Chance-constrained water supply operation of reservoirs using cellular automata

Abstract

This paper presents the application of a novel Cellular Automata approach for the efficient and effective solution of chance-constrained water supply reservoir operation problems. The method is based on the observation that a low value of the penalty parameter would lead to partial enforcement of the constraints. The constraints of the operation problem namely operation and reliability constraint, are dealt differently. A high enough value of the penalty parameter is used for the first set while a lower than enough value is used for the second set leading to complete enforcement of the first set of the constraint and partial fulfillment of the second set. Since the proper value of the penalty parameter to be used for the reliability constraints is not known a priori, an adaptive method is, therefore, proposed to find the proper value. For this, the problem is first solved for the optimal operation using a zero value of the penalty parameter. The value of the penalty parameter is then adjusted using the reliability of the optimal operation obtained. At each iteration, the penalty parameter is increased if the current reliability is less than the target reliability and deceased if otherwise. The proposed model is used for the optimal water supply operation of Dez reservoir in Iran over short, medium and long term for different target reliabilities and the results are presented and compared with those of a Genetic algorithm. The proposed model is shown to produce comparable results to the GA for water supply problem with much reduced computational effort.