基于改进粒子群优化算法的BOD-DO水质模型参数确定

摘要

[Objective] The improved particle swarm optimization algorithm was used to determine parameters of the BOD-DO water quality model.[Method] The population diversity was kept by adding mutation operators of differential evolution algorithms to the individual historical best position and strong local search ability of the simplex algorithm in the later phase of convergence.Then,the improved particle swarm optimization algorithm was established and adopted for determination of parameters for the BODDO water quality model.[Result] The improved particle swarm optimization algorithm can effectively determine the parameters for the BOD-DO water quality model.The wide range of initial values had less effect on the convergence,but larger range required large iteration number.Different initial populations had certain influences on the convergence,and the initial population formed by uniform distribution method effectively improved the convergence rate and speed.The random selection strategy also effectively improved the convergence rate and speed.Results demonstrated that the convergence precision was improved with convergence rate of 100％.The convergence speed of the improved particle swarm optimization algorithm was increased by more than 5 times.The standard deviation was about 10％ of the particle swarm optimization algorithm.[Conclusion] The improved particle swarm optimization algorithm can effectively avoid the premature or stagnation phenomenon of particle swarm optimization algorithm,and it provides a reliable method for different water quality models.%[目的]将改进的粒子群优化算法应用于BOD-DO水质模型参数求解,为水质模型参数求解提供支持.[方法]通过差异演化算法对各个体历史最佳位置进行变异,以保持种群多样性,并在搜索后期加入局部搜索能力强的单纯形算法,建立改进的粒子群优化算法,并用该算法对BOD-DO水质模型参数进行求解.[结果]改进的粒子群优化算法能有效地确定BOD-DO水质模型参数;参数取值范围的放宽对算法的收敛性影响较小,但迭代次数有所增加;均匀分布法生成的初始种群可以有效地提高算法的收敛率,加快收敛速度;交叉概率和缩放因子的随机选取策略,可以有效地提高算法的收敛率并加快收敛速度;比较计算结果可知,改进的粒子群优化算法的收敛精度有所提高,收敛率可达到100％,收敛速度可提高5倍以上,标准差约是粒子群优化算法的10％.[结论]改进的粒子群优化算法有效地避免了原算法的早熟或停滞,为不同类型的水质模型参数求解提供了一个可靠的方法.