Nonlinear Optimization Model for Screening Multipurpose Reservoir Systems

An optimization model for selecting and sizing potential reservoir and hydropower plant sites on river basins is presented. The model improves on the work of Lall and Miller by replacing the modified sequent peak algorithm for sizing reservoirs with a behavior analysis algorithm that allows operation of the reservoir system with realistic operating policies. A heuristic simulation procedure is used to assess the feasibility of developing self-sustainable hydropower plants at each site. A gradient (derivative) based optimization algorithm is employed to integrate these simulations into an optimal screening model. To enhance the efficiency of the algorithm, the approximate and expensive approach of evaluating derivatives by divided differences is replaced by automatic differentiation. A comparison of the computational performance shows that the optimal solution with automatic differentiation is obtained four times faster than with the divided difference method. The model is developed in the context of river basins in India.