Simultaneous Investment and Allocation Decisions Applied to Water Planning

Abstract

This paper develops models for water planning that simultaneously consider investment and allocation decisions. The framework is demonstrated for a region which includes the counties of San Luis Obispo and Santa Barbara in California. The quadratic discontinuous mathematical programming problems-resulting from such models are solved optimally by using generalized Benders decomposition. Benders decomposition is a method for breaking a larger problem into a series of manageable subproblems which yield feasible solutions to the overall system. In this case the decomposition is used to incorporate a quadratic programming allocation subproblem with an integer programming investment sequencing problem.