一种改进的遗传算法及其应用

An Improved Genetic Algorithm and Its Application

遗传算法由于其隐含并行性和全局搜索特性,使其具有其他常规优化算法无法拥有的优点.然而,标准遗传算法存在着收敛速度慢、易”早熟”等缺陷.针对应用标准遗传算法时所存在的局限性,从适应值、交叉和变异算子以及控制参数的选取等多方面进行了遗传算法的改进设计.这种改进的遗传算法可进一步改善算法的搜索能力、搜索效率和收敛性能.最后以(N+M)容错系统的优化模型作为优化目标,得到了费用模型的最优解.计算结果验证了算法的有效性和正确性.

Genetic algorithm has some advantages that other normal methods do not have because of its two characters-implicit parallelism and global searching. But SGA has defects of slow convergence and premature convergence. According to these characteristics, the improved genetic algorithm is designed from several aspects such as fitness, crossover and mutation operator as well as choice of control parameters in this paper. The algorithm can improve its searching space, searching efficiency and convergence performance. Finally, （N＋M） fault-tolerant system is used as the optimization objective, and the optimal solution of cost model is also obtained. Calculation result shows that the algorithm proposed in the paper is valid and correct.