混合模式搜索差分进化算法研究及其在高压容器概率优化中的应用

为求解高压容器的概率优化问题,提升差分进化算法的收敛速度,提出一种基于模式搜索算法的混合差分进化算法。该算法在差分进化算法的后期,引入模式搜索算子,以当前群体最优点作为模式搜索算子的基点进行模式搜索。仿真结果表明该混合算法收敛速度明显要优于差分进化算法。最后建立高压容器的概率优化设计模型,利用混合算法求解高压容器优化问题的确定型优化模型及概率优化模型,结果表明概率优化模型的优化结果更合理可靠。

In order to solve the probabilistic optimization problem of high pressure vessel and improve the convergence speed of differential evolution algorithm,a hybrid differential evolution algorithm based on pattern search algorithm is proposed.In the later stage of the differential evolution algorithm,the pattern search operator is introduced,and the best point of the current population is used as the base point of the pattern search operator to search the pattern.The simulation results show that the convergence speed of the hybrid algorithm is obviously better than that of the differential evolution algorithm.Finally,the probabilistic optimization design model of high pressure vessel is established,and the deterministic optimization model and probabilistic optimization model of high pressure vessel optimization problem are solved by hybrid algorithm.The results show that the optimization result of probabilistic optimization model is more reasonable and reliable.