论文部分内容阅读
在分析标准遗传算法的优越性与存在不足的基础上,提出了对遗传算法的改进方法.将能量熵的选择加入到遗传算法的退火选择中,以充分地探索解空间,保持种群的多样性.将伪梯度搜索应用于对个体的邻域搜索,利用当前种群的有效信息及系统信息,提高寻优速度.对典型的TSP问题及一实际电力网络故障恢复的仿真研究表明,改进算法全局优化性能优于启发式遗传算法及标准、退火遗传算法,同时使收敛速度有了较大的提高.
Based on the analysis of the advantages and disadvantages of the standard genetic algorithm, this paper proposes an improved method for genetic algorithms, which adds the choice of energy entropy to the annealing of genetic algorithm to fully explore the solution space and maintain the diversity of the population The pseudo-gradient search is applied to the neighborhood search of individuals and the effective information and system information of the current population are used to improve the optimization speed.The simulation study of a typical TSP problem and an actual power network fault recovery shows that the global optimization of the improved algorithm Performance is better than the heuristic genetic algorithm and standard annealing genetic algorithm, while the convergence rate has been greatly improved.