LARGE-SCALE OPTIMIZATION MODELS FOR GROUNDWATER MANAGEMENT.

For optimal management it is necessary that the decision maker be able to predict with reasonable accuracy the responses of the aquifer system to decisions affecting recharge and withdrawal. Simulation models have been developed for the major aquifers in Texas. Simulation alone, however, may not yield an optimal solution, since this depends upon a particular set of system variables chosen by the analyst. The best answer obtained after repeated experimentation cannot be guaranteed to be optimum for the model chosen. Additionally, the repeated experimentation to determine the best management plan for the aquifer can be quite time consuming and very expensive. A promising approach would be to develop an optimization model which could incorporate the simulation model and data bases that are already established. Such an optimization model could be very helpful in determining optimal recharge and pumping rates. The purpose of this paper is to describe such a model.