社会网络二元不对称关系预测中的非独立性检验问题

社会网络中普遍存在着二元不对称关系,在网络动态演化过程中对二元不对称关系预测是基于现有的大量网络结构信息,判定信息的有效性是提高预测效率和预测结果正确率的重要保证.文章针对二元不对称关系预测提出了变量非独立性假设以及检验算法,并通过建立正态性检验算法进一步提高非独立性检验的效率.实例验证表明本文所提出的方法是合理的.     

民主选举中的数学与公正性

首先介绍选举理论中的5种投票方法（简单多数、单轮决胜、系列决胜、Coombs法、Borda计数）和5条公平性准则（多数票、Condorcet获胜者、Condorcet失败者、无关候选人独立性、单调性）,并用政治和社会领域的若干实例给以解释。然后给出著名的Arrow不可能性定理的两种不同的版本,以及对Arrow一条公平性准则的修正;按照修正后的准则,存在满足所有公平性准则的投票方法。     

少数服从多数规则的特性

假定投票人的人数是奇数，E．S．Maskin引入传递性刻画了少数服从多数规则的特性．Campbell和Kelly试图将E．S．Maskin的理论推广到投票人的人数是任意整数的情形，用有限传递性来刻画少数服从多数规则的特性．我们发现存在这样的少数服从多数规则，它满足无关方案独立性条件并且认为任意两个不同的备选方案都是有差异的，但却不一定满足有限传递性．所以，用有限传递性来刻画少数服从多数规则的特性就有一定缺陷．本文对满足无关方案独立性的少数服从多数规则特性进行了重新刻画，并弱化了有关条件．     

非混合单调二元算子方程(组)解的存在唯一性

利用锥理论及Banach压缩映射原理,在不要求上、下解条件及算子紧性与连续性的条件下,建立了一类满足更一般序关系条件的非混合单调二元算子方程组(?)解的存在唯一性定理,以及非单调二元算子方程T(x,x)=x和非单调一元算子方程Lx=x解的存在唯一性定理,推广了最近相关文献的研究结果.     

对具有从属关系的社会选择函数进一步的探讨

分析了Hattori给出的三个条件的合理性，并建立了一个仅满足非限制域条件而不满足存在一个独裁与无关个人影响力独立性条件的具有从属关系的社会选择函数，使具有从属关系的社会选择函数族更加完备。     

群体多目标决策α-联合有效映射的性质

本文首先引入群体多目标决策的α-联合有效映射,然后讨论此映射满足Arrow的有关公理, Sen的强Pareto原则和局部非独裁性条件,以及May的匿名性、中立性以及正响应性条件的情况.     

例说独立性与不相关性的关系——兼说高等数学与概率统计的关系

本文讨论独立性与不相关性的关系,除了随机变量之间的独立性必定导致二个随机变量之间的不相关性外,以几个服从二元二点分布、二元均匀分布、二元指数分布、二元正态分布的二个随机变量之间的独立性与不相关性作为例子:说明独立性与不相关性之间的其他三种情形;不难发现,刻划独立性的是函数关系,刻划不相关性的是数字关系;由此,同时讨论高等数学与概率统计的关系,也列出了高等数学与概率统计之间的四种情形.     

二元Arnold-Strauss型指数分布的条件指数性及渐近独立性

本文讨论二元Arnold-Strauss型指数分布的条件指数性及渐近独立性,证明了给定X关于Y的条件密度和给定Y关于X的条件密度都是指数分布密度,求出了用于预报的条件概率;并证明了X,Y之间的渐近独立性.另外,讨论了它的识别性,若已知可识最小值的分布密度时,所有参数皆可识别.     

二元Lawrance-Lewis型指数分布的识别性及渐近不独立性

本文讨论了二元Lawrance-Lewis指数分布的识别性和渐近不独立性,若已知可识最小值的分布密度时,所有参数皆可识别;另外,还计算了X,Y之间的相关系数和尾部相关系数,证明了X,Y之间的渐近不独立性;且X,Y不相关当且仅当b=2时.     

普通选择函数的若干合理性条件的模糊化研究

以普通选择函数的有关结论为基础,利用模糊逻辑联结运算模糊化了Arrow的五个合理性条件,讨论了模糊化后的这些合理性条件之间的关系,研究了文献中给出的一个开问题,并对模糊情形下不成立的情况提供反例予以说明,从而推广了普通选择函数的相关结论。     

Between arrow and Gibbard-Satterthwaite; A representation theoretic approach

A central theme in social choice theory is that of impossibility theorems, such as Arrow’s theorem [Arr63] and the Gibbard-Satterthwaite theorem [Gib73, Sat75], which state that under certain natural constraints, social choice mechanisms are impossible to construct. In recent years, beginning in Kalai [Kal01], much work has been done in finding robust versions of these theorems, showing “approximate” impossibility remains even when most, but not all, of the constraints are satisfied. We study a spectrum of settings between the case where society chooses a single outcome (à-la-Gibbard-Satterthwaite) and the choice of a complete order (as in Arrow’s theorem). We use algebraic techniques, specifically representation theory of the symmetric group, and also prove robust versions of the theorems that we state. Our relaxations of the constraints involve relaxing of a version of “independence of irrelevant alternatives”, rather than relaxing the demand of a transitive outcome, as is done in most other robustness results.     

Assessing the Level of Compromise in Real Choice Models

We analyze the plurality runoff rule often used in national elections and prove that the corresponding choice function is Borda-rational. The compromise level attained in real voting systems is defined. The corresponding measures are calculated for the case of plurality runoff with three main candidates. We investigate the influence of the axiom of independence of irrelevant alternatives on the rationality type of choice functions realized in multicriterion choice systems.     

Coalitional power structure without the pareto principle

In this paper we have extended Arrow's analysis to a framework where for any given profile of individual preference orderings the decision procedure specifies a non-trivial probability distribution over possible social orderings. We have demonstrated that if the social decision procedure satisfies certain probabilistic versions of weak independence of irrelevant alternatives, then it is characterized by a ‘power’ structure for all possible coalitions of individuals without assuming either the Pareto Principle or its antecedents. A generalised version of Arrow's impossibility theorem follows as a special case of our result. We have weakened Arrow's independence condition, and have shown the existence of a hierarchy of dictators without imposing the Pareto criterion.     

阿罗不可能性定理的几个等价形式

在这篇短中。给出了关于社会福利函数F的防止策略性操纵的概念,并且证明了如果备选对象至少有三个。则下面结论是相互等价的：(1)F满足Pareto与IIA性质;(2)F满足Pareto与RID性质;(3)F是独裁的;(4)F是满的、正向响应的;(5)F是防止策略操纵的且F是满的。     

Binariness and rational choice

The concept of numerical representability of preferences together with maximality is at the heart of the concept of rationality embodied in classical optimization models. The difficulty of representing social preferences arises from inherent intransitivities thrown up by democratic voting procedures and by non-binary choice rules which need to be adopted to cope with these intransitivities. An alternative (weaker) concept of representability is developed and it is shown that this concept can partially accommodate intransitivity and non-binariness. ‘Weak transitivity’ and ‘weak binaries’ conditions are provided and it is shown that these conditions are necessary and sufficient for ‘weak representability’. While the weak transitivity condition would be violated by social aggregation procedures, the non-binary functions used by social choice theorists do indeed satisfy the condition of ‘weak binariness’.     

Symmetric social choices and collective rationality

A social choice function C defined on the m-element subsets of a set X for n voters assigns a non-empty subset C(A,R) of A to each pair (A,R) in which ¦A¦ = m,A X, and R is an n-tuple of voter preference orders on X. When n > 2, 3 m < ¦X¦, and C satisfies natural conditions of independence, symmetry among voters and alternatives, and collective rationality, it is proved that C is completely indecisive in the sense that C(A,R) = A for all (A,R). This non-binary result complements a binary (m = 2) result proved by Hanson. It suggests that there is a fundamental incompatibility between collective rationality and conditions designed to assure equitable treatment of voters and of alternatives.     

Bargaining conventions

This paper describes and defends a weaker version of Nash's axiom of independence of irrelevant alternatives which is called convention consistency.     

Single-plateaued choice

Single-plateaued preferences generalize single-peaked preferences by allowing for multiple best elements. These preferences have played an important role in areas such as voting, strategy-proofness and matching problems. We examine the notion of single-plateauedness in a choice-theoretic setting. Single-plateaued choice is characterized by means of a collinear interval continuity property in the presence of independence of irrelevant alternatives. Further results establish that, our notion of single-plateauedness conforms to the motivation underlying the term and we analyze the consequences of alternative continuity properties. The importance of basic assumptions such as closedness and convexity is discussed.     

The categorical approach to social choice theory

A number of different problems in social choice theory are given a common treatment in the framework of category theory. A social choice function is a map μ from a set Hom(A, PX) of profiles - in applications functions from a set A of agents to the set PX of linear orders on a set X of alternatives - to a set GPX the nature of which depends on the application. The main theorem established is applied to social welfare functions and strategy-proof voting schemes.