An Effective Optimization Technique for the Design of Sewer Systems

A mathematical model for the optimal design of sewer systems is presented. Pipe sizes, buried depth and the number and the locations of on-line pumping stations are optimised. Discrete differential dynamic programming (DDDP), based on the conventional dynamic programming (DP), is introduced and modified to solve the optimization model. Some useful guide on how to achieve a good optimum by DDDP is provided. A comparison between DDDP and DP is carried out through a field study which shows that DDDP has big advantages over DP in terms of computation efficiency while the global optimum could still be achieved by DDDP if the feasible domain of state and decision variables is appropriately discretized.