离散粒子群算法的波纹管优化设计研究

为了解决波纹管所涉及的非线性约束离散变量的优化设计问题,文中尝试将粒子群算法与惩罚函数法相结合,建立新型离散粒子群算法,实现离散变量与连续变量之间的转化。提出了一种新的离散惩罚因子更新策略,以保证离散解的精度及算法的收敛性。通过著名的容器设计算例验证,文中方法优于文献方法。用该方法对波纹管工程实例进行优化设计,优化目标值比在用产品提高了79.96%,且与理论值接近,离散解的精度满足要求,进一步证明了该方法在求解波纹管工程非线性约束离散变量优化设计问题时的有效性。

Study on bellow optimization design based on Discrete Particle Swarm Optimization

Bellow optimization design is a problem of nonlinear constraints discrete design variables.A new Discrete Particle Swarm Optimization was proposed to solve this kind of problem,in which the penalty function was employed to transform discrete design variables to continuous design variables;a new updating scheme of discrete penalty factor was proposed to improve the performance of approach and accuracy of optimum solution.The validity of proposed approach was examined by the famous benchmark——vessel design,the result of proposed approach is superior to those of other approaches from literatures.The proposed approach was applied in the engineering example of bellow optimum design,the discrete global optimum was obtained,and the optimal objective function value was 79.96% higher than that of products in-service.Therefore,the proposed approach can be successfully used to solve the problem of bellow optimum design which is problem of nonlinear constraints discrete design variables.