基于改进的自适应粒子群算法的给水管网优化设计

Optimal Design of a Water Supply System Based on Improved Self-adaptive Particle Swarm Algorithm

针对粒子群算法在优化给水管网设计时易陷入局部最优难以寻找到最优解的问题,提出改进的动态自适应粒子群算法(modified dynamically adaptive particle swarm optimization,M-DAPSO).定义趋同因子和参数调整函数,使算法能根据种群内部信息自适应调整参数,提出自适应变异策略增加种群多样性.最后,将M-DAPSO算法应用到Hanoi管网优化中,仿真结果表明:该算法能以最小的计算代价求得最优的工程造价;与其他优化算法相比,M-DAPSO算法具有较强的全局搜索能力和较快的收敛速度.

For the problem of easily getting in the local minimum and difficulty in finding the optimal solution when the water supply system are optimized by the particle swarm optimization （PSO） , the paper proposes a modified dynamically adaptive particle swarm optimization （M-DAPSO）. By defining the convergence factor and parameter adjustment function, the improved algorithm proposes the adaptive mutation strategy to increase the population diversity and adjust its parameters. The algorithm is finally applied to Hanoi network optimization. Result show that it can obtain the optimum cost by the minimum computational cost. Compared with other optimization algorithms, M-DAPSO has stronger ability of searching globally and faster convergence speed.