论文部分内容阅读
在优化配水管网的管径设计方案时,为了缩短计算时间、简化参数选择过程,需要选取速度快,参数敏感性低的算法。该文比较分析了标准遗传算法、稳态遗传算法和子群遗传算法进行管径优化计算时的求解速度和其速度对参数的敏感性。依据经典的管径优化问题框架,编写测试程序,利用3种算法对纽约隧道管网算例进行了多次管径优化计算。比较了不同的种群大小、重组率和变异率下3种遗传算法获得已知最优解时的水力计算次数和参数局部敏感性。结果表明:子群遗传算法达到已知最优解的平均水力计算次数少,不同的参数选取对其平均水力计算次数的影响较小。该算法求解速度快,参数敏感性低,推荐在管网设计时选用。
In order to shorten the calculation time and simplify the process of parameter selection, we need to select the algorithm with fast speed and low parameter sensitivity when optimizing the pipe diameter design of distribution network. This paper analyzes the speed of the solution and the sensitivity of the speed to the parameters when the standard genetic algorithm, the steady state genetic algorithm and the subgroup genetic algorithm are used to optimize the pipe diameter. According to the classic framework of pipe diameter optimization, a test program was written, and three kinds of algorithms were used to optimize the pipe diameter calculation of New York tunnel network. We compared the number of hydraulic calculations and the local sensitivity of the parameters obtained by the three genetic algorithms with different known population sizes, recombination rates and mutation rates. The results show that the average number of hydraulic calculations of subgroup genetic algorithm to reach the known optimal solution is small, and the influence of different parameters on the average number of hydraulic calculations is small. The algorithm solves fast, the parameter sensitivity is low, it is recommended to choose in the pipe network design.