提出用遗传一模拟退火算法(GASA)混合优化策略来求解生命线管网的拓扑优化问题.混合优化策略结合了遗传算法的并行搜索机制和模拟退火算法的概率突跳特性,提高了算法的优化性能、参数鲁棒性以及计算效率.数值仿真实验表明了算法的稳定性非常好,首次达到最优值的进化代数,且比单一遗传算法提高了26.5倍.
This paper presents a genetic algorithms-simulated annealing (GASA) hybrid optimization strategy for topology design of lifeline networks. The hybrid strategy oombines the paralld searching structure of genetic algorithms with the probabilistic jumping property of simulated annealing algorithms. As a result, the GASA hybrid optimization strategy is much more efficient and likely to find the global optimum which is proved by numerical examples. Compared with the simple genetic algorithm,the evolutionary generations approach the optimum for the first time and increase by 26.5 times.