AWWA ACE61876

This study proposes a groundwater management model that can determine the optimal network and pumping rates of the pumping wells by simultaneously considering fixed and operating costs. The proposed model involves a novel hybrid algorithm that combines the Genetic Algorithm (GA) and Constrained Differential Dynamic Programming (CDDP), a type of optimal control theorem. The main part of the hybrid algorithm is the GA, in which each chromosome represents a potential network design of the pumping wells. The CDDP is used to compute the optimal pumping rates and operating costs associated with each pumping network. A two dimensional finite element model is embedded in the CDDP to simulate the groundwater flow. Hypothetical cases are considered to demonstrate the computational capability. The results of the simulation reveal that the proposed algorithm can solve the complex problem using affordable computation resources. Comparing the management strategy obtained by the proposed algorithm with that computed using an algorithm which considers only the operating cost indicates that the proposed algorithm significantly reduces the costs of implementing management strategy. Accordingly, the proposed model can be used to obtain a cost-effective strategy, including network design and pumping policy, for managing a groundwater system. Includes 31 references, tables, figures.