USE OF CLOUD COMPUTING FOR OPTIMAL DESIGN OF URBAN WASTEWATER SYSTEMS

摘要

Optimal design and rehabilitation of urban wastewater systems (UWwS) is achieved by combining modelling tools with mathematical optimization in high dimensional decision space. However in case of using complex models the bottleneck is the computational time. Parallel computing requires infrastructure capable to deal with parallelized algorithms. This paper presents the experience of using parallel computing on the Cloud (a set of computers distributed across Internet) for optimal design of integrated UWwSs. Amazon's Elastic Compute Cloud was employed. SWMM software is used to implement the hydraulic model, and NSGA-Ⅱ as the multi-objective optimization algorithm that allows for parallelization. A comparison between virtual and local clusters is presented and the recommendations for practitioners are drawn.