可靠性多目标优化设计的自适应行为博弈算法

To solve the problem of reliability multi-objective design optimization, an adaptive behavior-based game algorithm was proposed by considering the dynamic characteristics of players' behaviors in the game process. Each design objective was regarded as a player, and the random design variable set was decomposed into the strategy sets of players through the fuzzy clustering analysis for the influence factors of random design variables about the objective functions. The adaptive behavior rule was established. Then, the behavior and the mapping relationship between the objective function and the payoff function of each player were adjusted before the new round of game based on the rule. After multi-round of games, the equilibrium solution was obtained based on the convergence criterion. The design results of the automobile brake indicated that the proposed algorithm had the higher calculation efficiency by comparing with the competitive game and cooperative game.