A hybrid model for irrigation planning using chance constrained programming and hydrologic simulation

Abstract

Abstract An irrigation planning model is proposed that builds from the water balance of a generalized irrigation system to examine the economic implications of alternative allocations of surface and ground waters. To reflect the stochastic nature of the planning problem, decision rules for canal diversions and well pumpages in each period are considered functions of (stochastic) rainfall and streamflow. The formulation of reasonable decision rules is discussed and a model incorporating stochastic demands and stochastic supplies is constructed. A series of more complex decision rules is tested. For the more complex rules, at each iteration of the algorithm a separation is made in a ‘ master ’ linear programme of the effects of stochastic supplies from those of stochastic demands. The algorithm utilizes a Monte Carlo procedure to adjust the ‘ supply and demand effects’ at each iteration in accordance with the desired probabilistic relations and the optimal solutions to the previous iteration's dual problem. Results are presented for a proposed irrigation scheme in Bangladesh.