数据挖掘在计算多目标博弈策略空间的运用——以四杆桁架设计为例  

The Use of Data Mining Techniques in Computing Strategy Spaces for Multi-Objective Games——Taking the Four-bar Truss Design As an Example

作　　者：韩晓晶 HAN Xiaojing(Ma'anshan Technical College,Ma'anshan Anhui 243000,China)

机构地区：[1]马鞍山职业技术学院,安徽马鞍山243000

出　　处：《佳木斯大学学报(自然科学版)》2025年第1期46-49,共4页Journal of Jiamusi University:Natural Science Edition

基　　金：2023年安徽省职业与成人教育课题(AZCJ2023124);2023年度安徽省质量工程(2023kcsz045)。

摘　　要：博弈方法在多目标优化问题中提供了高效且准确的最优解搜寻手段,但是在模型逐渐复杂的情况下,博弈空间的划分变得越来越困难,这直接影响了结果生成的效率和可信度。主要研究了基于数据挖掘技术的博弈方法在多目标优化问题中的技术路线和实际应用。首先,介绍了将多目标优化问题转化为博弈问题的基本框架.进一步地,通过改进的K-means聚类算法将设计变量分类对应到相应的策略空间。最后,通过四杆桁架结构优化设计问题,验证了该方法的有效性和计算效率。展示了数据挖掘技术在博弈优化中的应用前景,为优化问题提供了全新的视角。Game methods provide an efficient and accurate means of searching for optimal solutions in multi-objective optimization problems,but the division of the game space becomes more and more difficult under the gradual complexity of the model,which directly affects the efficiency and credibility of the result generation.In this paper,we mainly study the technical route and practical application of the game method based on data mining technology in multi-objective optimization problems.Firstly,the basic framework of converting multi-objective optimization(MOO)problem into a game problem is introduced.Further,the design variables are categorized to correspond to the corresponding strategy spaces by an improved K-means clustering algorithm.Finally,the effectiveness and computational efficiency of the method are verified by the four-bar truss structure optimization design problem.This paper shows the prospect of the application of data mining technology in game optimization,which provides a new perspective on the optimization problem.

关 键 词：K-MEANS聚类 博弈论 多目标优化 数据挖掘 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]