非稳态罚函数遗传算法及其用于机械/结构系统的健壮性设计

Robust Design for Mechanical/ Structural System by Using Non Stationary Penalty Function with GA

在机械 /结构的优化设计中 ,普遍存在约束的作用 ,且最优解往往位于可行域的边界上。由于外界环境的变化或人为因素造成设计变量扰动 ,可能使设计成为不可行。本文提出了一种基于设计变量敏感性的健壮性设计方法 ,并用非稳态罚函数遗传算法来实现多目标优化设计。

There exist constraints in optimization design of mechanical/structural system, which always makes the optimal points on the boundary of the feasible region.Small disturbances of design variable are likely to make the design infeasible and useless. A general method for robust design is presented based on sensitivity of objective and constraint functions of the design parameters. Genetic algorithm(GA) with Non stationary penalty function is used to solve the multiobjective optimization problem with constraints.