模拟动物行为的多目标可靠性优化设计博弈算法

By reforming players' behaviors based on animal behavior, a simulating animal behavior-based game algorithm was proposed to solve the problem of multi-objective reliability design optimization. Every objective was regarded as a player, and the design vector was decomposed into players' strategy sets through random vector mapping technology. Selfish and mutually beneficial behaviors of animals were conferred to corresponding players based on the behavior matching rule. Then, the mapping relation between the payoff function and the objective function of every player was constructed through the simulated animal behaviors. The players that simulated mutually beneficial behaviors leagued together, and competed with other players that simulated selfish behaviors. The equilibrium solution was obtained after multi-round games. The design results of the vehicle brake show that the proposed algorithm is more efficient than competitive and cooperative games.