1992. (In Hebrew).I. Arnon, R. Nirenburg and M. Sukenik. Teaching decimal numbers using concrete objects, The Second Conference of the Association for the Advancement of the Mathematical Education in Israel, Book of Abstracts, Jerusalem, Israel, p. 19, 1995. (In Hebrew).I. Arnon. Refining the use of concrete objects for teaching mathematics to children at the age of concrete operations, The Third Conference of the Association for the Advancement of the Mathematical Education in Israel, Book of Abstracts, Jerusalem, Israel, p. 69, 1996. (In Hebrew).I. Arnon. In the mind’s eye: How children develop mathematical concepts – extending Piaget's theory. Doctoral dissertation, School of Education, Haifa University, 1998a.I. Arnon. Similar stages in the developments of the concept of rational number and the concept of decimal number, and possible relations between their developments, The Fifth Conference of the Association for the Advancement of the Mathematical Education in Israel, Book of Abstracts. Be’er-Tuvia, Israel, p. 42, 1998b. (In Hebrew).The studies by Arnon and her colleagues listed above deal with the development ofmathematical concepts by elementary school children. Having created a framework thatcombines APOS theory, Nesher’s theory on Learning Systems, and Yerushalmy’s ideas ofmulti-representation, she investigates the introduction of mathematical concepts as concreteactions versus their introduction as concrete objects. She establishes developmental paths for certain fraction-concepts. She finds that students to whom the fractions were introduced asconcrete actions progressed better along these paths than students to whom the fractions were introduced as concrete objects. In addition, the findings establish the following stage in thedevelopment of concrete actions into abstract objects: after abandoning the concrete materials, and before achieving abstract levels, children perform the concrete actions in their imagination.This corresponds to the interiorization of APOS theory.M. Artigue, Enseñanza y aprendizaje del análisis elemental: ¿qué se puede aprender de las investigaciones didácticas y los cambios curriculares? Revista Latinoamericana de Investigación en Matiemática Educativa, 1, 1, 40-55, 1998.In the first part of this paper, the author discusses a number of student difficulties and tries toexplain them using various theories of learning including APOS Theory. Students’unwillingness to accept that 0.999… is equal to 1 is explained, for example, by interpreting the former as a process, the latter as an object so that the two cannot be seen as equal until thestudent is able to encapsulate the process which is a general difficulty. In the second part of the paper, the author discusses the measures that have been taken in France during the 20thCentury to overcome these difficulties.M. Asiala, A. Brown, D. DeVries, E. Dubinsky, D. Mathews and K. Thomas. A framework for research and curriculum development in undergraduate mathematics education, Research in Collegiate Mathematics Education II, CBMS Issues in Mathematics Education, 6, 1-32, 1996.The authors detail a research framework with three components and give examples of itsapplication. The framework utilizes qualitative methods for research and is based on a veryspecific theoretical perspective that was developed through attempts to understand the ideas of Piaget concerning reflective abstraction and reconstruct them in the context of college levelmathematics. For the first component, the theoretical analysis, the authors present the APOStheory. For the second component, the authors describe specific instructional treatments,including the ACE teaching cycle (activities, class discussion, and exercises), cooperativelearning, and the use of the programming language ISETL. The final component consists ofdata collection and analysis.M. Asiala, A. Brown, J. Kleiman and D. Mathews. The development of students’ understanding of permutations and symmetries, International Journal of Computers for Mathematical Learning, 3, 13-43, 1998.The authors examine how abstract algebra students might come to understand permutations of a finite set and symmetries of a regular polygon. They give initial theoretical analyses of what it could mean to understand permutations and symmetries, expressed in terms of APOS. Theydescribe an instructional approach designed to help foster the formation of mental constructions postulated by the theoretical analysis, and discuss the results of interviews and performance on examinations. These results suggest that the pedagogical approach was reasonably effective in helping students develop strong conceptions of permutations and symmetries. Based on thedata collected as part of this study, the authors propose revised epistemological analyses ofpermutations and symmetries and give pedagogical suggestions.M. Asiala, J. Cottrill, E. Dubinsky and K. Schwingendorf. The development of student’s graphical understanding of the derivative, Journal of Mathematical Behavior, 16(4), 399-431, 1997.In this study the authors explore calculus students’ graphical understanding of a function and its derivative. An initial theoretical analysis of the cognitive constructions that might be necessary for this understanding is given in terms of APOS. An instructional treatment designed to help foster the formation of these mental constructions is described, and results of interviews,conducted after the implementation of the instructional treatment, are discussed. Based on the data collected as part of this study, a revised epistemological analysis for the graphicalunderstanding of the derivative is proposed. Comparative data also suggest that students who had the instructional treatment based on the theoretical analysis may have more success indeveloping a graphical understanding of a function and its derivative than students fromtraditional courses.M. Asiala, E. Dubinsky, D. Mathews, S. Morics and A. Oktac. Student understanding of cosets, normality and quotient groups, Journal of Mathematical Behavior,16(3), 241-309, 1997.Using an initial epistemological analysis from Dubinsky, Dautermann, Leron and Zazkis(1994), the authors determine the extent to which the APOS perspective explains students’mental constructions of the concepts of cosets, normality and quotient groups, evaluate the。
times

0
0
0
fn(t1; : : : ; tn)d(Nt1 ? t1)
d(Ntn ? tn )
hIn(fn); Im(gm)iL2 ( ) = n!1fn=mg hfn; gmiL2(R+;dt) n ; fn 2 L2 (R + ; dt) n; gm 2 L2 (R &metric we let In(fn) = In(f~n), where f~n denotes the symmetrization of fn in n variables, hence (1) can be written as
for l = 0.
Z
with Sd ( ) = Sd (
We recall that if f 2 L2 ( d ; e?td dt1
Abstract
Key words: Poisson process, stochastic analysis, chaotic decompositions.
Mathematics Subject Classi cation. 60J75, 60H05.
1 Introduction
Let (Nt )t2R+ be a standard Poisson process with jump times (Tk )k 1, and T0 = 0. The underlying probability space is denoted by ( ; F ; P ), so that L2 ( ; F ; P ) is the space of square-integrable functionals of (Nt )t2R+ . Any F 2 L2( ; F ; P ) can be expanded into the series 1 X 1 (1) F = E F ] + n! In(fn) n=1 where In(fn) is the iterated stochastic integral
碳酸亚乙烯酯添加剂对锂离子电池性能的影响

(EC)、碳酸丙烯酯(PC)、二乙基碳酸酯(DEC)、二甲基碳酸酯(DMC)。
溶剂均采用精馏结合分子筛吸附的方法提纯至纯度(质量分数)>-99.95%(气相色谱仪为GC-14C,日本岛津)。
电解液的配制及电池的装配均在充满高纯氩气的手套箱中进行【畎H20)<I10-6】,LiPF6浓度为1moFL。
电解液中水和酸(I-IF)含量均低于20x10-6,分别用KarlFisher水分测定仪KF831和KarlFisher电位滴定仪798GPTTitrino(瑞士万通)测定。
1.2电池性能测试电池电化学性能测试用CR2032扣式半电池,以MCMB电极片为正极,金属锂片作为负极,电解液为不含与含(质量分数)2%VC的lmoi/LLiPFc,/(EC:PC:DEC:DMC)=(3:1:4:21。
组装成半电池后,在室温下用Land多通道充放电测试系统在0.01~2.0V电压区间进行恒电流充放电实验。
金属锂片作为对电极和参比电极,MCMB电极为工作电极,在多通道测试系统(SolartronAnalytical1480Multistat,英国)上对电池进行循环伏安测试,扫描速率为0.2mV/s。
1.3表面性质分析用JSM.6510(JEOL)扫描电镜观察石墨电极表面的形貌,加速电压为15kV。
用ATR-IRNicoletiSl0(ThermoFish.er)傅里叶变换红外光谱研究电极表面的SEI膜成分,扫描范围680~2000cm—i。
2结果与讨论2.1石墨电极的循环伏安行为图l为石墨电极在不含VC(a)与含VC添加剂(b)的电解液体系中的前三次循环伏安图。
从图l(a)可以看出,石墨电极在不含VC电解液的循环伏安图谱中具有以下特征:(1)在第一次负向扫描过程中,在电极电位为1.5V左右出现一个还原峰,对应于电极表面吸附的溶剂分子的还原,在1.13V左右出现一个还原峰,对应于电解液本体中溶剂组分的还原分解,并形成同体电解质相界面膜(SEI);随着电位的不断降低,电解液的阴极电流逐渐增大,对应于锂离子向石墨电极嵌入,生成锂碳嵌合物的量不断增加;(2)正向扫描时,在0.45V左右出现一个氧化峰,对应于锂碳嵌合物发生阳极氧化,锂离子从石墨电极中脱出。
基于VMD最优分解次数的研究

基于VMD最优分解次数的研究发表时间:2019-08-15T15:58:02.740Z 来源:《科技新时代》2019年6期作者:夏雄飞[导读] 本文通过对VMD分解信号后的若干个IMF模态分量进行希尔伯特变换求得该信号的平均瞬时频率分量,并根据平均瞬时频率分量确定最佳的信号分解次数。
民航湖南空管分局,421000,摘要:非平稳、非线性信号的常用处理方法有小波变换、短时傅里叶变换、魏格纳分布等,然而这些方法都有一定的限制。
最近EMD 分解算法和VMD算法对于非平稳、非线性信号具有非常好分解效果。
本文通过对比VMD分解算法和EMD分解算法,发现VMD分解算法能够很好地抑制杂波信号干扰并能够在一定程度上抑制模态混叠现象,能够很好地避免EMD分解算法带来的缺陷。
其次VMD分解算法更容易表达信号波动的大趋势,而EMD分解则不容易观察到该特性。
最后针对VMD算法的分解次数是有人为设定的,因此如何选择最佳分解层数,对于信号分解是否彻底也具有至关重要的作用,本文通过计算VMD算法分解后的IMF模态分量的平均瞬时频率,确定VMD分解算法的最佳分解次数。
关键字:非线性信号;非平稳信号;EMD分解;VMD分解;Abstract: Commonly used methods for non-stationary and nonlinear signals are wavelet transform, short-time Fourier transform, and Wigner distribution. However, these methods have certain limitations. Recently, the EMD decomposition algorithm and the VMD algorithm have very good decomposition effects for non-stationary and nonlinear signals. By comparing the VMD decomposition algorithm and the EMD decomposition algorithm, it is found that the VMD decomposition algorithm can suppress the clutter signal interference and suppress the modal aliasing phenomenon to a certain extent, which can avoid the defects caused by the EMD decomposition algorithm. Secondly, the VMD decomposition algorithm is more likely to express the general trend of signal fluctuation, and EMD decomposition is not easy to observe this characteristic. Finally, the number of decompositions for the VMD algorithm is artificially set. Therefore, how to choose the optimal decomposition layer is also crucial for the complete decomposition of the signal. This paper calculates the average of the IMF modal components after the decomposition of the VMD algorithm. The instantaneous frequency determines the optimal number of decompositions of the VMD decomposition algorithm.Keywords: nonlinear signal; non-stationary signal; EMD decomposition; VMD algorithm;1. 引言传统时频分析方法由于是将信号以规定的频率成分做线性变化处理取得了比较好的效果。
量子信息

2.1 Show that (1,-1),(1,2) and (2,1) are linearly dependent.Solution: we have (1,-1)+(1,2) -(2,1)=0, so (1,-1),(1,2) and (2,1) are linearly dependent.2.2 Suppose V is a vector space with basis vectors 01and ,and A is a linear operator from V to V such that 0110and A A ==. Give a matrix representation for A, with respect to the input basis 0,1,and the output basis 0,1.Find input and output basis which give rise to a different matrix representation of A.Solution: 0001A =+,1001A =+,so 0110A ⎛⎫⎪⎝⎭=.BecauseV is a vector space with basis vectors 01and ,we can get another basis of V arevectors))0101and +-,)))010101A A +=+=+,))))01011001A A -=-=-=,so now 1001A ⎛⎫ ⎪-⎝⎭=2.3 Suppose A is a linear operator from vector space V to vector space W, and B is a linear operator from vector space W to vector space X. Let ,j k w and x v i be bases for the vector space V ,W, and X, respectively. Show that the matrix representation for linear transformation BA is the matrix product of the matrix representatitions for B and A, with respect to the appeopriate bases.Solution: ,i jij j kj k jkA v AB B x ωω==∑∑i ji j ji j ji kj k ki k jjjkkBA v B A A B A B x BA x ωω====∑∑∑∑∑So the matrix representation for linear transformation BA is the matrix product of the matrix representatitions for B and A.2.4 Show that the identity operator on a vector space V has a matrix representation which is one along the diagonal and zero everywhere else, if the matrix representation is taken with respect to the same input and output bases. This matrix is known as the identity matrix.Solution: let v i be the basis for the vector space V, if the matrix representation is taken with respecttothesameinputandoutputbases,wehave121100...00...0i i nA v v v v v v v v i i i -+=+++++++=,so 100 (00)10 (00)01...0:::0000...1A ⎛⎫⎪ ⎪ ⎪ ⎪ ⎪ ⎪⎝⎭=2.5 Verify that (. , .) just defined is an inner product on nC Solution:let12,,,......i nv ωωωω are vectors innC, we have(1)(),,i i iiiv i v λωλω⎛⎫=∑ ⎪⎝⎭∑, (2)()()****,,,*i i i i i i ii i v v vv v ωωωωω⎛⎫=== ⎪⎝⎭∑∑∑so()()*,,v vωω=. (3) ()*,0,i iiv v v v=≥∑ only if 0v =, (),0v v =So (. , .) just defined is an inner product on n C2.6 Show that any inner product (. , .) is conjugate-linear in the first argument,()*,,i i i i i iv v λωλω⎛⎫= ⎪⎝⎭∑∑ Solution:()()()*****,,,,,i i i i i iiii ii i iiiv v v v vλωλωλωλωλω⎛⎫⎛⎫==== ⎪⎪⎝⎭⎝⎭∑∑∑∑∑2.7 Verify that ()1,1ω≡ and ()1,1v ≡- are orthogonal. What are the normalized forms ofthese vectors.Solution: because ((1,1),(1,-1))=1-1=0,()1,1ω≡ and ()1,1v ≡- are orthogonal.v ω==))1,11,1- 2.8 Prove that the Gram-Schmidt procedure produces an orthonormal basis for V .Solution:()121211212112111212121211212112121,,v v v v v v v v v v v v v v v v ωωωωωωωωωωωωωωωωω---====---121111212110v v v ωωωωωωω-==-, let2.9 The Pauli matrices can be considered as operators with respect to an orthonormal basis 0 and 1 for a two-dimensional Hilbert space.Express each of the Pauli operators in the outer product notation. Solution:10010010000,1,01,10,0,110100100001⎛⎫⎛⎫⎛⎫⎛⎫⎛⎫⎛⎫====== ⎪ ⎪ ⎪ ⎪ ⎪ ⎪⎝⎭⎝⎭⎝⎭⎝⎭⎝⎭⎝⎭So00011,0110,0110,0011x y z i i σσσσ=+=+=-+=-2.10 Suppose i v is an orthonormal basis for an inner product space V . What is the matrix representation for the operator j k v v , with respect to the i v basis.Solution:,,,j i i k i i j k i iii i g i g iiiii gv m v v n v v v m v nm n v v ====∑∑∑∑∑2.11 Find the eigenvectors, eigenvalues, and diagonal representations of the Pauli matrices X,Y ,Z.Solution: X: eigenvalues:1 and -1, eigenvectors : and ⎛⎪ ⎪⎝⎭X ⎛ ⎛ =- ⎪⎝ ⎪⎝⎭Y: eigenvalues:1 and -1, eigenvectors: andY =- Z: eigenvalues:1 and -1, eigenvectors : 01and 0011Z ==-2.12 Prove that the matrix 1011⎛⎫⎪⎝⎭is not diagonalizable. Solution: for matrix 1011⎛⎫⎪⎝⎭, there is only one eigenvalue 1λ=, the eigenvectors is 01⎛⎫⎪⎝⎭,because ()01001111⎛⎫⎛⎫≠ ⎪ ⎪⎝⎭⎝⎭, the matrix 1011⎛⎫ ⎪⎝⎭is not diagonalizable2.13 Ifω and v are two vectors, show that ()†v v ω=Solution:()(),,vv v v v ωωωωωω==, according to (2.32), we have()†v v ωω=.2.14 Show that the adjoint operation is anti-linear, †*i i i i i i a A a A +⎛⎫= ⎪⎝⎭∑∑Solution:()11221122,......i i n n n n i v a A v a A a A a A a v A a A a v A ωωωωω⎛⎫⎛⎫=+++=+++ ⎪ ⎪⎝⎭⎝⎭∑ ()()()*1122,,...,,n n i i i a A v a A v a A v a A v ωωωω++++⎛⎫=+++= ⎪⎝⎭∑So, according to (2.32), †*i i i i i i a A a A +⎛⎫= ⎪⎝⎭∑∑2.15 Show that ()A A ++=Solution: ()()(),,v Av A A v A v ωωωω+++===, so ()AA ++=2.16 Show that any projector P satisfies the equation 2P P = Solution: 11122...ni P in n ===+++∑,()()21122 (1)122...P n n n n=++++++11112222...n n n n =++1122....n n P =++=2.17 Show that a normal matrix is Hermitian if and only if it has real eigenvalues. Solution:2.18 Show that all eigenvalues of a unitary matrix have modulus I , that is, can be written in theform i e θfor some real θ.Solution: let the eigenvector of a unitary matrix is α, so U αλα=, U αλαλα==,Because U αα=, we have 1λ=2.19 Show that the Pauli matrices are Hermitian and unitary. Solution: X **010*********X X XX ⎛⎫⎛⎫⎛⎫===⎪⎪ ⎪⎝⎭⎝⎭⎝⎭, *X X = Y **00100001i i Y Y YY i i --⎛⎫⎛⎫⎛⎫=== ⎪⎪ ⎪⎝⎭⎝⎭⎝⎭, *Y Y =Z ***101010,010101Z Z ZZ Z Z ⎛⎫⎛⎫⎛⎫====⎪⎪ ⎪--⎝⎭⎝⎭⎝⎭So the Pauli matrices are Hermitian and unitary.2.20 Suppose 'A and ''A are matrix representation of an operator A on a vector space V withrespect to different orthonormal bases, i j v and ω. Then the element of 'A and ''A are'''ij i j ij i j A v A v andA A ω==. Characterize the relationship between 'A and ''ASolution:2.21 Repeat the proof of the spectral decomposition in Box 2.2 for the case when M is Hermitian, simplifying the proof wherever possible. Solution:2.22 Prove that two eigenvectors of a Hermitian operator with different eigenvalues are necessarily orthogonal.Solution: 12,letU U αλαβλβ==, we have*2U βλβ+++=, *21U U βαλβλα+++=()*2110λλβα+-=,***12112212211,1,1,1,,1because λλλλλλλλλλ====≠≠()*2110λλ-≠, 0βα+= ,so two eigenvectors of a Hermitian operator with differenteigenvalues are necessarily orthogonal.2.23 Show that the eigenvalues of a projector P are all either 0 or 1.Solution: 22P PP P P ααλαλαλα====, 2P P =, we have 2P P ααλα== So 2λαλα=, ()20λλα-=, the eigenvalues of a projector P are all either 0 or 1.2.24 Show that a positive operator is necessarily Hermitian.(Hint: Show that an arbitrary operator A can be written A=B+Ic where B and C are Hermitian.)Solution: let operator A is positive, so A is 合同 with I. **A C IC C C == where C is ainvertible matrix. **A C C A ==, so a positive operator is necessarily Hermitian2.25 Show that for any operator A, A A +is positive.Solution: ()0AX A AX AX AX +++=≥, so A A +is positive.2.26 Let (=0+1ψ23and ψψ⊗⊗ explicitly, both in terms of tensorproducts like 01, and using the Kronecker product. Solution: ((()20+10+1000110112ψ⊗==+++()(3000110110+1ψ⊗=+++⊗=2.27 Calculate the matrix representation of the tensor products of the Pauli operators (a) X and Z; (b) I and X; (c) X and I. Is the tensor product commutative?Solution: (a) 001001100001100110000100X Z ⎡⎤⎢⎥-⎡⎤⎡⎤⎢⎥⊗=⊗=⎢⎥⎢⎥⎢⎥-⎣⎦⎣⎦⎢⎥-⎣⎦(b) 010010011000011000010010I X ⎡⎤⎢⎥⎡⎤⎡⎤⎢⎥⊗=⊗=⎢⎥⎢⎥⎢⎥⎣⎦⎣⎦⎢⎥⎣⎦(c) 00100110000110011000010X I ⎡⎤⎢⎥⎡⎤⎡⎤⎢⎥⊗=⊗=⎢⎥⎢⎥⎢⎥⎣⎦⎣⎦⎢⎥⎣⎦the tensor product is not commutative. 2.28 Show that the transpose , complex conjugation, and adjoint operations distribute over thetensor product, ()()()***;;TT T A B A B A B A B A B A B +++⊗=⊗⊗=⊗⊗=⊗Solution:()*******1112111121*******21222**21222******1212............=::::::::......n n n n m m mn m m mn A B A B A B A B A B A B A B A B A B A B A B A B A B A B A BA B A B A B A B A B ⎡⎤⎡⎤⎢⎥⎢⎥⎢⎥⎢⎥⊗==⊗⎢⎥⎢⎥⎢⎥⎢⎥⎢⎥⎣⎦⎣⎦()T T 1112111211TT T 2122212222TT T 1212............::::::::......TTn m Tn T T m m m mn n n mn A B A B A B A B A B A B A B A B A B A B A B A B A B A B A BA B A B A B A B A B ⎡⎤⎡⎤⎢⎥⎢⎥⎢⎥⎢⎥⊗===⊗⎢⎥⎢⎥⎢⎥⎢⎥⎢⎥⎣⎦⎣⎦So ()A B A B +++⊗=⊗2.29 Show that the tensor product of two unitary operators is unitary. Solution :let U,V are two unitary operators.()()()()m n mn U V U V U V U V U U V V I I I +++++⊗⊗=⊗⊗=⊗=⊗=So the tensor product of two unitary operators is unitary.2.30 Show that the tensor product of two Hermitian operators is Hermitian. Solution: let A,B are two Hermitian operators.()A B A B A B +++⊗=⊗=⊗ so the tensor product of two Hermitian operators is Hermitian.2.31 Show that the tensor product of two positive operators is positive. Solution: let A,B are two positive operators.()0T T T X A B X X AX X BX ⊗=⊗≥,so he tensor product of two positive operators ispositive.2.32 Show that the tensor product of two projectors is a projector. Solution: let A,B are two projectors()()22A B A B A B A B ⊗⊗=⊗=⊗,so the tensor product of two projectors is a projector2.33TheHadamardoperatorononequbitmaybewrittenas()()010011H ⎤=++-⎦. Show explicitly that the Hadamard transform on nqubits, nH ⊗, may be writtenas (),1x ynx yHx y ∙⊗=-. Write out an explicit matrixrepresentation for 2H ⊗.Solution: ()()(),0100111x yx yH x y ∙⎡⎤=++-=-⎣⎦Let(),1x yk x yH x y∙⊗=-,()()()1,1010011x yk k x yH H H x y ∙⊗+⊗⎡⎤=⊗=-++-⎣⎦2.34 Find the square root and logarithm of the matrix 4334⎡⎤⎢⎥⎣⎦Solution: the eigenvalues of matrix 4334⎡⎤⎢⎥⎣⎦are 121,7λλ== For11λ=, the eigenvector is 11-⎡⎤⎢⎥⎣⎦, for 27λ=, the eigenvector is 11⎡⎤⎢⎥⎣⎦So 1111P -⎡⎤=⎢⎥⎣⎦, 11221122P -⎡⎤-⎢⎥=⎢⎥⎢⎥⎢⎥⎣⎦114311101111112273411071111222222⎡⎤-⎢⎥--⎡⎤⎡⎤⎡⎤⎡⎤⎡⎤⎡⎤⎡⎤==-+⎢⎥⎢⎥⎢⎥⎢⎥⎢⎥⎢⎥⎢⎥⎢⎥⎣⎦⎣⎦⎣⎦⎣⎦⎣⎦⎣⎦⎣⎦⎢⎥⎢⎥⎣⎦2.35 let v be any real , three-dimensional unit vector and θ a real number. Prove that()()()exp cos sin i v I i v θθσθθσ∙=+∙ where 31i i i v v σσ=∙=∑Solution:2.36 Show that the Pauli matrices except for I have trace zero. Solution: tr[X]=0+0=0, tr[Y]=0+0=0, tr[Z]=1-1=0.2.37 If A and B are two linear operators show that tr(AB)=tr(BA). Solution: 112211111()...m mm m nii ii ni in ji ij i i i i j Tr AB A B AB A B A B ======+++=∑∑∑∑∑112211111()...n n n m ni i i i mi im ji ij i i i i j Tr BA B A B A B A A B ======+++=∑∑∑∑∑, so ()()Tr AB Tr BA =2.38 If A and B are two linear operators, show that ()()()Tr A B Tr A Tr B +=+ And if z is a arbitrary complex number show that ()()Tr zA zTr A = Solution: ()iii Tr A A≡∑ ()iii Tr B B≡∑()()()()ii ii ii ii iiiTr A Tr B A B A B Tr A B +=+=+=+∑∑∑()()ii ii iiTr zA zA z A zTr A ≡==∑∑2.39 The set V L of linear operators on Hilbert space V is obviously a vector space —the sum of two linear operators is a linear operator , zA is a linear operator is A is a linear operator and z is a complex number, and there is a zero element 0 . An important additional result is that the vector space V L can be given a natural inner product structure ,turning it into a Hilbert space. (1) Show that the function (),., on V V L L ⨯ defined by(),()A B tr A B +≡ is an innerproduct function. This inner product is known as the Hilbert —Schmidt or trace inner product. (2) If V has d dimensions show that V L has dimension 2d(3) Find an orthonormal basis of Hermitian metrices for the Hilbert space V L . Solution: (1) (),()(),iiiiiiiiA B tr A B tr A B A B λλλλ++⎛⎫≡== ⎪⎝⎭∑∑∑∑ (),()A B tr A B +≡, ()(),()(),B A tr B A tr A B A B ++++===()***1122111,()...0n n ni i i i im im i i i A A tr AA A A A A A A +=====+++≥∑∑∑If and only if 0A =, (),A A =0, so (),()A B tr A B +≡ is an inner product function.(2) (3)2.40 Verify the commutation relations [][][],2;,2;,2X Y iZ Y Z iX Z X iY ===Solution: []01000110,2210001001i i X Y XY YX i iZ i i --⎡⎤⎡⎤⎡⎤⎡⎤⎡⎤=-=-==⎢⎥⎢⎥⎢⎥⎢⎥⎢⎥-⎣⎦⎣⎦⎣⎦⎣⎦⎣⎦[]01010001,2200101010i i Y Z YZ ZY i iX i i --⎡⎤⎡⎤⎡⎤⎡⎤⎡⎤=-=-==⎢⎥⎢⎥⎢⎥⎢⎥⎢⎥--⎣⎦⎣⎦⎣⎦⎣⎦⎣⎦[]1001011002,20110100120Z X ZX XZ iY ⎡⎤⎡⎤⎡⎤⎡⎤⎡⎤=-=-==⎢⎥⎢⎥⎢⎥⎢⎥⎢⎥---⎣⎦⎣⎦⎣⎦⎣⎦⎣⎦2.41 Verify the anti-commutation relations {},0i j σσ= where i j ≠ are both chosen from the set 1,2,3. Also verify that (i=0,1,2,3)2i I σ=Solution:{}12010001,0100010i i i i σσ--⎡⎤⎡⎤⎡⎤⎡⎤=+=⎢⎥⎢⎥⎢⎥⎢⎥⎣⎦⎣⎦⎣⎦⎣⎦{}23010100,0001010i i i i σσ--⎡⎤⎡⎤⎡⎤⎡⎤=+=⎢⎥⎢⎥⎢⎥⎢⎥--⎣⎦⎣⎦⎣⎦⎣⎦{}3110010110,001101001σσ⎡⎤⎡⎤⎡⎤⎡⎤=+=⎢⎥⎢⎥⎢⎥⎢⎥--⎣⎦⎣⎦⎣⎦⎣⎦2221230101001010,,1010000101i i I I I i i σσσ--⎡⎤⎡⎤⎡⎤⎡⎤⎡⎤⎡⎤======⎢⎥⎢⎥⎢⎥⎢⎥⎢⎥⎢⎥--⎣⎦⎣⎦⎣⎦⎣⎦⎣⎦⎣⎦2.42 Verify that []{},,2A B A B AB +=Solution:[]{},,22A B A B AB BA AB BA AB +-++==2.43 Show that for ,1,2,3j k =,31j k jk jkl l l I i σσδεσ==+∑Solution: when j=k ,j k jk I I σσδ==; when j k ≠ 31j k jkl l l i σσεσ==∑So for ,1,2,3j k =, 31j k jk jkl l l I i σσδεσ==+∑2.44 Suppose []{},0,,0A B A B ==, and A is invertible. Show that B must be 0. Solution: []{},0,,0A B AB BA A B AB BA =-==+= so 20,0AB AB == Because A is invertible, 0,0A AB B -==2.45 Show that [],,A B B A +++⎡⎤=⎣⎦Solution: [](),,A B AB BA B A A B B A ++++++++⎡⎤=-=-=⎣⎦2.46 Show that [][],,A B B A =-Solution: []()[],,A B AB BA BA AB B A =-=--=-2.47 Suppose A and B are Hermitian. Show that [],i A B is Hermitian. Solution:[]()()[],,,i A B iAB iBA iB A iA B i A B B A i A B i A B ++++++++++++⎡⎤=-=-+=-==⎣⎦2.48 What is the polar decomposition of a positive matrix P? Of a unitary matrix U? Of a Hermitian matrix , H?Solution: if P is a positive matrix, the polar decomposition of P is P=IP If U is a unitary matrix, the polar decomposition of U is U=UI If H is a Hermitian matrix, the polar decomposition of H is2.49 Express the polar decomposition of a normal matrix in the outer product representation. Solution:2.50 Find the left and right polar decompositions of the matrix 1011⎡⎤⎢⎥⎣⎦. Solution: let A= 1011⎡⎤⎢⎥⎣⎦, 101010111121A A +⎡⎤⎡⎤⎡⎤==⎢⎥⎢⎥⎢⎥⎣⎦⎣⎦⎣⎦,the eigenvalues of A A + is 1λ=,the eigenvector of λ is 01⎡⎤⎢⎥⎣⎦ 约当标准型:1010,22010P P -⎡⎤⎡⎤⎢⎥==⎢⎥⎢⎥⎣⎦⎣⎦ 11001110221200110P A AP -+⎡⎤⎡⎤⎡⎤⎡⎤⎢⎥==⎢⎥⎢⎥⎢⎥⎢⎥⎣⎦⎣⎦⎣⎦⎣⎦110121101P P -⎡⎤⎡⎤⎢⎥==⎢⎥⎢⎥⎣⎦⎣⎦=JA= 1011⎡⎤⎢⎥⎣⎦1011I ⎡⎤=⎢⎥⎣⎦2.51 Verify that the Hadamard gate H is unitary.Solution: 111111,111111H H H H HH I +++⎤⎡⎤=====⎥⎢⎥---⎦⎣⎦So the Hadamard gate H is unitary.2.52 Verify that 2H I =Solution: 2H H H I +==2.53 What are the eigenvalues and eigenvectors of H? Solution: the eigenvalues of H are121,1λλ==-For 11λ=,the eigenvector is 11⎡⎤⎥⎦, for 21λ=-, the eigenvector is11⎡⎤⎢⎥⎣⎦2.54SupposeAandBarecommutingHermitianoperators.Provethat()()()exp exp exp A B A B =+.Solution: ()()2211exp exp ......2!2!A B I A A I B B ⎛⎫⎛⎫=++++++ ⎪⎪⎝⎭⎝⎭()()()2232231133...2!3!I A B A AB BA B A A B AB B =+++++++++++ ()()()()2311...exp 2!3!I A B A B A B A B =+++++++=+2.55 Prove that ()12,U t t defined in equation (2.91) is unitary.Solution: ()()()()21211212,exp ,,exp iH t t iH t t U t t U t t h h +---⎡⎤⎡⎤==⎢⎥⎢⎥⎣⎦⎣⎦()()()()()()212121211212,,exp exp exp iH t t iH t t iH t t iH t t U t t U t t Ih h h h +------⎡⎤⎡⎤⎡⎤==+=⎢⎥⎢⎥⎢⎥⎣⎦⎣⎦⎣⎦(()21iH t t h- and()21iH t t h-- are commuting )So ()12,U t t defined in equation (2.91) is unitary.2.56 Use the spectral decomposition to show that ()log K i U ≡- is Hermitian for any unitaryU, and thus exp()U iK = for some Hermitian K. Solution:2.57 Suppose {}{}l m L and M are two sets of measurement operators. Show that a measurement defined by the measurement operators{}l L followed by a measurement defined by themeasurement operators {}m M is physically equivalent to a single measurement defined by measurement operators {}lm N with the representation lm m l N M L = Solution: Suppose the state of the quantum system is ϕ, after operators {}lm N measure, thestate of the system isIf a measurement defined by the measurement operators {}l L followed by a measurement defined by the measurement operators {}m M , after the measurement by operators {}l L , thestate become{}m M ,the state become==So we can see that a measurement defined by the measurement operators {}l L followed by a measurement defined by the measurement operators {}m M is physically equivalent to a single measurement defined by measurement operators {}lm N with the representation lm m l N M L =2.58 Suppose we prepare a quantum system in an eigenstateϕ of some observable M, withcorresponding eigenvalue m .What is the average observed value of M , and the standard deviation?(system state ?)Solution: the average observed value of M is2.59 Suppose we have qubit in the state 0, and we measure the observable X. What is theaverage value of X? What is the standard deviation of X?Solution: X⎛⎛=-⎪⎝⎪⎝⎭It gives the result +1 with probability1002=It gives the result -1 with probability1002⎛⎛=⎪⎝⎪⎝⎭the average value of X is 0. the standard deviation of X is 1.2.60 Show that vσ∙has eigenvalues 1±, and that the projectors onto the corresponding eigenspaces are given by ()2P I vσ±=±∙Solution: 312123v v ivvv iv vσ-⎡⎤∙=⎢⎥+-⎣⎦,321213v iv vviv v vλλσλ---∙==--+So 222222221231230,1,1,1v v v v v vλλλ---=++===±,,0mP P P IandP P P P P P P P+-+++---+-=+====∑v P Pσ+-∙=-So the projectors onto the corresponding eigenspaces are given by ()2P I vσ±=±∙2.61 Calculate the probability of obtaining the result +1 for a measurement of vσ∙, given that the state prior to measurement is 0.What is the state of the system after the measurement if +1 is obtained?Solution: after the measurement if +1 is obtained, the state of the system is312012I vvv ivσ+∙+⎡⎤=⎥+⎦2.65 Express the states ((0101+-in a basis in which they are not the same up to a relative phase shift.Solutiong: 1211,11v v ⎤⎤==⎥⎥-⎦⎦((1212010,010v v v v +=+-=+they are not the same up to a relative phase shift.2.66 Show that the average value of the observable 12X Z for a two qubit system measured in the state(0011+ is zero.Solution: 10100000,11,0011000011⎡⎤⎡⎤⎡⎤⎢⎥⎢⎥⎢⎥⎢⎥⎢⎥⎢⎥==+=⎢⎥⎢⎥⎢⎥⎢⎥⎢⎥⎢⎥⎣⎦⎣⎦⎣⎦ ,120010000110000100X Z ⎡⎤⎢⎥-⎢⎥=⎢⎥⎢⎥-⎣⎦the average value of the observable 12X Zis12X Z ϕϕ]0010100010100101000001001⎡⎤⎡⎤⎢⎥⎢⎥-⎢⎢⎥==⎢⎢⎥⎢⎢⎥-⎣⎦⎣⎦2.68 Prove thata b ψ≠ for all single qubit state a and b .Solution: 11111222212210,,01a b a b a b let a and b a b a b a b a b ψ⎡⎤⎡⎤⎢⎥⎢⎥⎡⎤⎡⎤⎢⎥⎢⎥====⎢⎥⎢⎥⎢⎥⎢⎥⎣⎦⎣⎦⎢⎥⎢⎥⎣⎦⎣⎦11122122,0a b a b f a b then a b a b ψ⎧=⎪=⎪=⎨=⎪⎪=⎩, it is impossible, so a b ψ≠ for all single qubit state a and b .2.69 Verify that the Bell basis forms an orthonormal basis for the two qubit state space.Solution:00011001,0110ββ⎡⎤⎡⎤⎢⎥⎢⎥⎥⎥====⎥⎥⎥⎥⎣⎦⎣⎦10011001,0110ββ⎡⎤⎡⎤⎢⎥⎢⎥⎢⎥⎢⎥====⎢⎥⎢⎥-⎢⎥⎢⎥-⎣⎦⎣⎦So the Bell basis forms an orthonormal basis for the two qubit state space.2.71 Let ρbe a density operator. Show that ()21trρ≤, with equality if and only if ρis a pure state.Solution: 22i j i j ji j i j ji j j i i j j j jρλλλλλ====∑∑∑∑()2221j jj jtr tr j jρλλ⎛⎫==≤⎪⎝⎭∑∑. If ρis a pure state, λ=1, ()21trρ=2.72 The Bloch sphere picture for pure states of a single qubit was introduced in Section 1.2. This description has an important generalization to mixed states as follows:(1) Show that an arbitrary density matrix for a mixed state qubit may be written as()2I rρσ=+∙, where r is a real three-dimensional vector such that 1r≤.This vector is known as the Bloch vector for the state ρ(2) What is the Bloch vector representation for the state 2Iρ=?(3) Show that a state ρis pure if and only if 1r=(4) Show that for pure states the description of the Bloch vector we have given coincides with that in Section 1.2.Solution: (1) 312123r r irrr ir rσ-⎡⎤∙=⎢⎥+-⎣⎦()3123121222122r r irI rrr irσ+-⎡⎤⎢⎥+∙=⎢⎥-+⎢⎥⎢⎥⎣⎦Because 1r≤,we have 2221231r r r++≤,3122221233312111220014222r r irr r r randrr ir+----+=≥≥-+,so()312312122122r r irI rrr irσ+-⎡⎤⎢⎥+∙=⎢⎥-+⎢⎥⎢⎥⎣⎦is a self-adjoint non-negative operator with unit trace.So an arbitrary density matrix for a mixed state qubit may be written as ()2I rρσ=+∙.(2) the Bloch vector representation for the state2I ρ= is r=0(3) ()222312322223123212*4412*4r r r r I r r r r r ρσ+⎡⎤+⎢⎥⎢⎥==-⎢⎥⎢⎥⎣+∙+⎦++++()222123212tr r r r ρ+++=,if and only if 1r =,()21tr ρ=So a state ρ is pure if and only if 1r =. (4)2.73 Let ρ be a density operator and supposei ψ is some linearly independent set of purestates spanning the support of ρ.(The support of a Hermitian operator A is the vector space spanned by the eigenvectors of A with non-zero eigenvalues.)Show that there is a unique probability distribution ()i p such thati i i ip ρψψ=∑, and moreover, the probabilitiesi p are given by 11i i ip ψρψ-=, where1ρ- is defined to be the inverse of ρ, when ρis considered as an operator acting only on the support of ρ.(This definition removes the problem that ρ may not have an inverse.) Solution:2.74 Suppose a composite of systems A and B is in the state a b , where a is a pure state of system A,and b is a pure state of system B.Show that the reduced density operator of system A alone is a pure state. Solution:()()()()(),AB A AB B B B a b a b tr tr a b a b tr a a b b a aρρρ====⊗=2.75 For each of the four Bell states, find the reduced density operator for each qubit. Solution: for Bell state(0011+,00001100001111112ρ+++==()()()()()1111100001100001111112tr tr tr tr tr ρ+++=000010100101111122I+++==()2tr ρ=。
磷石膏低温分解研究及进展

磷石膏低温分解研究及进展谢龙贵马丽萍*陈宇航戴取秀崔夏许文娟(昆明理工大学环境科学与工程学院,云南昆明650093)摘要:磷石膏是磷化工中产生量最大的工业废渣,其大量堆存导致严重的环境污染,同时也造成土地和硫资源的大量浪费。
将其用于制硫酸联产水泥能很好的解决磷石膏的堆存问题,然而,磷石膏热分解的高能耗成为了该项技术推广应用的瓶颈,因此,不少学者对磷石膏的低温分解做了一定的研究,研究表明,反应气氛和外加助剂对磷石膏分解温度的降低具有显著贡献。
关键词:磷石膏;低温分解;研究进展The Research Progress of Low-temperature Decomposition ofPhosphogypsumXIE Longgui, MA Liping*,CHEN Y uhang, DAI Quxiu, CUI Xia, XU Wenjuan(Faculty of Environmental Science and Engineering ,Kunming University of Science and Technology,Kunming 650093,China)Abstract: Phosphogypsum is the largest industrial solid waste of phosphorous chemical industry. Its store up in quantity causes serious environmental pollution, moreover, great amount of land and sulfur are wasted. To use phosphogypsum to produce sulphuric acid and cement can decrease its storage volume at a large extent, however, the decomposition of phosphogypsum under high temperature becomes the limits to popularize this technology. Lots of scholars have done a great amount of work to lower the decomposition temperature, the results showed that the reaction atmosphere and some additives do a significant contribution to lower the decomposition temperature.Keywords: Phosphogypsum, Low-temperature decomposition, Research progress基金项目:国家高技术发展计划“863”项目(2007AA06Z321)。
Cholesky-like Factorizations of Skew-Symmetric Matrices

AMS(MOS) subject classi cation: 15A23, 65F05
decompositions of the form (1.1) where (1.2)
1. Introdric matrix. We will study
B = RT JR;
n J = ?0 I0 : In As will be shown in this paper, there always exists an R such that (1.1) holds. We present a stable O(n3 ) algorithm that computes an R that has the form of a permuted triangular matrix. Our motivation comes from eigenvalue problems with Hamiltonian structure. A matrix H 2 R2n;2n is said to be Hamiltonian if (JH )T = JH and skew-Hamiltonian if (JH )T = ?JH . Example 1. The study of corner singularities in anisotropic elastic materials 5, 6, 11, 9] leads to generalized eigenvalue problems of the form 0 M M 0 v ?K 0 ? G M w = 0;
CHOLESKY-LIKE FACTORIZATIONS OF SKEW-SYMMETRIC MATRICES
Abstract. Every real skew-symmetric matrix B admits Cholesky-like factorizations B = RT JR i h Keywords: skew-symmetric matrices, matrix factorizations, Hamiltonian eigenproblems, com-
英语可数与不可数名词研究

英语可数与不可数名词研究一、引言在英语教学中,可数名词与不可数名词一直是让老师尴尬,学生痛苦的部分。
每次提及bread,chocolate,paper,water,money等不可数名词时,老师说的最多的就是:记下来,这个词不可数,没有复数,不能加s,做主语时,谓语动词用单数。
但当学生问及原因时,却总也不能做出很好的回答。
根据英语传统语法,如:张道真(1995),薄冰(1998),章振邦(1999),Quirk(1985)等学者以名词能否计数的特征来区分可数名词与不可数名词,这也确实给学习者带来了帮助但还是有其缺陷。
因为这种区分只能解决人们较熟悉的典型的可数名词和不可数名词但并不能解决很多的例外。
比如说bread 这个词,国内学生很难理解为什么它是不可数的。
与传统观念不同,美国语言学家Langacker(1987)从认知的视角出发系统地探讨了名词数的问题他依据有界性(Boundedness),同质性(Homogeneity),分割性(Contractibility)以及叠加性(Replicability)区分可数名词和不可数名词。
例如,car作为一个典型的可数名词,占有一定的空间,具有明确边界;car由引擎车轮方向盘座位车门等部件构成,内部不同质;car不可切分,拆卸后的car就不再是car;而car与car的叠加其结果为两个car。
与此相反,不可数名词water没有明确边界;内部同质,无论切分还是叠加,其结果依旧是water。
名词可数性的这种认知解释确实让人耳目一新,这种提法也看似可以解决所有的问题。
但它一方面过于抽象,另一方面也有其不可触及之处。
比如还是同样的问题,bread 一词是有界的,可分割的,但它却是不可数名词。
本文基于以上学者对可数与不可数名词区分的探讨,提出了一种可有效区分可数名词与不可数名词的通俗易懂的方法——分割法,一种顺应学生汉语思维方式来诠释可数名词与否的方法,以期让中学英语教师摆脱那种必须讲又讲不通的尴尬处境,让学生对英语学习更加充满激情,学起来更加轻松愉快。
- 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、"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、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)。
Short paths inε-regular pairs and small diameter decompositions of dense graphsJoanna Polcyn joaska@.plAndrzej Ruci´n ski rucinski@.plDepartment of Discrete MathematicsAdam Mickiewicz UniversityPozna´n,PolandAbstractIn this paper we give optimal optimal vertex degree conditions that guar-antee connection by short paths inε-regular bipartite graphs.We also studya related question of decomposing an arbitrary graph into subgraphs of smalldiameter.Keywords:ε-regular graphs,diameter,edge decomposition1IntroductionSince its discovery in late1970’s,the Szemer´e di Regularity Lemma has been a powerful tool in modern graph theory.Roughly speaking,this deep result from [16]provides a decomposition of every large graph into bipartite subgraphs most of which areε-regular pairs.Thus,naturally,there has been a growing interest in studying the properties of such graphs(see,e.g.,[1,2,3,4,5,6,7,10]).The goal of this paper is two-fold.First,in Section3we broaden our knowledge onε-regular pairs by determining the minimum degree of vertices that guarantees small diameter.It is quite easy to show(cf.Corollary3.3(i))that in anε-regular pair of order2n and density d≥εevery two vertices of degree at leastεn are con-nected by a short path.However,our main result in this direction(Theorem3.4) shows that this bound can be replaced by a quantity of order(ε2/d)n.Moreover, this is essentially optimal.The other result proved in this paper deals with graph decomposition into subgraphs of small diameter.Such decompositions have a potential application in distributed computing(see,e.g.,[12]).Here we approximate a given graph bya union of subgraphs with small diameter.It is easy to prove(cf.Proposition4.1),that for everyγ>0,the set of all but at mostγn2edges of every n-vertex graph G can be split into no more than1/γsubgraphs with diameter bounded from above by3/γ.In Theorem4.5,using the Szemer´e di Regularity Lemma,we1push the diameter down to four on the cost of an increase in the number of parts of the decomposition.Similar results for quasi-random3-uniform hypergraphs have been proved in[15]. 2PreliminariesLet G=(V,E)be a graph.For E0⊆E we denote by G[E0]=(V0,E0),where V0= e∈E0e,the subgraph of G induced by E0.Note that G[E0]does not need to be an(vertex)induced subgraph in the usual sense.By dist G(x,y)we denote the distance of vertices x,y∈V,that is,the length of a shortest path connecting them,if such a path exists.Otherwise we set dist G(x,y)=∞.By the diameter of G we mean diam(G)=max x,y∈V dist G(x,y).In particular, if G is not connected,then diam(G)=∞.For A⊆V,let N(A)be the set of all vertices adjacent in G to at least one vertex in A.In particular,if A={v}we write N(v).The number deg(v)=|N(v)| is the degree of vertex v in graph G.The minimum vertex degree is denoted by δG.In Section4we will use the inequality3|V|diam(G)<|U||W|.The number d G(U,W)is called the density of the graph G between U and W,or simply,the density of the pair(U,W).In the remainder of this section let G denote a bipartite graph with bipartition V=V1∪V2.A simple averaging argument yields the following fact.Fact2.1If d G(V1,V2)<d(resp.,d G(V1,V2)>d),then for all natural numbers ℓ1≤|V1|andℓ2≤|V2|there exist subsets U⊂V1,|U|=ℓ1and W⊂V2,|W|=ℓ2 with d G(U,W)<d(resp.,d G(U,W)>d).Note that eachε-regular graph isε′-regular for allε′≥ε.The last definition of this section deals with pairs of subsets of vertices with no edge in between.Definition2.3A pair of sets U⊆V1and W⊆V2with e G(U,W)=0is called a hole or a(|U|,|W|)-hole in G.A bipartite graph G is called(ℓ1,ℓ2)-holeless if there is neither an(ℓ1,ℓ2)-hole nor an(ℓ2,ℓ1)-hole in G.Note that,trivially,each(ℓ1,ℓ2)-holeless graph is also(ℓ′1,ℓ′2)-holeless for allℓ′i≥ℓi, i=1,2.Also,it follows from the definition ofε-regularity that,ifε≤d,then each (d,ε)-regular graph is(⌈ε|V1|⌉,⌈ε|V2|⌉)-holeless.However,it is proved in[13],that even much smaller holes are forbidden.Lemma2.4([13])(i)For all0<ε<d<1,d+ε≤1,there exists n0such that every(d,ε)-regularbipartite graph G with|V1|=|V2|=n≥n0is(ℓ,ℓ)-holeless for allℓ≥2ε(√d−εn+1.(ii)For all0<ε<d<1,d+ε<1,ε<(1−d)2/d and for allα<2ε(√d−εthere exists n0=n0(d,ε,α)such that for all n≥n0there exists a(d,ε)-regular graph with|V1|=|V2|=n containing an(⌈αn⌉,⌈αn⌉)-hole.3Short paths between vertices of sufficiently large degreesThis section is devoted to the problem of what minimum degree of vertices in anε-regular graph guarantees a small diameter.As a starting point recall an old result from[9]which says that almost all bipartite graphs with density d have diameter three.Although anε-regular pair resembles a random bipartite graph,in general nothing can be said about its diameter,because it may contain isolated vertices. On the other hand,it seems reasonable that vertices of sufficiently large degrees are connected by short paths.Indeed,below wefirst prove that degrees at least, roughly,(2ε3/2/√Proposition3.1Letℓ≥1and G be an(ℓ,ℓ)-holeless balanced bipartite graph.If deg(v)≥ℓand deg(w)≥ℓ,then dist G(v,w)≤4.Proof Without loss of generality we may assume that v∈V1.Considerfirst the case when w∈V2.By the assumption we have|N(v)|≥ℓand|N(w)|≥ℓ,and therefore,due to the absence of(ℓ,ℓ)-holes,there exists an edge e={x,y}with x∈N(v)and y∈N(w).Thus,vertices v,x,y,w form in G a path from v to w of length three(see Figure1).Considerfirst the case when v∈V1and w∈V2.Ifℓ≤⌈n/2⌉then there must be an edge between these two sets yielding a path of length at mostfive(see Figure 3).n+1d−εthen dist G(v,w)≤4.In particular,ifδG≥2ε(√εd−ε)/(d−ε)and sufficiently large n,there exists a(d,ε)-regular balanced bipartite graph G with2n vertices containing two vertices u,w∈V,deg(u)=deg(w)=⌊αn⌋,with dist G(u,w)≥5. Proof We only need to prove part(ii).By Lemma2.4(ii)there exists a(d,ε+ o(1))-regular graph with|V1|=|V2|=n−1containing an(⌈αn⌉,⌈αn⌉)-hole between sets A⊂V1and B⊂V2.We add two vertices,u and w,and all the edges connecting u with the vertices of B and w with the vertices of A.This way we obtain a new graph with the desired property.(We omit technical details here.)Again,the obtained bound is nearly optimal,namely there exist(d,ε)-regular graphs containing a vertex with degree only slightly smaller than2ε2n/(d+ε), which is disconnected from all vertices other than its neighbors.Theorem3.4(i)Let G be a(d,ε)-regular balanced bipartite graph with2n vertices,where0<ε≤1/2,ε<d<1,and n is sufficiently large.If2ε2nmin{deg(v),deg(w)}≥(|B|−|A|)(d+ε)|B|2<d+ε−αand note that0<γ≤ε−α.Let G′=(V′1∪V′2,E)be a(d+ε−γ,γ/2)-regular graph,where|V′1|=n−⌊αn⌋,|V′2|=n−1and n is sufficiently large.The existence6of such graphs can be proved easily using random graphs (see [13]).Notice that 0<d +ε−γ<1.Then,we add to V ′1a set A of ⌊αn ⌋vertices and a vertex w to V ′2,adjacent to all vertices of A .We claim that the graph G =(V 1∪V 2,E ),where V 1=V ′1∪A and V 2=V ′2∪{w },has the desired property.Since G [A ∪{w }]is an isolated star in G ,it remains to show that G is (d,ε)-regular.Let U ⊂V 1,W ⊂V 2,|U |,|W |≥εn .Set U ′=U \A and W ′=W \{w }and observe that |U ′|≥(ε−α)n >(γ/2)|V ′1|,and,for large n ,|W ′|>(γ/2)|V ′2|(see Figure4).2γ+O (n −1)<d +ε,and d G (U,W )≥d G ′(U ′,W ′) 1−|A ||W | > d +ε−3ε−O (n −1)>(d +ε−2γ) 1−2εε =d −ε+γ(d +ε)d +ε+2γd +ε+2γε.4An approximate decomposition into few sub-graphs with small diameterThis section provides two results about approximating a given graph by a union of subgraphs with small diameter.As a consequence of Proposition4.1below,for a givenγ>0,we can split the set of all but at mostγ|V(G)|2edges of G into no more than1/γsubgraphs with diameter bounded from above by3/γ.Then in Theorem4.5,using the celebrated Szemer´e di Regularity Lemma,we decrease the bound on the diameter to four,at the price of increasing the number of subgraphs in the partition.Thus,in a sense,every graph can be decomposed into a bounded number of subgraphs with small diameter,provided a small set of edges and/or vertices can be ignored.Proposition4.1Let0<γ<1.The set of edges of every graph G=(V,E)with |V|=n can be partitioned into k+1subsets,E=E0∪E1∪···∪E k,where k≤1/γand|E0|≤γn2,in such a way that for each1≤s≤k we have diam(G[E s])≤3/γ.Proof Keep removing from G,one by one,vertices of degree less thanγn,until the minimum degree of the obtained induced subgraph G′is at leastγn.Note that we have removed at mostγn2edges.Since for each connected component C of G′we have|V(C)|>δC≥γn,there are,obviously,at most1/γcomponents. Moreover,for each such component C,by(1)we havediam(C)<3|V(C)|γn=3Note that,in fact,the above proof yields a partition into vertex-disjoint sub-graphs.In Proposition4.1both bounds,on the diameter and on the number of sub-graphs G[E s],depend linearly on1/γ,and thus grow to infinity when the precision of approximation,γ,tends to zero.However,one can compromise on one of these bounds,improving the other to the extent that it becomes independent ofγ.In-deed,using the Szemer´e di Regularity Lemma[16]and Corollary3.3(ii),we will put the cap of four on the diameter,at the cost of letting the number of subgraphs in the partition to be an enormous constant.Before we make this precise,let us quote the Regularity Lemma.Definition4.2Let0<ε<1,t be a positive integer,and let G=(V,E)be a graph.We call a partition V=V0∪V1∪···∪V tε-regular,if|V1|=|V2|=···=|V t|, |V0|<t,and all but at mostε t2 pairs(V i,V j),1≤i<j≤t,areε-regular. Theorem4.3(Szemer´e di Regularity Lemma,[16])Let0<ε<1and let t0be a positive integer.There exist integers N=N(ε,t0)and T=T(ε,t0)such that every graph G=(V,E)with|V|≥N vertices admits anε-regular partition V=V0∪V1∪···∪V t with t0≤t≤T.8In the proof of Theorem4.5below,we will also need a simple observation that forε≤d/3,every(d,ε)-regular graph can be approximated by a subgraph of diameter at most four.Fact4.4If0<3ε≤d<1then G contains an induced subgraph G′such that |E(G)\E(G′)|≤2εn2and diam(G′)≤4.Proof For each i=1,2,letU i={v∈V i:deg(v)≤2εn}.Since3ε≤d,it follows easily from the(d,ε)-regularity of G that|U i|<εn,i=1,2. Let W i be such that U i⊆W i⊆V i and|W i|=⌊εn⌋:=m,i=1,2.Finally,let G′=G−(W1∪W2)be the graph obtained from G by removing all vertices of W1∪W2,together with their incident edges.Let E0be the set of all removed edges.Note that G′=G[E\E0]and that |E0|≤(2m)n≤2εn2.Moreover,it is easy to see that G′is(d,ε′)-regular,where ε′=εn/(n−m)<d,whileδG′≥2εn−εn=εn=ε′(n−m)≥...n.Therefore,by Corollary3.3(i),diam(G′)≤4.n2.5Let F2be the set of all edges of G contained in one of the sets V i,i=1,2,...,t. We have⌊n/t⌋2 <n22t0≤γ|F2|≤tt 2<ε10n2.9Finally,let F 4be the set of all edges belonging to the pairs (V i ,V j )with density d G (V i ,V j )≤4ε.Then|F 4|≤ t 2 4ε n 5n 2.Consequently,setting E ′0=F 1∪F 2∪F 3∪F 4,we have|E ′0|<γ10n 2+γ5n 2=4t 2=2t 2,such that diam(G [E ′s \E 0s ])≤4.Altogether we have deleted a set E ′′0=ks =1E 0s of at most t 2 2t2<15γn 2+1Note that the proofs of Proposition 4.1and Theorem 4.5both yield algorithms of complexity O (n 2)which construct the respective partitions.The latter is based on a O (n 2)-time algorithm found in [11]which builds an ε-regular partition in every n -vertex graph.Acknowledgements:We would like to thank both Referees for their valuable comments on the presentation of our paper.Most importantly,Referee A suggested to look at the problem of minimum degree guaranteeing paths of length at most four (Corollary 3.3)and sketched a solution to it.He also supplied us with an alternative construction for the proof of Theorem 3.4(ii)and observed the algorithmic aspects of Proposition 4.1and Theorem 4.5discussed in the previous paragraph.References[1]N.Alon,R.A.Duke,H.Lefmann,V.R¨o dl &R.Yuster,The algorithmicaspects of the regularity lemma,Journal of Algorithms 16(1)(1994)80-109.[2]N.Alon,V.R¨o dl &A.Ruci´n ski,Perfect matchings in ε-regular graphs TheElectr.J.of Combin.5(1)(1998)#R13.10[3]A.Czygrinow&B.Nagle,Matrix-free proof of a regularity characterization,bin.10(2003)#R39.[4]A.Czygrinow&B.Nagle,Strong edge colorings in uniform graphs DiscreteMathematics,286(3)(2004)219-223.[5]R.A.Duke,H.Lefmann&V.R¨o dl,A fast approximation algorithm for com-puting the frequencies of subgraphs in a given graph.SIAM put.24 (1995)598-620.[6]A.Frieze,On the number of perfect matchings and Hamilton cycles in epsilon-regular non-bipartite graphs Electr.J.of Combin.7(2000)R57.[7]A.Frieze&M.Krivelevich,On packing Hamilton cycles in epsilon-regulargraphs J.of Combin.Th.B94(2005)159-172.[8]S.Janson,T. L uczak&A.Ruci´n ski,Random Graphs,John Wiley and Sons,New York(2000).[9]V.L.Klee,rman&E.M.Wright,The diameter of almost all bipartitegraphs,Studia Scientiarum Mathematicarum Hungarica15(1980)39-43. [10]Y.Kohayakawa&V.R¨o dl,Regular pairs in sparse random graphs I,RandomStruct.Algorithms22(4)359-434(2003).[11]Y.Kohayakawa,V.R¨o dl&L.Thoma,An optimal algorithm for checkingregularity,SIAM put32(2003)2010-2035.[12]N.Linial&M.Saks,Low diameter graph decompositions,Combinatorica13(4)(1993)441-454.[13]J.Polcyn,Large holes in quasi-random graphs,Electr.J.of Combin.To ap-pear.[14]J.Polcyn,Short paths in3-uniform quasi-random hypergraphs,DiscussionesMathematicae:Graph Theory24(2004)469-484.[15]J.Polcyn,V.R¨o dl,A.Ruci´n ski&E.Szemer´e di,Short paths in quasi-randomtriple systems with sparse underlying graphs,J.of Combin.Th.,Series B96 (2006)584-607.[16]E.Szemer´e di,Regular partitions of graphs,Probl`e mes en Combinatoire etTh´e orie des Graphes,Proc.Colloque RS,(Bermond,J.-C.,Fournier, J.-C.,Las Vergnas,M.,Sotteau,D.,eds),(1978)399-401.11。