Document Type

Report

Publication Date

January 1976

Abstract

The feasibility of optimizing large regional water resource planning problems by means of integer programming algorithms is analyzed. Two types of integer programming models are developed: (1) A water supply model including 23 separate but geographically related community systems; and (2) a river basin water quality model including 15 point sources of wastewater, 4 types of pollutants, 6 surveillance points, and 7 alternative treatment processes. The water supply model was structured as a mixed integer problem (some continuous variables included) while the water quality model was an all integer problem. Four integer programming algorithms were tested on the sample problems as follows: (1) MXINT – The Burroughs B6700 TEMPO package algorithm; (2) FMPS-MIP – The UNIVAC 1108 MPS package algorithm; (3) GMINT – A proprietary algorithm authored by A. M. Geoggrion and R.D. McBride; and ($) AIP – A 0,1 algorithm which uses the Balas additive concept. Several versions (sizes) of both problems were successfully solved by one or more of the algorithms with computational efforts ranging from less than 1 to more than 40 minutes of CPU time.