混合多目标骨干粒子群优化算法在污水处理过程优化控制中的应用

Optimal control of wastewater treatment process using hybrid multi-objective barebones particle swarm optimization algorithm

通过对污水生化处理过程的分析,选取能耗和罚款最低为优化目标,建立污水生化处理过程多目标优化控制模型。为了提高Pareto最优解集的收敛性和多样性,提出一种基于Pareto支配和分解的混合多目标骨干粒子群优化算法(HBBMOPSO)。该方法采用带自适应惩罚因子的分解方法选取个体引导者,采用Pareto支配和拥挤距离法维护外部档案和选取全局引导者。此外,采用精英学习策略增强粒子跳出局部Pareto前沿的能力。最后,将HBBMOPSO与自组织模糊神经网络预测模型和自组织控制器相结合,实现污水生化处理过程溶解氧和硝态氮设定值的动态寻优、智能决策和底层跟踪控制。利用国际基准仿真平台BSM1进行实验验证,结果表明所提HBBMOPSO方法在保证出水水质参数达标的前提下,能够有效降低污水处理过程的能耗。

Through analysis of biological wastewater treatment process（WWTP）, a multi-objective optimal control strategy was developed with targets of minimizing both energy consumption and amercement. A hybrid multi-objective barebones particle swarm optimization（HBBMOPSO） algorithm based on Pareto dominance and decomposition was proposed to improve convergence and diversity of optimized set of Pareto solutions. In HBBMOPSO, selection of personal leaders was determined from self-adaptive penalty factor decomposition while maintenance of external dossiers and selection of global leaders were determined from dominance and crowded distance. Furthermore, elitism learning strategy was adopted to facilitate particle escaping from local Pareto fronts. Finally, HBBMOPSO was combined with self-organizing fuzzy nerve network modeler and controller to realize dynamic optimization, intelligent decision, and background monitoring on dissolved oxygen and nitrate nitrogen in biological WWTP. Experimental study on international standardized simulator platform BSM1 showed that HBBMOPSO method can effectively reduce energy consumption under the premise of ensuring effluent to meet quality standard.