基于混合行为博弈的多目标仿生设计方法
Multi-Objective Game Design of Bionics Based on Mixed-Behavior
-
摘要:将多个设计目标视为不同的博弈方,通过计算设计变量对目标函数的影响因子和模糊聚类,将设计变量集合分割为各博弈方拥有的策略空间. 对蜥蜴种群的繁衍生存机理进行仿生,将3种蜥蜴的行为方式分别定义为利己主义、集体主义和投机主义,并赋予相应的博弈方,各博弈方根据所仿生蜥蜴的行为特点,建立自身博弈得益函数与目标函数之间的映射关系.各博弈方分别以自身博弈得益函数为目标,在各自的策略空间中进行单目标优化,获得本博弈方对其余博弈方的最佳对策,所有博弈方的最佳对策形成一轮博弈的策略组合,并根据收敛判别,通过多轮博弈,获得最终的博弈解. 以白鹤滩拱坝体型的三目标优化设计为例,设计结果显示坝体体积方量减少了16.412万方,或2.38%; 最大主拉应力降低了0.036MPa, 或0.31%; 整体应变能下降了0.167GJ 或4.47% 体现了基于混合行为博弈方式的多目标仿生设计方法的有效性.Abstract:Taking the design objectives as different game playersand calculating the factors of design variable to objective function andfuzzy clustering the design variable set is divided into the strategicspace owned by each game player. According to the characteristics of threekinds of lizard's behavior style they are defined as egoism collectivismand opportunism and endowed with each game player. For the bionic researchon lizard's survival of multiply mechanism this paper takes three kinds oflizard's behavior as game player and establishes the mapping relationshipbetween game players' payoff and objective functions. In their own strategicspace each game player takes their payoff as mono-objective foroptimization. It gives the best strategy upon other players. All the beststrategies are combined to be a game strategic set. With convergence andmulti-gaming the final game solution is obtained. Taking tri-objectiveoptimization of Baihetan arch dam shape design as example the results showthat the dam volume decreased by 16.412\times 10^4m3 which is2.38%; and the maximum tensile principal stress decreased by 0.036MPawhich is 0.31%; the strain energy of dam decreased by 0.167GJ which is4.47%; these prove that the method is effective.