Hanging Gardens Algorithm to Generate Decentralized Layouts for the Optimization of Urban Drainage Systems

Abstract

Traditional urban drainage systems rely heavily on centralized network-based infrastructures. Recently, the idea of centralized urban drainage networks has increasingly been questioned. The latest investigations suggest a transition from centralized to decentralized or hybrid schemes. Therefore, there is a need for tools and methodologies to evaluate and optimize drainage networks with arbitrary degrees of centralization (DC). For this purpose, this paper introduces an algorithm called the hanging gardens algorithm to generate all possible sewer layouts and to explore different degrees of decentralization. The proposed algorithm starts with generating a centralized layout and introducing a list of outlet candidates. Next, it adds arbitrary outlets from candidates to the generated layout and uses a graph theory–based approach to assign parts of the layout to different outlets. This procedure is iterated until all (combinations) outlet candidates have been included. To demonstrate the performance of the proposed algorithm in enumerating all different DC and generating realistic layouts, the algorithm is coupled with an optimization engine in order to optimize the stormwater collection network of a section of the city of Ahvaz, Iran. The number and location of outlets, the layout configuration of each part and the size of pipes are used as optimization variables to minimize costs subject to hydraulic and feasibility constraints. The proposed algorithm performs well in exploring different DC and finding near-optimum solutions.