混合GA与SA求解非线性约束优化

Hybrid GA and SA for Solving Nonlinear Constrained Optimization Problems

在非线性约束优化中,处理好约束条件和增强局部搜索能力是解决这类问题的关键.本文在给出问题一般形式的基础上,设计了一个模拟退火和遗传算法结合的算法.它用模拟退火算法来增强局部搜索能力,用线性交叉来处理约束以外的解,将可行解与不可行解用适应值的正负来区分.仿真试验表明,该算法收敛速度快、搜索能力强、稳健性好,本方法是对应用遗传算法求解非线性约束优化问题的又一次深入探索.

For solving the problems of nonlinear constrained optimization, it is very important to handle the constraints and enhance the ability of searching in the part space. With an ordinary model given for the problems, an algorithm was designed by combining genetic algorithm (GA) with simulated annealing (SA).It enhances the ability of local searching by using Simulated Annealing, deals with the solutions out of constrains by linear crossing, and differentiates feasible solutions from unfeasible solutions by positive or negative fitness. Simulation results show that convergence is fast,searching ability is strong and stability is good. The algorithms designed is a further step forward in research about genetic algorithm's application in solving nonlinear constrained optimization problems. 