An integrated methodology for optimum operation and economic capacity expansion of stochastic multi-facility systems

Abstract

The objective of this research is to develop a computer methodology for optimum operating criteria and economic capacity expansion decisions for large multi-facility systems with stochastic inputs and demands. This work is based on a previous theoretical study by Curry and Helm. The goal is to minimize the total cost of operation and expansion during the entire planning horizon while satisfying the demand constraints, the storage capacity constraints and other constraints characterized by the particular system. The specific systems considered in the present study are a multiple multi-purpose linked reservoir system with stochastic inflows and demands, and a multi-plant, multi-warehouse system with stochastic demands. Based on the chance constrained models for stochastic multi-reservoir system by Curry and Helm, a computational scheme for operational decisions for large systems is developed. The operational problem is a large linear programming problem. The problem is formulated in Fortran and the solution is obtained by MPS/360. Interfacing between the Fortran and MP/360 is accomplished by the use of Read Communications Format (READCOMM). Expansion of the water resources is also considered. The expansion decision problem is formulated in "yes-no" type binary variables with values 1 or 0. Balas' zero-on integer linear programming is applied for its solution. Then the operational and expansion problem is integrated. The integrated model is a mixed continuous-integer programming problem. Bender's portioning procedure is applied for the solution of the integrated model. Since expansion may not take place in all periods of the planning horizon due to financial or construction lag time, a preferred time expansion scheme is incorporated into the solution methodology. The concept of multi-reservoir operation and expansion is extended to multi-plant multi warehouse system. A chance constrained model for operation and expansion of this system is developed. Several example systems are used to demonstrate the methodology at its various stages.