Reservoir management via reliability programing

Abstract

The possibility of applying reliability programing for determining the monthly contract volumes to be released by a reservoir is discussed in this paper. While in standard chance constraint programing the risk of the solution is fixed a priori, here constraint reliabilities are considered as extra decision variables. Hence the optimal operation results from a compromise between profit and risk. It is shown that under quite general assumptions the reliability program enjoys a concavity property that allows the application of an efficient solution algorithm.