基于粒子群优化算法的复杂系统可靠性分配与优化

摘要

In order to obtain the result of complex system reliability optimization allocation speedily under minimal cost restriction. On the basis of analyzing the function of reliability cost and the data of reliability, the model of three parameters reliability cost function was selected, which parameters were obtained easily. Then, the parameter selection methods of particle swarm optimization were studied based on these works. A new method of parameter selection in particle swarm optimization—varying acceleration coefficients was presented, which fits for complex system reliability allocation. Consequently the subjectivity in process of reliability allocation was reduced, and the convergence rate of reliability optimization was speeded. Through instance analyzing and comparing with genetic algorithm, the result proves that this method is efficient and superiority.%为了能快速获得费用最小约束条件下的复杂系统可靠性优化分配结果,在对可靠性成本函数及可靠性数据分析研究的基础上,选择了一种工程上易于获取的三参数可靠性成本函数模型作为可靠性优化分配模型;在此基础上,研究了粒子群优化算法的参数选择方法,提出了一种适合于复杂系统可靠性分配的改进型粒子群优化算法--变加速系数的粒子群优化算法,减少了可靠性分配过程中的主观因素,加快了可靠性优化分配的收敛速度.通过实例分析并与遗传算法进行了比较,证明了该方法的有效性和优越性.