An interactive sorting method for additive utility functions

In this paper, we consider the problem of placing alternatives that are defined by multiple criteria into preference-ordered categories. We consider a method that estimates an additive utility function and demonstrate that it may misclassify many alternatives even when substantial preference information is obtained from the decision maker (DM) to estimate the function. To resolve this difficulty, we develop an interactive approach. Our approach occasionally requires the DM to place some reference alternatives into categories during the solution process and uses this information to categorize other alternatives. The approach guarantees to place all alternatives correctly for a DM whose preferences are consistent with any additive utility function. We demonstrate that the approach works well using data derived from ranking global MBA programs as well as on several randomly generated problems.     

Selection of a representative value function in robust multiple criteria sorting

We introduce the concept of a representative value function in robust ordinal regression applied to multiple criteria sorting problems. The proposed approach can be seen as an extension of UTADIS^GMS, a new multiple criteria sorting method that aims at assigning actions to p pre-defined and ordered classes. The preference information supplied by the decision maker (DM) is composed of desired assignments of some reference actions to one or several contiguous classes—they are called assignment examples. The robust ordinal regression builds a set of general additive value functions compatible with the assignment examples and results in two assignments: necessary and possible. The necessary assignment specifies the range of classes to which the action can be assigned considering all compatible value functions simultaneously. The possible assignment specifies, in turn, the range of classes to which the action can be assigned considering any compatible value function individually. In this paper, we propose a way of selecting a representative value function among the set of compatible ones. We identify a few targets which build on results of the robust ordinal regression and could be attained by a representative value function. They concern enhancement of differences between possible assignments of two actions. In this way, the selected function highlights the most stable part of the robust sorting, and can be perceived as representative in the sense of robustness preoccupation. We envisage two possible uses of the representative value function in decision support systems. The first one is an explicit exhibition of the function along with the results of the UTADIS^GMS method, in order to help the DM to understand the robust sorting. The other is an autonomous use, in order to supply the DM with sorting obtained by an example-based procedure driven by the chosen function. Three case studies illustrating the use of a representative value function in real-world decision problems are presented. One of those studies is devoted to the comparison of the introduced concept of representativeness with alternative procedures for determining a single value function, which we adapted to sorting problems, because they were originally proposed for ranking problems.     

A PROMETHEE-based classification method using concordance and discordance relations and its application to bankruptcy prediction

Outranking relation theory has been widely used to study pattern classification. Here we propose a classification method with concepts from the flows used in PROMETHEE methods, which are extensively applied in multi-criteria decision aids. PROMETHEE uses a flow, generated on the basis of a preference index and measured by various preference functions for each criterion, to represent the preference intensity for one pattern over another pattern. However, only criteria that are concordant with the preference contribute to a preference index. In the present study, the opinions from discordant criteria are also taken into account. The proposed method newly defines an overall preference index using both concordance and discordance relations for ordinal sorting problems. The final classification decision for a new pattern depends on its net flow. The criteria weights are determined using a genetic-algorithm-based approach. Empirical results obtained for a real-world problem regarding bankruptcy prediction demonstrate that the proposed method performs well compared to other well-known classification methods.     

Interval-valued fuzzy multiple criteria decision-making methods based on dual optimistic/pessimistic estimations in averaging operations

This study aimed to consider the effect of dispositional optimism and pessimism to provide simple and useful decision models and methods for multiple criteria decision analysis within an interval-valued fuzzy environment. Uncertain and imprecise assessment information is usually present in many practical decision-making situations. Interval-valued fuzzy sets are useful for modeling impressions and quantifying the ambiguous nature of subjective judgments in a convenient way. Based on measurement tool estimations defined on interval-valued fuzzy sets, dual optimistic and pessimistic point operators were utilized in this study, and this paper discusses several important properties of optimistic/pessimistic averaging operations. Two algorithmic procedures were developed to address the effects of optimism and pessimism, involving changes in overall judgments and in the separate evaluations of alternatives with respect to each criterion. Furthermore, this study explored the practical problem of medical decision making to demonstrate the feasibility and applicability of the proposed method and to make a comparison with other existing methods. Finally, computational experiments were designed using enormous amounts of simulation data, and a comparative analysis of rank orders yielded by dual optimistic and pessimistic averaging operations was conducted.     

A multi-criterion interaction-oriented model with proportional rule for designing supply chain networks

Supply chain management has offered a way to improve the industrial environment becomes more competitive. While, the commonly seen methodologies may be effective in solving the production–distribution problem only from supplier- or customer-oriented consideration, those cannot present the interactive relationship between upstream and downstream enterprises. In the competitive semiconductor industry environment, considering the viewpoints of the supplier and consumer simultaneously is particularly required, because multiple manufacturing and demanding steps are performed at separate situations, concurrently. In this paper, we propose an interaction-oriented approach, which bases on the analytic hierarchy process (AHP) methodology and proportional rule, to solve the semiconductor distribution problem with multiple quantitative and qualitative criteria. The developed approach gives an expected satisfaction for the all participators of the whole chain while the cooperative information is shared perfectly and effectively. Analysis results demonstrate the proposed methodology is efficient and effective through a real world case study.     

Multiple criteria classification with an application in water resources planning

A new kind of multiple criteria decision aid (MCDA) problem, multiple criteria classification (MCC), is studied in this paper. Traditional classification methods in MCDA focus on sorting alternatives into groups ordered by preference. MCC is the classification of alternatives into nominal groups, structured by the decision maker (DM), who specifies multiple characteristics for each group. Starting with illustrative examples, the features, definition and structures of MCC are presented, emphasizing criterion and alternative flexibility. Then an analysis procedure is proposed to solve MCC problems systematically. Assuming additive value functions, an optimization model with constraints that incorporate various classification strategies is constructed to solve MCC problems. An application of MCC in water resources planning is carried out and some future extensions are suggested.     

Greedy algorithms for a class of knapsack problems with binary weights

In this article we identify a class of two-dimensional knapsack problems with binary weights and related three-criteria unconstrained combinatorial optimization problems that can be solved in polynomial time by greedy algorithms. Starting from the knapsack problem with two equality constraints we show that this problem can be solved efficiently by using an appropriate partitioning of the items with respect to their binary weights. Based on the results for this problem we derive an algorithm for the three-criteria unconstrained combinatorial optimization problem with two binary objectives that explores the connectedness of the set of efficient knapsacks with respect to a combinatorial definition of adjacency. Furthermore, we prove that our approach is asymptotically optimal and provide extensive computational experiments that shows that we can solve the three-criteria problem with up to one million items in less than half an hour. Finally, we derive an efficient algorithm for the two-dimensional knapsack problems with binary constraints that only takes into account the results we obtained for the unconstrained three-criteria problem with binary weights.     

基于主动规则的多用户交互模型

论述了多用户交互的特性,提出了一个多用户交互模型(E*Fb*Ft)nCAn。针对该模型的特点,建立起ECA规则模型,讨论了交互事件、回调事件和复合事件。分析了将(E*Fb*Ft)nCAn模型转换为主动机制系统中的事件、规则的方法。用例表明基于主动机制的(E*Fb*Ft)nCAn模型是有效的。     

基于规则归纳的信息抽取系统实现

面对Web信息的迅猛增长,信息抽取技术非常适合于从大量的文档中抽取需要的事实数据。通过文档对象模型（DOM）解析以及检索、抽取、映射等规则的定义,设计并实现了一种具有规则归纳能力的信息抽取系统,用于Web信息的自动检索。在用于抽取规则归纳的框架下,还重点对用于生成抽取模式的WHISK学习算法进行了实验对比分析,结果表明系统对于单槽和多槽数据都具有不错的归纳学习能力。     

Learning model for adaptive behaviors as an organized group of swarm robots

This paper describes a novel organizational learning model for multiple adaptive robots. In this model, robots acquire their own appropriate functions through local interactions among their neighbors, and get out of deadlock situations without explicit control mechanisms or communication methods. Robots also complete given tasks by forming an organizational structure, and improve their organizational performance. We focus on the emergent processes of collective behaviors in multiple robots, and discuss how to control these behaviors with only local evaluation functions, rather than with a centralized control system. Intensive simulations of truss construction by multiple robots gave the following experimental results: (1) robots in our model acquire their own appropriate functions and get out of deadlock situations without explicit control mechanisms or communication methods; (2) robots form an organizational structure which completes given tasks in fewer steps than are needed with a centralized control mechanism. This work was presented, in part, at the Second International Symposium on Artificial Life and Robotics, Oita, Japan, February 18–20, 1997     

Multiple criteria decision-making for the resolution of conflicting ergonomic knowledge in manual materials handling

The ergonomic design of manual materials handling tasks typically utilizes one or more of three approaches; biomechanical, physiological, and psychophysical. Unfortunately, a comprehensive design is rarely obtained because of the difficulty of resolving the conflicting guidelines and criteria for each of the three approaches. This paper describes a methodology utilizing the multiple criteria decision-making (MCDM) process for resolving this conflict. This approach calculates the ratio of the contribution from each type of stress to an overall stress level as determined by the unacceptability of this stress to the user. An example is presented to demonstrate this technique.     

考虑STDP学习律的树突整合型神经元网络的放电同步

生物神经网络的同步被认为在大脑神经信息的处理过程中发挥了重要作用。本文以树突整合型(DHH)神经元网络为研究对象,在网络中考虑脉冲时间依赖可塑性(STDP)学习律,分析网络连接概率、外部扰动以及STDP学习律对网络同步性的影响。研究结果表明,STDP学习律对网络同步的影响不仅与连接概率有关,也与外部扰动有关。在兴奋性神经元网络中,连接概率增大可以增强网络的同步,外部扰动增大会减弱网络的同步,且在较大的外部扰动下STDP学习律对网络同步特性的增强作用减弱;而在抑制性DHH神经元网络中,连接概率增大,也可以增强网络的同步,且在较大的连接概率下,外部扰动增大,可以增强网络的同步,而STDP学习律对网络同步的增强作用也更加明显。     

基于多智能体强化学习的乳腺癌致病基因预测

通过分析基因突变过程, 提出利用强化学习对癌症患者由正常状态至患病状态的过程进行推断, 发现导致患者死亡的关键基因突变. 首先, 将基因视为智能体, 基于乳腺癌突变数据设计多智能体强化学习环境; 其次, 为保证智能体探索到与专家策略相同的策略和满足更多智能体快速学习, 根据演示学习理论, 分别提出两种多智能体深度Q网络: 基于行为克隆的多智能体深度Q网络和基于预训练记忆的多智能体深度Q网络; 最后, 根据训练得到的多智能体深度Q网络进行基因排序, 实现致病基因预测. 实验结果表明, 提出的多智能体强化学习方法能够挖掘出与乳腺癌发生、发展过程密切相关的致病基因.     

基于强化学习的煤矸石分拣机械臂智能控制算法研究

针对传统煤矸石分拣机械臂控制算法如抓取函数法、基于费拉里法的动态目标抓取算法等依赖于精确的环境模型、且控制过程缺乏自适应性,传统深度确定性策略梯度(DDPG)等智能控制算法存在输出动作过大及稀疏奖励容易被淹没等问题,对传统DDPG算法中的神经网络结构和奖励函数进行了改进,提出了一种适合处理六自由度煤矸石分拣机械臂的基于强化学习的改进DDPG算法。煤矸石进入机械臂工作空间后,改进DDPG算法可根据相应传感器返回的煤矸石位置及机械臂状态进行决策,并向相应运动控制器输出一组关节角状态控制量,根据煤矸石位置及关节角状态控制量控制机械臂运动,使机械臂运动到煤矸石附近,实现煤矸石分拣。仿真实验结果表明:改进DDPG算法相较于传统DDPG算法具有无模型通用性强及在与环境交互中可自适应学习抓取姿态的优势,可率先收敛于探索过程中所遇的最大奖励值,利用改进DDPG算法控制的机械臂所学策略泛化性更好、输出的关节角状态控制量更小、煤矸石分拣效率更高。     

A comparative analysis of score functions for multiple criteria decision making in intuitionistic fuzzy settings

The purpose of this paper was to conduct a comparative study of score functions in multiple criteria decision analysis based on intuitionistic fuzzy sets. The concept of score functions has been conceptualized and widely applied to multi-criteria decision-making problems. There are several types of score functions that can identify the mixed results of positive and negative parts in a bi-dimensional framework of intuitionistic fuzzy sets. Considering various perspectives on score functions, the present study adopts an order of preference based on similarity to the ideal solution as the main structure to estimate the importance of different criteria and compute optimal multi-criteria decisions in intuitionistic fuzzy evaluation settings. An experimental analysis is conducted to examine the relationship between the results yielded by different score functions, considering the average Spearman correlation coefficients and contradiction rates. Furthermore, additional discussions clarify the relative differences in the ranking orders obtained from different combinations of numbers of alternatives and criteria as well as different importance conditions.     

Evaluation criteria for blog design and analysis of causal relationships using factor analysis and DEMATEL

The purpose of this study is to find the key factors influencing blog design, and explore the causal relationships between the criteria for each factor. Since design is a multiple criteria decision-making (MCDM) problem, this study adopts a model which is a hybrid of factor analysis and the Decision Making Trial and Evaluation Laboratory method (DEMATEL). The DEMATEL method is used to simplify and visualize the interrelationships between criteria in making a decision. This study found five core factors that influence blog design: visual clarity, interface and usability, content and searchability, programming, and sociability. In addition, the key criteria for each factor were identified and the impact-relation maps obtained. The results of this study can provide useful guidance to blog designers for developing better blog platforms.     

基于小样本集弱学习规则的KNN分类算法*

KNN及其改进算法使用类标号已知的数据集 对类标号未知的数据集 进行类别标识,如果 中的数据数量过少,将会影响最后的分类精度。基于小样本弱学习规则的KNN分类算法旨在提高基于小样本集的KNN算法的分类精度,它首先对 中的数据对象进行学习,从中选取一些数据,利用学到的标签知识对其进行类别标号,然后将其加入到 中,最后利用扩展后的 对 中的数据对象进行类别标识。通过使用标准数据集的测试发现该算法能够提高KNN的分类精度,取得了较满意的结果。     

一种基于队列机制的神经密码学的新学习规则

针对神经密码学中现有队列机制学习规则安全性不足的问题,提出了一种新的安全性更高的学习规则.该学习规则修改了原学习规则改变权值的幅度,使权值修改的幅度不会因为突触深度的增加而增加,从而避免了因权值修改的幅度过大使神经密码的安全性降低.最后利用概率统计实验对该学习规则在简单攻击和几何攻击下的安全性进行仿真.实验结果证明该学习规则与改进前的学习规则和经典学习规则相比,安全性均有所提高.     

关联学习:关联关系挖掘新视角

关联关系挖掘与发现是大数据挖掘与分析的重要基础,现有的关联关系挖掘方法多是对数据进行统计分析,对未知数据缺少关联判别作用.尝试从学习的角度进行关联关系挖掘,给出了关联学习的形式化定义和相关概念,并根据关联学习定义构建学习数据集.具体地构建了2类关联图像数据集(two class associated image data sets, TAID),利用卷积神经网络提取关联特征,然后分别用softmax函数和K近邻算法判别关联关系,基于此提出3种关联关系判别器:关联图像卷积神经网络判别器(associated image convolutional neural network discriminator, AICNN)、关联图像LeNet判别器(associated image LeNet discriminator, AILeNet)和关联图像K近邻判别器(associated image K-nearest neighbor discriminator, AIKNN).3种关联判别器在TAID数据集上进行测试,AICNN在64×64像素90 000个训练样本上的判别精度达0.821 7,AILeNet在256×256像素22 500个训练样本上的判别精度达0.845 6,AIKNN在256×256像素22 500个训练样本上的判别精度达到0.866 4.这3种关联判别器有效地证明了学习角度挖掘关联关系的可行性.     

基于STDP学习规则的视网膜神经回路的特性

基于视网膜的生理解剖结构,构建了包括视锥细胞、水平细胞、双极细胞、AII无长突细胞、神经节细胞、外侧膝状体核和ON通路与OFF通路的视网膜神经回路模型,并在神经节细胞层和外侧膝状体核层的突触连接中引入STDP(Spike-Timing Dependent Plasticity)学习规则,通过添加单一图形刺激和交替图形刺激,比较神经节细胞和外侧膝状体核的电位发放、发放频率以及两者之间突触权重的变化,研究视网膜神经回路的信息传递特性.结果表明:构建的神经回路模型可有效地将光照强度信息转化为发放时序频率信息,且表现出生物视网膜的信息结构特性;STDP学习规则的引入使得外侧膝状体核层接收了相应的刺激模式并学习记忆了这种模式,且ON通路和OFF通路表现出学习独立性;STDP学习规则可以对交替出现的图形刺激,在突触权重的空间分布上进行叠加,且重叠部分的学习效果更加显著.