适用于环形供水网络优化设计的混合智能算法

摘要：本文设计了一种混合智能算法来解决环形供水网络的优化问题。基于概率学习法和种群的增量学习法（PBIL）与粒子群优化算法（PSO）相结合，PBIL算法中的概率矩阵用PSO算法中的快速升级策略来改进。通过结合两种算法，能提高搜索能力，将供水网络问题用程序代码完整地描述出来。同时，通过调整信息熵值来减少算法的复杂性，并且估计其收敛趋势。本文将修改了的智能升级算法用在实际工程中，并与遗传算法进行了比较，分析其适应性，有效性和稳定性。研究表明，两者的结合能大大提高效率，并且提高全球搜索能力。该算法更适合解决流量分配和供水泵站组合等问题。

Hybrid Intelligent Algorithm for Optimization Design of Annular Water Supply Network

Authors: Wan Shanshan, Sun Lei

Keywords: PSO; intelligent algorithm; optimization; annular water supplu network

Abstract: One hybrid intelligent algorithm is designed to solve the annular water supply network optimization. The model to minimize the objective function of the annual reduced cost with the constraints of hydraulic conditions. The intelligent optimization algorithm Population Based Incremental Learning -PBIL based on probability learning strategy is combined to Particle Swarm Optimization algorithm-PSO. The probability matrix in PBIL algorithm is modified with the quick velocity update strategy of PSO algorithm. An integer encoding representation is given according to the water supply network problem. Also, information entropy is adopted to reduce the algorithm’s complexity and estimate the convergent tendency. The modified intelligent evolutionary algorithm is tested on engineering projects and compared with Genetic Algorithm. The good adaptability, validity and stability performance are fully shown by the results.