基于改进混沌粒子群算法的管网优化

Pipe Network Optimization Based on Improved Chaotic Particle Swarm Optimization

针对混沌粒子群运算初期的无目的性、固定的控制参量在运算后期不利于跳出局部最优等进行了分析和改进,利用混沌运动的随机性、遍历性和规律性特点对粒子群体中的最优粒子进行混沌寻优,根据给水管网管径选取的离散化特殊性,对混沌粒子群的混沌参量μ进行公式化规定,提出一种改进型混沌粒子群算法(HMCPSO),提高粒子群算法摆脱局部极优的能力。通过引入混沌算法启动机制,有效提高种群初期的收敛能力,通过引入粒子群位置的历史数据判断机制,减少多余的混沌运算,有效缩短算法运行时间。将本改进算法应用于给水管网的模型中,仿真效果表明文中提出的改进算法与PSO和CPSO算法相比,找到的结果更优且稳定性较好,运算时间得到有效减少。

This paper analyses and improves the no purpose in initial and ineffective for jumping out of the local optimum in the late on Chaotic Particle Swarm Optimization（ CPSO）, based on the randomness, stochastic property and regularity of chaos to find a new superi- or individual by chaotic searching on the global best individual. Defined chaos parameters/x according to the particularity of the water distribution network diameter, introduces an improved method of CPSO （HMCPSO）, and improves the ability for seeking the global ex- cellent result. By introducing the mechanism of startup of CPSO, the convergence ability in early stage is improved effectively. By introducing the mechanism of.judge the historical data of particle position, reduces the excess operations of chaos, the calculation time is re- duced effectively. HMCPSO is then applied in the pipe network; comparing with PSO and CPSO, the result shows HMCPSO has higher speed and more stable outcomes.