基于多目标进化算法和SWMM的LID设施空间布局优化研究OACSTPCD

In response to the problem that traditional algorithms cannot meet the performance requirements of low impact development(LID)facility spatial layout optimization models,taking the construction pilot of Xiushan sponge city in Chongqing as an example.The platEMO4.0 platform based on MATLAB software was used to compare and analyze the solution results and performance evaluation indicators of seven multi-objective evolutionary algorithms,including NSGA-Ⅱ,NSGA-Ⅲ,MOEAD,PICEA-g,MOEAPSL,CCMO,and CAMOEA,for the optimization problem of LID facility spatial layout,and the best solution was proposed.The results show that most performance indicators of new algorithms are better than traditional algorithms,with CCMO algorithm having the best diversity and convergence,while MOEAPSL algorithm has the fastest solving speed,strongest search ability,and the highest number of optimal solutions.The CCMO and MOEAPSL algorithms can be used to obtain the Pareto approximate frontier,which is the optimal solution set for the spatial layout of LID facilities,under different rainfall recurrence periods in the study area.The optimal plan with runoff reduction as the control objective has a total runoff control rate of 67.23%～76.70%,a peak flow reduction rate of 66.42%～77.86%,and a construction cost of 203.90～245.23 yuan/m2 for LID under a rainfall return period of 5～100 years.