2013年高教社杯全国大学生数学建模竞赛B题优秀论文资料

合集下载

全国大学生数学建模竞赛题目B题

全国大学生数学建模竞赛题目B题

B题公交车调度
公共交通是城市交通的重要组成部分,作好公交车的调度对于完善城市交通环境、改进市民出行状况、提高公交公司的经济和社会效益,都具有重要意义。

下面考虑一条公交线路上公交车的调度问题,其数据来自我国一座特大城市某条公交线路的客流调查和运营资料。

该条公交线路上行方向共14站,下行方向共13站,第3-4页给出的是典型的一个工作日两个运行方向各站上下车的乘客数量统计。

公交公司配给该线路同一型号的大客车,每辆标准载客100 人,据统计客车在该线路上运行的平均速度为20公里/小时。

运营调度要求,乘客候车时间一般不要超过10分钟,早高峰时一般不要超过5分钟,车辆满载率不应超过120%,一般也不要低于50%。

试根据这些资料和要求,为该线路设计一个便于操作的全天(工作日)的公交车调度方案,包括两个起点站的发车时刻表;一共需要多少辆车;这个方案以怎样的程度照顾到了乘客和公交公司双方的利益;等等。

如何将这个调度问题抽象成一个明确、完整的数学模型,指出求解模型的方法;根据实际问题的要求,如果要设计更好的调度方案,应如何采集运营数据。

2013年数学建模国赛b题

2013年数学建模国赛b题

2013年数学建模国赛b题
摘要:
1.背景介绍:2009 年3 月合肥市非国有建筑专业职称资格评审通过人员名册
2.名册内容:通过人员名单、职称、资格等信息
3.意义:对非国有建筑行业的专业人才的肯定和鼓励
正文:
2009 年3 月,合肥市对非国有建筑专业职称资格进行了评审,并通过了一份详细的名册。

这份名册包含了通过人员名单、职称、资格等信息,是对非国有建筑行业的专业人才的肯定和鼓励。

在这个名册中,我们可以看到各位通过人员的姓名、工作单位、评审职称以及资格等信息。

他们经过了严格的评审,最终脱颖而出,获得了相应的职称资格。

这不仅是他们个人努力的结果,也是他们所在单位和行业的认可。

这份名册的意义不仅在于对个人的肯定,更在于对整个非国有建筑行业的
推动。

它鼓励了更多的专业人士积极投身于建筑行业,提高了整个行业的专业水平。

同时,它也为行业内外提供了一个参考,让人们更好地了解非国有建筑行业的发展和人才状况。

2013年美国大学生数学建模大赛B题获奖论文

2013年美国大学生数学建模大赛B题获奖论文

For office use onlyT1________________ T2________________ T3________________ T4________________Team Control Number22599Problem ChosenAFor office use onlyF1________________F2________________F3________________F4________________ 2013Mathematical Contest in Modeling(MCM/ICM)Summary Sheet(Attach a copy of this page to your solution paper.)Heat Radiation in The OvenHeat distribution of pans in the oven is quite different from each other,which depends on their shapes.Thus,our model aims at two goals.One is to analyze the heat distri-bution in different ovens based on the locations of electrical heating cubes.Further-more,a series of heat distribution which varies from circular pans to rectangular pans could be got easily.The other is to optimize the pans placing,in order to choose a best way to maximize the even heat and the number of pans at the same time.Mathematically speaking,our solution consists of two models,analyzing and optimi-zing.In part one,our whole-local approach shows the heat distribution of every pan.Firstly,we use the Stefan-Boltzmann law and Fourier theorem to describe the heat distribution in the air around the electrical heating tube.And then, based on plane in-tercept method and simplified Monte Carlo method,the heat distribution of different shapes of pans is obtained.Finally,we explain the phenomenon that the corners of a pan always get over heated with water waves stirring by analogy.In part two,our discretize-convert approach optimizes the shape and number of the pans.Above all,we discre-tize the side length of the oven, so that the number and the average heat of the pans vary linearly.In the end,the abstract weight P is converted into a specific length,in order to reach a compromise between the two factors.Specially,we create a unique method to convert the variables from the whole space to the local section.The special method allows us to draw the heat distribution of every single section in the oven.The algorithm we create does a great job in flexibility,which can be applied to all shapes of pans.Type a summary of your results on this page.Do not includethe name of your school,advisor,or team members on this page.Heat Radiation in The OvenSummaryHeat distribution of pans in the oven is quite different from each other,which depends on their shapes.Thus,our model aims at two goals.One is to analyze the heat distri-bution in different ovens based on the locations of electrical heating cubes.Further-more,a series of heat distribution which varies from circular pans to rectangular pans could be got easily.The other is to optimize the pans placing,in order to choose a best way to maximize the even heat and the number of pans at the same time.Mathematically speaking,our solution consists of two models,analyzing and optimi-zing.In part one,our whole-local approach shows the heat distribution of every pan. Firstly,we use the Stefan-Boltzmann law and Fourier theorem to describe the heat distribution in the air around the electrical heating tube.And then,based on plane in-tercept method and simplified Monte Carlo method,the heat distribution of different shapes of pans is obtained.Finally,we explain the phenomenon that the corners of a pan always get over heated with water waves stirring by analogy.In part two,our discretize-convert approach optimizes the shape and number of the pans.Above all, we discre-tize the side length of the oven,so that the number and the average heat of the pans vary linearly.In the end,the abstract weight P is converted into a specific length,in order to reach a compromise between the two factors.Specially,we create a unique method to convert the variables from the whole space to the local section.The special method allows us to draw the heat distribution of every single section in the oven.The algorithm we create does a great job in flexibility, which can be applied to all shapes of pans.Keywords:Monte Carlo thermal radiation section heat distribution discretizationIntroductionMany studies on heat conduction wasted plenty of time in solving the partial differential equations,since it’s difficult to solve even for computers.We turn to another way to work it out.Firstly,we study the heat radiation instead of heat conduction to keep away from the sophisticated partial differential equations.Then, we create a unique method to convert every variable from the whole space to section. In other words,we work everything out in heat radiation and convert them into heat contradiction.AssumptionsWe make the following assumptions about the distribution of heat in this paper.·Initially two racks in the oven,evenly spaced.·When heating the electrical heating tubes,the temperature of which changes from room temperature to the desired temperature.It takes such a short time that we can ignore it.·Different pans are made in same material,so they have the same rate of heat conduction.·The inner walls of the oven are blackbodies.The pan is a gray body.The inner walls of the oven absorb heat only and reflect no heat.·The heat can only be reflected once when rebounded from the pan.Heat Distribution ModelOur approach involves four steps:·Use the Fourier theorem to calculate the loss energy when energy beams are spread in the medium.So we can get the heat distribution around each electrical heating tube.The heat distribution of the entire space could be go where the heat of two electrical heating tubes cross together.·When different shapes of the pans are inserted into the oven,the heat map of the entire space is crossed by the section of the pan.Thus,the heat map of every single pan is obtained.·Establish a suitable model to get the reflectivity of every single point on the pan with the simplified Monte Carlo method.And then,a final heat distribution map of the pan without reflection loss is obtained.·A realistic conclusion is drawn due to the results of our model compared with water wave propagation phenomena.First of all,the paper will give a description of the initial energy of the electrical hea-ting tube.We see it as a blackbody who reflects no heat at all.Electromagnetic know-ledge shows that wavelength of the heat rays ranges from um 110−to um 210as shown below[1]:Figure 1.Figure 2.We apply the Stefan-Boltzmann’s law[2]whose solution is ()1/512−=−T c b e c E λλλ(1)()λλλλλd e c d E E T c b b ∫∫∞−∞−==0/51012(2)Where b E means the ability of blackbody to radiate. 1c and 2c are constants.Obviously,,the initial energy of a black body is )(0122398.320m w e E b ×+=.Combine Figure 1with Figure 2,we integrate (1)from 1λto 2λto get the equation as follow:λλλλλλd E E b b ∫=−2121)((3)Figure 3.From Figure 3,it can be seen how the power of radiation varies with wavelength.Secondly,based on the Fourier theorem,the relation between heat and the distance from the electrical heating tubes is:dxdt S Q λ−=(4)Where Q is the power of heat (W s J =/),S is the area where the energy beamradiates (2m ),dxdt represents the temperature gradient along the direction of energy beam.[3]It is known that the energy becomes weaker as the distance becomes larger.According to the fact we know:dxdQ =ρ(5)Where ρis the rate of energy changing.We assume that the desired temperature of electrical heating tube is 500k.With the two equations,the distribution of heat is shown as follow:Figure 4.(a)Figure 4.(b)In order to draw the map of heat distribution in the oven,we use MATLAB to work on the complicated algorithm.The relation between the power of heat and the distance is shown in Figure4(a).The relation between temperature and distance is presented in Figure4(b).The spreading direction of energy beam is presented in Figure5.Figure5.The shape of electrical heating tube is irregular.The heat distribution of a single electrical heating tube can be draw in3D space with MATLAB.The picture is shown in Figure6.After superimposing,the total heat distribution of two tubes is shown below in Figure7and Figure8.Figure6.Figure7.Figure8.The pictures above show the energy in an oven with no pan.We put in a rectangular pan whose area is A,and intercept the maps with MATLAB.The result is show in Figure9.Figure9.Figure10.Put in a circular pan to intercept the maps,whose area is A,also.The distribution of heat is shown in Figure11.Figure11.When put in a pan in transition shape,which is neither rectangular nor circular.The area of it is A,also.The heat distribution on such a pan is shown as follow:Figure12(a)Figure12(b).Figure13.Next,learning from the Monte Carlo simulation[4],a model is established to get obtain the reflectivity.We generate a random number between0and1to determine if the energy beam on certain point is reflected.•Firstly,to demonstrate the question better,we construct a simple model:Figure 14.Where θis the viewing angle from electrical heating tube to the pan.360θ=R is the proportion of the beams radiated to the pan.•What is more,we assume the total beam is 1M .Ideally,the number of absorption is3601θ×M .Then,each element of the pan is seen as a grid point.Each grid point can generate a-3601θ×M -random-number vector between 0and 1in MATLAB.•After MATLAB simulating,the number of beams decreased by 2M ,due to thereflection.So we define a probability θρ12360M M ×=to describe the number of beams reflected.The conclusion is :•If R ≤ρ,the energy beam is absorbed.•If R >ρ,the energy beam is reflected.[5]Based on the analysis above,our model get a final result of heat-distribution on the pan as shown below:Figure 15(a)Figure15(b).The conclusion is known that the closer the shape of pans is to circle,the more evenly the heat is distributed.Moreover,the phenomenon that the corners always get over heated can be explained by water wave propagation in different containers.When there is a fluctuation in the center of the water,the ripples will fluctuate and spread in concentric circles,as shown in Figure16.The fluctuation stirs waves up when contacting the pared with the waves with one boundary,the waves in corner make a higher amplitude.The thermal conduction on the pan is exactly the inverse process of the waves propagation.The range of thermal motion is much smaller than it on the side.That’s why the corners is easy to get over heated.In order to make the heat evenly distributed on the pan,the sides of the pan should be as few as possible.Therefore,if nothing is considered about the utilization of space,a circle pan is the best choice.Figure16,the water waves propagation[6]According to the analysis above and Figure7,the phenomenon shows that the heat conduction is similar to water waves propagation.So it is proved that heatconcentrates in the four corners of the rectangular pan.The Super Pan ModelAssumptions•The width of the oven(W)is mm100,the length is L.•There are three pans at most in vertical direction.•Each pan’s area is A.The first part.Calculate the maximum number of pans in the oven.Different shapes of pan have different heat distribution which affects the number of pans,judging from the previous solution.According to the conclusion in the first model,the heat is distributed the most evenly on a circular pan rather than a rectangular one.However, the rectangular pans make fuller use of the space the space than circular ones.Both factors considered,a polygonal pan is chosen.A circle can be regard as a polygon whose number of boundaries tends to infinity. Except for rectangle,only regular hexagon and equilateral triangle can be closely placed.Because of the edges of equilateral triangle,heat dissipation is worse than rectangle.So,hexagonal pans are adopted after all the discussion.Considering the gaps near boundaries,we place the hexagonal pans closely attached each other on the long side L.There are two kinds of programs as shown below.Program1.Program2.Obviously,Program2is better than Program1when considering space utilization.So scheme 1is adopted.Then,design a size of each hexagonal pan to make the highest space utilization.With the aim of utilization,hexagonal pans has to be placed contact closely with each other on both sides.It is necessary to assume a aspect ratio of the oven to work out the number of pans(N ).Assume that the side length of a regular hexagon is a ,the length-width ratio of the oven is λand L ∆is the increment in discretization.Because the number of pans can not change continuously when ⋅⋅⋅=+∈3,2,1),1,(m m m n ,the equations would be as follows.⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎩⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎨⎧⎥⎥⎥⎥⎦⎤⎢⎢⎢⎢⎣⎡⋅−====∆⋅+<⎥⎦⎤⎢⎣⎡∆⋅+−∆⋅+≤⋅=∆∆⋅+=<<=a W L n W L N Lk L W W L k L W L k L a L L k L L L W aW 23,810;23105000λ(6)Result:⎪⎪⎩⎪⎪⎨⎧⋅⋅⋅==⋅+=⋅⋅⋅=−=+−⋅+=3,2,1,2233,2,1,1212130201k k n n N N k k n n N N Where 1N represents the number of pans when n is odd,2N represents the number of pans when n is even.The specific number of pans is depended on the width-length ratio of oven.The second part.Maximum the heat distribution of the pans.We define the average heat(H )as the ratio of total heat and total area of the pans.Aiming to get the most average heat,we set the width-length ratio of the oven λ.Space utilization is not considered here.A conclusion is easy to draw from Figure 8that a square area in the oven from 150mm to 350mm in length shares the most heat evenly.So the pans should beplaced mainly in this area.From model1we know that the corners of the oven are apt to gather heat.Besides,four more pans are added in the corners to absorb more heat. Because heat absorbing is the only aim,there is no need to consider space utilization. Circular pans can distribute heat more evenly than any other shape due to model1.So circular pans are used in Figure17.Figure8.Figure17.We set the heat of the pans in the most heated area(the middle row)as Q.Pans in the corners receive more heat but uneven theoretically.And the square of the four pans in the corners is so small compared with the total square that we set the heat of the four as Q too.When the length of oven(L)increases,the number of pans increases too. It makes the square of the gaps between pans bigger,meanwhile.If each pan has a same radius(r)and square(A),the equation about average heat,length-width ratio and number of pans would be(7).⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎩⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎨⎧=+=⎥⎦⎤⎢⎣⎡⋅−====∆⋅+<⎦⎤⎢⎣⎡∆⋅+−∆⋅+≤⋅=∆∆⋅+=<<⋅=⋅⋅=...3,2,12;71021053410002k nN N r W L n W L N L k L W W L k L W L k L r L L k L L L W r A W r λπ(7)Here we get the most average heat (H ):29400WQ H ⋅=πThe third part.We discussed two different plans in the previous parts of the paper.One is aimed to get the most average heat,while the other aimed to place the most pans.The two plans are contradictory with each other,and can not be achieved together.Firstly,the weight of plan 1is P and the weight of plan 2is P −1.Obviously,this kind optimization has difficulty in solving and understanding.So we turn to another way to make it a easier and linear question.It has been set that the width of the oven is a constant W and there should be three pans at most in vertical direction.We make the weight P a proportion of the two plans.Thus the two plans could be achieved together due to proportion P and P −1,as shown in Figure18.Figure 18.As been told in model 1,the corners have a higher temperature than other parts of the oven.So plan 1is used in district 1(in Figure 10)and plan 2is used in district 2(in Figure 10).A better compromise could be reached in this way,as shown in Figure 19.Figure 19.Every pan has a square of A .Radius of circular ones is r .Side length of regular hexagon is a .1.1:23322=⇒⋅=⋅r a a r π(8)Based on the equation (8),if the pans are placed as shown in Figure 19,regular hexagons are placed full of district 1,the circular ones will be placed beyond the border line.If the circular ones are placed full of district 2,there will be more gaps in district 1,which will be wasted.So we change our plan of placing pans as Figure 20.Figure 20.The number of circular ones decreases by two,but the space in district 1is fully used,and no pan will be placed beyond the borderline.We assume that P is bigger than P −1,so that,the heat in district 1will be fully used.By simple calculating,we know that the ratio of the heat absorbed in circular pan (1H )and in regular hexagon (2H )is 1.2:1.Figure21.So,based on the pans placing plan,a equation on heat can be got as follow:⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎩⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎨⎧=⎥⎦⎤⎢⎣⎡−⋅−=⎥⎦⎤⎢⎣⎡−⋅====∆⋅+<⎥⎦⎤⎢⎣⎡∆⋅+−∆⋅+≤=∆∆⋅+=<<⋅=⋅==≈...3,2,1)1(,911233;23212kxWLPnxWLPnWLNLkLWWLkLWLkLxLLkLLLWraAxraλπ(9)Resolution:⎪⎪⎪⎪⎪⎪⎪⎩⎪⎪⎪⎪⎪⎪⎪⎨⎧=⋅⋅+⋅+⋅⋅+=⋅⋅+⋅+⋅⋅−+==⋅+⋅+=−=+−⋅+⋅+=...3,2,1)24(2.1)325()24(2.1)3215()2(232)12(12132221212111121211201k A N n Q Q n H A N n Q Q n H k n n n N N k n n n N N (10)1N and 1H means the number of pans and average heat absorbed when n is odd.2N and 2H means the number of pans and average heat absorbed when n is even.For example:(1)When 37.0=λ,6.0=P :16=N ,AQ H 075.1=.The best placing plan is:(2)When 37.0=λ,7.0=P :18=N ,AQ H ⋅=044.1.The best placing planis:(3)When 58.0=λ,6.0=P :12=N ,AQ H ⋅=067.1.The best placing plan is:A conclusion is easy to draw that when the ratio of width and length of the oven (λ)is a constant,the number of pans increases with an increasing P,but the average heat decreases (example (1)and (2)).When the weight P is a constant,the number of pans decreases with an increasing λ,and the average heat decreases also.So,the actual plan should be base on your specific needs.ConclusionIn conclusion,our team is very certain that the method we came up with is effective in heat distribution analysis.Based on our model,the more edges the pan has,the more evenly the heat distribute on.With the discretize-convert approach,we know that when the ratio of width and length of the oven (γ)is a constant,the number of pans increases with an increasing P ,but the average heat decreases.When the weight P is a constant,the number of pans decreases with an increasing γ,and the average heat decreases also.So,the actual plan should be base on your specific needs.Strengths &WeaknessesStrengths•Difficulties Avoided Avoided..In model 1,we turn to another way to work simulate the heat distribution instead of work on heat conduction directly.Firstly,we simulate heat radiation not heat conduction to keep away from the sophisticated partial differential equations.Then,we create a unique method to convert every variable from the whole space to section.In other words,we work everything out in heat radiation and convert them into heat contradiction.•Close to Reality.Our model considers both the thermal radiation and surface reflection,which is relatively close to the actual situation.•Flexibility Provided.Our algorithm does a great job in flexibility.The heat distribution map on sections are intercepted from the heat distribution maps of the entire space.All shapes of sections can be used in the algorithm.The heat distribution in the whole space is generated based on the location of the electrical heating tubes and the decay curve of the heat, which can be modified at any time.•Innovation.Based on our model,the space of an oven can be divided into six parts with different hear distribution.In order to make full use of the inner space,we invent a new pan which allows users to cook six different kinds of food at same time.An advertisement is published in the end of the paper.WeaknessesPan’’s Thermal Conductivity Ignored.•PanThe heat comes from not only the electrical heating tubes,but also heat conduction of the pans themselves.But the pan’s thermal conduction is ignored in the model,which may cause little inaccuracy.•Thermal Conductivity of Electrical Heating Tubes IgnoredIgnored..it is assumed that there are two electrical heating tubes in the oven and placed in a specific location.The initial temperature of the tubes is a desired constant temperature. In other words,the time electrical heating tubes spend to heating themselves is ignored.The simplification can cause some inaccuracy.simplification..•Linear simplificationIn model2,the length of the oven is discretized,so that the number of pans will changes linearly.calculating through simple integer linear method.This will lead to the result of our model is not accurate enough.ApplicationWe have discussed the heat distribution in the oven in model1.The heat distributionis shown in figure1and figure2.Figure1Figure2As shown,the edges of the oven are distributed the most heat.Areas on both sides of the,is distributed the least heat.While the middle area absorbs little less than theedges.So,we can separate the oven area into six parts,as shown bellow.Part1and part2are distributed the least heat and located the furthest from the heat source(the electrical heating tubes locate on the bottom of the oven).So these two parts absorb the least heat.Part3and part4are distributed the least heat but locating the nearest to the heat source.Part5located far from the bottom but distributed the most heat.So simply,we regard the heat of part3,part4and part5as the same.Part6 is distributed the most heat,and locating nearest to the bottom.So,the heat part6 absorbs is the most in the oven.Based on our conclusion above,we invent the iPan,a new combined pan,which can bake three kinds of food at the same time.For example,one wants to have a little bread,pieces of sausage,a chicken wing and a pizza for lunch.He will have to wait 30minutes at least for his lunch,if he just has one oven.As the Chinese saying goes,‘Bear paws and fish never come together’.By using iPan can solve the issue for him,he could put the bread in pan1,pizza in pan2,sausage in pan5and chicken wing in pan6,and power on.Thus,he can have his delicious lunch in at least10minutes.So,bear paws and fish come together.We make an advertisement for Brownie Gourmet Magazine in the end of the paper.Advertising SheetsReferences[1]Heat Radiation,/view/f5ed1619cc7931b765ce1599.html, Page.4[2]G.S.Ranganath,Black-body Radiation,/article/10.1007%2Fs12045-008-0028-7?LI=true#,February, 2013[3]Kaiqing Lu,The Chemical Basis of Heat Transfer,Journal of Higher Correspondence Education(Natural Sciences Edition),Vol.3:p.33,1996[4]Mark M.Meerschaert,Mathematical Modeling(Third Edition),China:China Machine Press,May.2009[5]Jianzhong Zhang,Monte Carlo Method,Mathematics in Practice and Theory,Vol.1p.28,1974[6]Shallow water equations,/wiki/Shallow_water_equations。

大学生数学建模竞赛B题优秀论文

大学生数学建模竞赛B题优秀论文

关于高等教育学费标准的评价及建议摘要本文通过对近几年来学费变化的研究,综合分析影响学费变化的五个要素,引入了三个变因:学校属性、专业类型、地域差异对学费的影响,对其合理性进行了定量的分析和评价。

首先,我们基于层次分析法建立了模型一。

模型一以五个要素,即教育市场供求关系、全国家庭支付承受力、国家财政及相关社会捐助、个人收益率、教育成本为方案层。

对于教育市场的供求关系我们用灰色预测GM(1,1)模型预测出未来几年的招生人数,用蛛网模型求解稳定的价格点为3225.51 元;对于国家财政及相关社会捐助,我们用回归分析得出其效应关系。

模型一以效率和公平两个标准作为准则层,应用极差归一化思想,构造指标函数,综合建立成对比较矩阵。

我们定义学费合理化指数为目标层,经准则层,得出五个要素对学费合理化指数的组合权重向量。

考虑到成对比较矩阵仍有一定主观因素,我们用熵值取权法修正组合权重向量。

最后,拟合出最佳学费曲线及其波动区间,其中 2007 年的结论值为 3370.75 元。

模型一的突出优点是客观可信,美中不足的是结论为一个平均最优值,没有考虑其他变因的影响,使用的局限性较大。

然后,我们基于学校属性、专业类型、地域差异三个变因对结论的影响建立了模型二。

评价了这三个变因对五个要素的综合影响,修正了五个要素对学费合理化指数的影响,使得结论更趋于合理,应用范围更加广泛。

修正后通过若干数据的检验,得出平均最佳学费约为 3000 元。

基于这两个模型,以及对高校学费现状的了解,我们提出三点主要建议: 1.鼓励高校开拓资金来源渠道,学习国外筹款方式,如发行教育彩票等; 2.建议国家增加助学贷款发放力度,并能够分类别基于不同金额的贷款,并出台一些补贴政策弥补不同地区的差异; 3.大力扶持民办高等院校发展,实现高等教育大众化,这样不仅缓解高等院校招生压力,并且能够促进高校教育健康发展。

本文的特色在于基于翔实丰富的资料,根据五个要素及三个变因的分析,建立了一种合理的高校学费评价体系,其拥有适用性广,稳定性好,灵敏度高等特点,对三个变因,即学校属性、专业类型、地域差异进行了深入定量的分析,并根据模型结论给提出了我们的一些可行性建议。

全国大学生数学建模竞赛B题优秀论文

全国大学生数学建模竞赛B题优秀论文
2.2 模型的符号说明
(1) 表示客流量随时间的变化值,R、RW、RG分别表示上海国际旅游入境人数本底值、外国游客入境人数本底值、港澳台游客入境人数本底值;
(2)R1表示2010年1、2、3、4、11、12月上海国际旅游入境实际人数,R2表示世博会期间上海国际旅游入境实际人数,RZ表示2010年上海国际旅游总入境实际人数;
最后,通过对模型结果的分析,量化评估上海世博会的影响力。从世博会对以上各个指标的贡献率可以看出:世博会极大地促进了旅游业的发展,并且对上海的财政收入做出了巨大的贡献。在分析所得结果的基础上,客观评价此模型,并指出其优点和缺点。
关键词:上海 世博会 影响力 本底趋势线 内插值
1.问题重述
2010年上海世博会是首次在中国举办的世界博览会。从1851年伦敦的“万国工业博览会”开始,世博会正日益成为各国人民交流历史文化、展示科技成果、体现合作精神、展望未来发展等的重要舞台。请你们选择感兴趣的某个侧面,建立数学模型,利用互联网数据,定量评估2010年上海世博会的影响力。
2.模型的假设与符号说明
2.1模型的假设
2010年上海世博会作为一场世界级的盛宴,要对其影响力进行定量评估,尚存在一些不确定因素。故为了研究方便,我们给出以下假设:
(1)假设世博会不受偶然事件严重冲击和干扰;
(2)假设旅游人数只受主要因素影响,其他一些因素可以忽略,比如天气等因素;
(3)假设世博会期间每月游览总人数波动不大,非世博会期间每月游览总人数波动也不大。
第二步,用Excel的指数模型、乘幂模型和SPSS的指数-三角函数复合模型 、直线-逻辑线增长复合模型 、直线-三角函数复合模型 对各个指标进行拟合,确定有关参数,获得各个指标的趋势线模型和方程,并计算各年的本底值;

全国数学建模B题论文

全国数学建模B题论文

全国数学建模B题论文 TYYGROUP system office room 【TYYUA16H-TYY-TYYYUA8Q8-2004高教社杯全国大学生数学建模竞赛题目B题电力市场的输电阻塞管理摘要本文是基于电力市场交易规则和输电阻塞管理原则,对电力市场的输电阻塞管理进行研究,提出合理假设,建立了较完善的模型,很好的解决了在电力市场中存在的一些问题。

问题一:首先作图分析数据,得出单个机组与线路潮流有线性关系,建立多元线性回归模型,用matlab进行拟合求出各线路上有功潮流关于各发电机组出力的近似表达式。

问题二:序内容量不出力的部分按清算价与修改后方案中该机组最后被选入段价之差补偿,序外容量多出力的部分根据该机组最后一个被选入的段价为进行补偿,两部分的补偿即阻塞费用。

问题三:先找到各机组在爬坡速率限定下下一时段出力值的变化范围,在此基通过调整预案使输电阻塞完全消除,故建立非线性优化模型,使每条线路上潮流的绝对值超过限值的百分比尽量小。

【关键词】一、问题重述我国电力系统的市场化改革正在积极、稳步地进行。

2003年3月国家电力监管委员会成立,2003年6月该委员会发文列出了组建东北区域电力市场和进行华东区域电力市场试点的时间表,标志着电力市场化改革已经进入实质性阶段。

可以预计,随着我国用电紧张的缓解,电力市场化将进入新一轮的发展,这给有关产业和研究部门带来了可预期的机遇和挑战。

电力从生产到使用的四大环节——发电、输电、配电和用电是瞬间完成的。

我国电力市场初期是发电侧电力市场,采取交易与调度一体化的模式。

电网公司在组织交易、调度和配送时,必须遵循电网“安全第一”的原则,同时要制订一个电力市场交易规则,按照购电费用最小的经济目标来运作。

市场交易-调度中心根据负荷预报和交易规则制订满足电网安全运行的调度计划――各发电机组的出力(发电功率)分配方案;在执行调度计划的过程中,还需实时调度承担AGC(自动发电控制)辅助服务的机组出力,以跟踪电网中实时变化的负荷。

数学建模美赛B题论文

2013建模美赛B题思路数学建模美赛B题论文摘要水资源是极为重要生活资料,同时与政治经济文化的发展密切相关,北京市是世界上水资源严重缺乏的大都市之一。

本文以北京为例,针对影响水资源短缺的因素,通过查找权威数据建立数学模型揭示相关因素与水资源短缺的关系,评价水资源短缺风险并运用模型对水资源短缺问题进行有效调控。

首先,分析水资源量的组成得出影响因素。

主要从水资源总量(供水量)和总用水量(需水量)两方面进行讨论。

影响水资源总量的因素从地表水量,地下水量和污水处理量入手。

影响总用水量的因素从农业用水,工业用水,第三产业及生活用水量入手进行具体分析。

其次,利用查得得北京市2001-2008年水量数据,采用多元线性回归,建立水资源总量与地表水量,地下水量和污水处理量的线性回归方程yˆ=-4.732+2.138x1+0.498x2+0.274x3根据各个因数前的系数的大小,得到风险因子的显著性为rx1>rx2>rx3(x1, x2,x3分别为地表水、地下水、污水处理量)。

再次,利用灰色关联确定农业用水、工业用水、第三产业及生活用水量与总用水量的关联程度ra =0.369852,rb= 0.369167,rc=0.260981。

从而确定其风险显著性为r a>r b>r c。

再再次,由数据利用曲线拟合得到农业、工业及第三产业及生活用水量与年份之间的函数关系,a=0.0019(t-1994)3-0.0383(t-1994)2-0.4332(t-1994)+20.2598;b=0.014(t-1994)2-0.8261t+14.1337;c=0.0383(t-1994)2-0.097(t-1994)+11.2116;D=a+b+c;预测出2009-2012年用水总量。

最后,通过定义缺水程度S=(D-y)/D=1-y/D,计算出1994-2008的缺水程度,绘制出柱状图,划分风险等级。

我们取多年数据进行比较,推测未来四年地表水量和地下水量维持在前八年的平均水平,污水处理量为近三年的平均水平,得出2009-2012年的预测值,并利用回归方程yˆ=-4.732+2.138x1+0.4982x2+0.274x3计算出对应的水资源总量。

2013年数学建模美赛B题论文

2013建模美赛B题思路摘要水资源是极为重要生活资料,同时与政治经济文化的发展密切相关,北京市是世界上水资源严重缺乏的大都市之一。

本文以北京为例,针对影响水资源短缺的因素,通过查找权威数据建立数学模型揭示相关因素与水资源短缺的关系,评价水资源短缺风险并运用模型对水资源短缺问题进行有效调控。

首先,分析水资源量的组成得出影响因素。

主要从水资源总量(供水量)和总用水量(需水量)两方面进行讨论。

影响水资源总量的因素从地表水量,地下水量和污水处理量入手。

影响总用水量的因素从农业用水,工业用水,第三产业及生活用水量入手进行具体分析。

其次,利用查得得北京市2001-2008年水量数据,采用多元线性回归,建立水资源总量与地表水量,地下水量和污水处理量的线性回归方程yˆ=-4.732+2.138x1+0.498x2+0.274x3根据各个因数前的系数的大小,得到风险因子的显著性为r x1>r x2>r x3(x1, x2,x3分别为地表水、地下水、污水处理量)。

再次,利用灰色关联确定农业用水、工业用水、第三产业及生活用水量与总用水量的关联程度r a=0.369852,r b= 0.369167,r c=0.260981。

从而确定其风险显著性为r a>r b>r c。

再再次,由数据利用曲线拟合得到农业、工业及第三产业及生活用水量与年份之间的函数关系,a=0.0019(t-1994)3-0.0383(t-1994)2-0.4332(t-1994)+20.2598;b=0.014(t-1994)2-0.8261t+14.1337;c=0.0383(t-1994)2-0.097(t-1994)+11.2116;D=a+b+c;预测出2009-2012年用水总量。

最后,通过定义缺水程度S=(D-y)/D=1-y/D,计算出1994-2008的缺水程度,绘制出柱状图,划分风险等级。

我们取多年数据进行比较,推测未来四年地表水量和地下水量维持在前八年的平均水平,污水处理量为近三年的平均水平,得出2009-2012年的预测值,并利用回归方程yˆ=-4.732+2.138x1+0.4982x2+0.274x3计算出对应的水资源总量。

高教社杯全国大学生数学建模竞赛题目 穿越沙漠

高教社杯全国大学生数学建模竞赛题目(请先阅读“全国大学生数学建模竞赛论文格式规范”)B 题 穿越沙漠考虑如下的小游戏:玩家凭借一张地图,利用初始资金购买一定数量的水和食物(包括食品和其他日常用品),从起点出发,在沙漠中行走。

途中会遇到不同的天气,也可在矿山、村庄补充资金或资源,目标是在规定时间内到达终点,并保留尽可能多的资金。

游戏的基本规则如下:(1)以天为基本时间单位,游戏的开始时间为第0天,玩家位于起点。

玩家必须在截止日期或之前到达终点,到达终点后该玩家的游戏结束。

(2)穿越沙漠需水和食物两种资源,它们的最小计量单位均为箱。

每天玩家拥有的水和食物质量之和不能超过负重上限。

若未到达终点而水或食物已耗尽,视为游戏失败。

(3)每天的天气为“晴朗”、“高温”、“沙暴”三种状况之一,沙漠中所有区域的天气相同。

(4)每天玩家可从地图中的某个区域到达与之相邻的另一个区域,也可在原地停留。

沙暴日必须在原地停留。

(5)玩家在原地停留一天消耗的资源数量称为基础消耗量,行走一天消耗的资源数量为基础消耗量的2倍。

(6)玩家第0天可在起点处用初始资金以基准价格购买水和食物。

玩家可在起点停留或回到起点,但不能多次在起点购买资源。

玩家到达终点后可退回剩余的水和食物,每箱退回价格为基准价格的一半。

(7)玩家在矿山停留时,可通过挖矿获得资金,挖矿一天获得的资金量称为基础收益。

如果挖矿,消耗的资源数量为基础消耗量的3倍;如果不挖矿,消耗的资源数量为基础消耗量。

到达矿山当天不能挖矿。

沙暴日也可挖矿。

(8)玩家经过或在村庄停留时可用剩余的初始资金或挖矿获得的资金随时购买水和食物,每箱价格为基准价格的2倍。

请根据游戏的不同设定,建立数学模型,解决以下问题。

1. 假设只有一名玩家,在整个游戏时段内每天天气状况事先全部已知,试给出一般情况下玩家的最优策略。

求解附件中的“第一关”和“第二关”,并将相应结果分别填入Result.xlsx 。

2013年全国研究生数学建模竞赛B题论文

参赛密码(由组委会填写)第十届华为杯全国研究生数学建模竞赛学校南京邮电大学参赛队号10293015队员姓名1.仲伟奇2.卢诗尧3.江爱珍参赛密码(由组委会填写)第十届华为杯全国研究生数学建模竞赛题 目 功率放大器非线性特性及预失真建模摘 要:本文根据函数逼近Weierstrass 定理对功放的非线性特性建立多项式数学模型。

对于无记忆功放,直接用matlab 中polyfit 函数或矩阵运算求解,用NMSE 值来评价不同阶数所得的多项式模型,最终将多项式模型的阶数定为4,此时47.13NMS dB E -=,系数详见4.1.3;根据线性原则和两个约束条件建立预失真的多项式模型,采用查表法求得预失真器的输入和输出,建立目标误差函数21ˆmin |()()|Nn GE z n z n ==-∑,用polyfit 函数或矩阵运算求解,最终根据GE 值最小确定多项式阶数为12, 此时-50.877B NMSE d =,系数详见4.2.3。

对于有记忆功放,在无记忆的基础上建立模型,增加延迟项来表征记忆效应,通过矩阵运算求解,然后用NMSE 值评估确定记忆效应多项式阶数为4,记忆深度为3,此时44.3839NMSE dB =-,系数详见4.3.3;根据功放的非线性模型,,建立预失真器的有记忆效应多项式模型,利用功放的输入输出数据间接得到预失真器的输入输出,再用矩阵运算,用NMSE 值来评估确定阶数为4,记忆深度为3,系数详见4.4.3,此时19.0058NMSE dB =-。

运用自相关函数和功率谱密度是一对傅里叶变换对的性质,对自相关函数作傅里叶变换求得功率谱密度,分析得出传输信道范围,最终得出输入信号、有无预失真补偿三类信号的A C P R 值分别为47.1212dB-,37.4586dB -,38.7557dB -,得出预失真补偿后的ACPR 值要比补偿前要小。

关键词:数据拟合 查表法 NMSE/EVM 评价 矩阵运算 多项式模型功率放大器非线性特性及预失真建模一问题重述1.1 问题引入信号的功率放大是电子通信系统的关键功能之一,其实现模块称为功率放大器(PA,Power Amplifier),简称功放。

  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)。

碎纸片的拼接复原摘要本文主要解决碎纸片拼接复原问题。

利用附件所给碎纸片的数据,运用蚁群优化算法、Adaboost算法、Harris角点检测算法,利用Matlab软件编程求解,得到碎纸片拼接复原结果。

针对问题一,依据文字所在行的几何特征,先将文字进行二值化处理,得到文字的数据信息。

运用蚁群优化全局匹配方案完成整体匹配,利用回溯的Best-First搜索算法,得到最佳候选匹配对,由于碎纸片形状相似,Best-First搜索算法会大大降低拼接效率,最后建立蚁群优化算法模型对复原结果进行优化,得到中、英文拼接复原图(见附录一)及顺序表(见表2、表3)。

针对问题二,先对附件3、附件4中的碎纸片进行像素特征分析,将每一个矩形像素特征区域的白色区域设为0、黑色区域设为1,利用Adaboost算法对碎纸片进行分类处理,再依据矩形像素特征进行匹配,得到拼接复原中文、英文图片。

对每次匹配循环进行人工干预得出碎纸片的拼接复原顺序图(见附录二)及顺序表(见表4、表6)。

针对问题三,在对比经典角点检测算法的基础上,利用附件5中图片的信息,运用Harris角点检测的多层匹配图像拼接算法,得到图片的角点信息。

采用标准互相关联法和互信息法对Harris角点进行粗匹配,之后根据特征点周围的边缘信息过滤为匹配点,再用RANSAC进行精确匹配,得到一幅完整的拼接复原图像。

最后,运用神经网络边缘检测算法进行优化,快速的获取准确的碎纸片的拼接复原顺序图(见附录三)及顺序表(见表8、表9)。

关键词:蚁群优化算法 Adaboost算法 Harris角点检测神经网络1 问题重述破碎文件的拼接在司法物证复原、历史文献修复以及军事情报获取等领域都有着重要的应用。

传统上,拼接复原工作需由人工完成,准确率较高,但效率很低。

特别是当碎片数量巨大,人工拼接很难在短时间内完成任务。

随着计算机技术的发展,人们试图开发碎纸片的自动拼接技术,以提高拼接复原效率。

请讨论以下问题:(1)对于给定的来自同一页印刷文字文件的碎纸机破碎纸片(仅纵切),建立碎纸片拼接复原模型和算法,并针对附件1、附件2给出的中、英文各一页文件的碎片数据进行拼接复原。

如果复原过程需要人工干预,请写出干预方式及干预的时间节点。

复原结果以图片形式及表格形式表达(见【结果表达格式说明】)。

(2)对于碎纸机既纵切又横切的情形,请设计碎纸片拼接复原模型和算法,附件3、附件4给出的中、英文各一页文件的碎片数据进行拼接复原。

如果复原并针对附件过程需要人工干预,请写出干预方式及干预的时间节点。

复原结果表达要求同上。

(3)上述所给碎片数据均为单面打印文件,从现实情形出发,还可能有双面打印文件的碎纸片拼接复原问题需要解决。

附件5给出的是一页英文印刷文字双面打印文件的碎片数据。

请尝试设计相应的碎纸片拼接复原模型与算法,并就附件5的碎片数据给出拼接复原结果,结果表达要求同上。

2 问题分析2.1问题一的分析若对中、英文各一页文件的纵切碎纸片拼接复原,需考虑复原的碎纸片之间的边界衔接、字迹线条、图片属性、文字的行高、文字行间距及文字断线等信息。

利用碎纸片文字所在行的几何特征,先对文字进行二值化处理。

借助蚁群优化算法的全局匹配方案,完成整体匹配。

再运用回溯Best-First搜索算法,得到最佳候选匹配对。

考虑到碎片形状特征相似度高等原因,不合理的候选匹配对可能大量存在,导致拼接效率大大降低。

再建立蚁群优化算法进行优化,利用Matlab编程得到碎纸片拼接复原结果。

2.2问题二的分析若将中、英文各一页文件的纵切碎纸片拼接复原,利用Adaboost算法,先将附件3、附件4中给出的碎纸片分别进行编号,再将所有待匹配碎纸片进行初始化,然后进行错误率分析,根据误差分析结果来调整权重,得到分类结果。

对所有碎片分类结果进行矩形像素特征分析,最后由矩形像素特征匹配出碎纸片的原图,对于匹配的循环进行人工干预得出碎纸片的排列顺序表。

2.3问题三的分析若对双面打印文件的碎纸片拼接复原,需考虑正反面文字的对应情况,先用人工干预的方式对文件的部分碎纸片拼接复原。

图像拼接复原需考虑图像特征,本文基于碎纸片的角点特征,利用附件5给出的一页英文印刷文字双面打印文件的碎纸片数据,在对比经典角点检测算法的基础上,运用Harris角点检测的多层匹配的图像拼接算法,采用标准互相关联法对Harris角点进行粗匹配,之后根据特征点周围的边缘信息过滤为匹配点,再用RANSAC进行精确匹配,得到一幅完整的拼接复原图像。

出于对结果准确性以及拼接快速性的考虑,运用BP神经网络边缘检测算法将结果进行优化。

最后,利用数学期望、方差等统计特性参数对检测出来的边缘图像进行质量评价。

3模型假设(1)假设碎纸片在复原过程中没有损坏、丢失的现象;(2)假设碎纸机横、纵切割文件规则,不出现异常状况;(3)假设模型计算的碎纸片拼接复原结果是唯一的;(4)假设附件所给中、英文文件碎纸片数据不混淆;(5)假设同一附件中纸片切割的大小、形状相同;4符号说明5 模型建立与求解5.1模型一的建立与求解5.1.1模型一的建立常规文档碎纸片计算机拼接方法一般利用碎纸片边缘的尖点特征、尖角特征、面积特征,搜索与之匹配的相邻碎纸片进行拼接,这种基于边界几何特征的拼接方法并不适用于边缘性状相似的碎纸片。

题中给定的来自同一页印刷文字文件的碎纸机破碎纸片(仅纵切)的形状相似,拼接时如果只利用碎纸片的边界特征,拼接效果并不理想。

对边缘相似的碎纸片的拼接,理想的计算机拼接过程应与人工拼接过程类似,即拼接时不但要考虑待拼接碎纸片边缘是否匹配,还要判断碎片内的字迹断线或碎片内的文字内容是否匹配以及理解碎片内文字图像含义。

利用现有的技术,完全可以获取碎纸片文字所在行的几何特征信息,比如文字行的行高、文字行的间距及文字断线等信息。

现有技术对碎纸片拼接复原的流程图,如图一所示:图一碎纸片拼接复原流程图二维碎片自动拼接问题主要有两种方案:基于轮廓的拼接和基于内容的拼接。

本文利用文字的特征运用蚁群优化的全局匹配对碎纸片自动拼接。

正确合理参与拼接的碎片之间有一定的空间约束,不合理的拼接则可能会不满足这些约束而在候选匹配队之间造成矛盾。

全局匹配是指在局部拼接的基础上,利用候选匹配队之间的矛盾剔除一些不合理的匹配,从而构建一个彼此两两相容的候选匹配对集合,从而完成整体的匹配。

这里提到的不合理,是从全局的角度考虑的。

初步全局匹配完成后,用回溯的Best-First搜索算法将局部匹配得到的候选匹配对按照相似性度量进行排序,找出最佳候选匹配对。

再用蚁群优化算法进行优化。

回溯的Best-First搜索算法流程图如图二所示:图二 基于回溯的Best-First 搜索算法在实际的碎片拼接过程中,由于碎片形状特征相似等原因,不合理的候选匹配对大量存在。

因此回溯会大大降低拼接的效率。

本文再用蚁群优化算法进行优化求得结果。

◆ 基于蚁群优化的全局匹配如果完成了局部匹配,则构建了这样一个候选匹配对集合。

但对于一组碎片的全局匹配来说,用于全局匹配的候选匹配对只需要一部分的正确候选匹配对,而不是全部。

可以采用蚁群优化形成一个匹配路径,来选取用于全局匹配的候选匹配对。

这个路径所含的候选匹配对刚好包含多个碎片。

当所考虑的问题服从问题相关约束时,可用成分的集合的一个子集来表示,这类问题被称作子集问题。

这里所求的匹配路径上的候选匹配对是候选匹配对集合的一个子集,因此,全局匹配也可以看作有特殊约束的子集问题。

从另一个角度,全局匹配也可以看做是一个求解整体没有矛盾的最小拼接代价的最有问题。

所以,全局匹配路径的构建,可以借鉴ACO 算法用于解决自己问题的方法。

设M 为候选匹配对集合,i M 表示其中一个候选匹配对。

i c 表示i M 的匹配代价,i y 表示在路径构建中i M 被选中的状态,如果i M 被选中,那么i y 的值为1,否则为0:2(1)/2l n C l l ==-表示候选匹配对的个数,l 表示碎片的个数。

则基于ACO 的全局回复重建问题定义如下:1min ()ni i i f y c y ==∑,{0,1}i y =,( 1,2....i n =)被选中的候选匹配对之间必须是相容的。

◆ 图的构建以及数据初始化在构建图(,)c G C L =中,成分集合C 对应着候选匹配对的集合,而连接集合L 不完全连接候选匹配对,与一个相容性列表相关联。

用ij λ表示候选匹配对i M 和j M 之间的相容系数,i S 表示候选匹配对i M 的局部相似性度量,i R 表示i M 的信息素信息素初始化如下:minmax min0.10.8i i S S R S S -=+⨯-其中min S 表示所选匹配对中相似性度量的最小值,max S 为最大值[0.1,0.9]i R ∈。

候选匹配对的匹配代价也是同相似性度量相关联的,ACO 算法中的路程代价通常表示成整数所以本文中的匹配代价也采用类似表示:minmax min9080i i S S c S S -=-⨯- [10,90]i c ∈◆ 路径构建与信息素更新人工蚂蚁根据信息素的浓度的大小,在路径中节点的领域并集中选择下一个访的节点。

当前位于节点i 的人工蚂蚁k 选择j 作为下一个访问节点的概率是:ki i ijk iji ijl N R P R ηη∈=∑k i j N ∈其中: 1i k k il l N G ==,max{}ij i j c c αη=k l G 是人工蚂蚁k 访问过的节点l 的领域,α是一个调节启发式比重的参数,,i c 和jc 是节点i 和j 的匹配代价。

每次路径构建完成后,人工蚂蚁只在当前最优路径的节点上释放信息素。

释放信息素的多少可用下式表示。

mini i a R R c ←+max i R R ← max i ifR R >min c i T ∈其中,a 为信息素释放比例系数,min c 为当前最优路径代价,max R 信息素最大值,min c T 为当今最优路径。

路径中的候选匹配对必须是彼此两两相容的,而矛盾的普遍存在导致很多人工蚂蚁构建路径失败。

路径构建失败必然产生很多不完全路径, 对这些路径上的节点进行信息素更新, 以此来实现信息的蒸发:(1)j j R R ρ←-min max ,j j R R ifR R ←<min c j T ∈其中,ρ为信息素蒸发系数,满足(0,1)ρ∈,min R 为信息素的最小值。

路径构建过程中,下一节点的选取都是在已选节点的邻域中进行的,如果以构建路径中候选匹配对已包含所有碎片,则构建路径成功。

每迭代一次,信息素列表都更新一次如果迭代次数超过迭代次数下限min k , 并且信息素列表的标准差与上一次迭代相比没有变化, 则停止迭代,或者迭代次数已经达到 迭代上限max k, 则停止迭代。

相关文档
最新文档