Optimal design of multi-storage network for combined sewer overflow management using a diversity-guided, cyclic-networking particle swarm optimizer – A case study in the Gunja subcatchment area, Korea

Multiple small-scale, distributed storage facilities have recently received much attention owing to their effectiveness for combined sewer overflow (CSO) mitigation. In this line of research, designing the optimal configuration of storage tanks in a sewer network is very challenging, and thus relatively few studies have been made to this day. To solve such a large-scale complex multimodal optimal design problem, a meta-heuristic particle swarm optimization-based design methodology of complex sewer networks for CSO management is developed. This search engine includes two mechanisms: a diversity-guided three-phase velocity update law and restricted social best searching based on the cyclic network structure. It allows regions of the design space to be explored efficiently by driving each particle to share information in switching the velocity update mechanism only with a set of neighboring particles via a fixed near-neighbor interaction structure. Therefore, the movement of a particle is no longer driven by the global best position of the entire swarm, which enhances the diversification attitude of the scheme. Its implementability under an actual environment is demonstrated by applying it to a combined sewer network case study of a complex large-scale multi-storage network in the Gunja subcatchment area located in Seoul, Republic of Korea. The simulation results indicate that the developed particle swarm optimization–based design methodology exhibits not only superior reliability but also high practicality, simplicity, and implementability for optimal planning of real-life CSO storage facilities.