基于博弈论客户需求多目标推理技术的研究

发布时间：2018-01-26 18:53

本文关键词： 动态需求引导多目标推理模型产品配置Nash均衡决策优化　出处：《合肥工业大学》2016年硕士论文　论文类型：学位论文

【摘要】：在个性化产品定制中,大数据背景下,针对客户需求信息的动态、不完备等情况,企业越来越关注对客户需求最大程度的满足,实现产品的最优配置。如何既实现对客户动态需求的实时引导,又能够解决产品配置中客户满意度和企业产品成本冲突问题,是当前企业急需解决的问题。结合计算机技术,对客户需求信息进行处理并采用合理的需求引导机制,然后利用合理的产品配置推理方法,是解决这类问题的重要手段,也是实现个性化定制生产的关键。本文研究了客户需求的相关问题,引入了蚁群算法机制和博弈理论,建立了基于蚁群算法的客户动态需求引导机制和基于Bayes-Nash均衡的产品配置多目标推理模型,对产品配置Nash均衡的求解、配置方案的决策优化进行了深入研究,具体内容如下：1)针对客户需求信息的特征,研究客户需求信息的获取方法；对客户需求进行了规范化处理,将客户需求分为静态需求和动态需求,建立多域间功能层次清晰的需求节点交互关联结构并进行赋权处理,将模糊、不完备的客户需求信息转化为明确、可用的客户需求向量。2)针对不完备的客户动态需求,为了客户需求正确表达及保证产品配置的可行性,引入蚁群算法机制,建立了基于蚁群算法的客户动态需求引导机制。提出产品可配置节点和功能需求节点,研究节点间的映射关系,建立产品可配置模型。研究蚁群算法的行为机制,实现对客户动态需求信息的实时引导。3)针对机械产品配置中客户满意度和企业产品成本冲突问题,引入Bayes-Nash均衡理论,建立产品多域节点之间的均衡模型;以企业和客户作为决策主体,企业产品成本和客户满意度作为博弈的收益函数,在结构、性能和成本域相似度最大值搜索的基础上确定博弈双方的策略集,最后,给出产品配置Nash均衡的检验标准,对产品配置方案进行决策分析。4)提出了产品配置Nash均衡求解方法,改进模拟退火算法,对产品配置多目标推理中Nash均衡点进行实时求解。5)提出基于变型设计决策优化方法,解决产品配置多目标推理中无Nash均衡点的问题从而实现产品配置决策的优化；提出基于Pareto客户需求引导决策优化方法,解决引导过程中空路径的问题从而实现需求引导决策的优化。
[Abstract]:In product customization, under the background of big data according to the customer demand information, dynamic, incomplete, enterprises pay more and more attention to the maximum to meet customer needs, to achieve the optimal allocation of products. How to realize the real-time guidance of customer demand and dynamic, can solve in the configuration of product customer satisfaction and enterprise product cost the conflict is the current urgent. Combined with computer technology, customer demand for information processing and using reasonable demand guide mechanism, product configuration reasoning method and reasonable use, is an important means to solve this problem, it is also the key to customization production. This paper introduces the issues about customer demand the introduction of the ant colony algorithm, mechanism and game theory, established the dynamic customer demand based on ant colony algorithm and guide mechanism based on Bayes-Nash equilibrium production Product configuration multi-objective reasoning model of product configuration is Nash balanced, in-depth research on decision making optimization program, the specific contents are as follows: 1) according to the characteristics of customer demand information acquisition method of the customer demand information; customer demand for the standardization process, customer demand will be divided into static and dynamic demand the establishment of functional requirements, a clear hierarchy of needs interaction structure and node weighted fuzzy, multi domain, customer demand for incomplete information into clear, available customer demand vector.2) according to the customer demand dynamic incomplete, in order to correct customer demand expression and ensure the feasibility of the product configuration, the ant colony algorithm mechanism the establishment of the guide mechanism, dynamic customer demand based on ant colony algorithm is proposed. Product configuration node and the functional requirements of node mapping relationship between nodes, construction Can establish product configuration model. The behavior mechanism of ant colony algorithm, to achieve real-time guide.3 on dynamic customer demand information) for customer satisfaction and product configuration in machinery enterprise product cost conflict, the introduction of Bayes-Nash equilibrium theory, establish the equilibrium model between product multi domain nodes; to the enterprise and the customer as the decision-making body, the enterprise product cost and customer satisfaction as the revenue function, the game in the structure, performance and cost of search domain similarity value is determined based on the both sides of the game strategy set, finally, gives the product configuration of Nash equilibrium test, decision analysis of.4 product configuration) proposed product configuration Nash equilibrium method, the improved simulated annealing algorithm real time.5 to solve the Nash equilibrium point of product configuration reasoning) proposed multi-objective decision optimization method based on variant design, product configuration solution In the multi-objective reasoning, there is no Nash equilibrium point, so as to achieve the optimization of product configuration decision. Based on the Pareto customer needs guidance decision-making optimization method, we solve the problem of the hollow path in the guidance process, so as to achieve the optimization of demand guidance decision-making.

【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：F273;TP18;F224.32

【参考文献】