Multiobjective Optimization Using Goal Programming for Industrial Water Network Design

Abstract

The multiobjective optimization (MOO) of industrial water networks through goal programming is studied using a mixed-integer linear programming (MILP) formulation. First, the efficiency of goal programming for solving MOO problems is demonstrated with an introductive mathematical example and then with industrial water and energy networks design problems, formerly tackled in literature with other MOO methods. The first industrial water network case study is composed of 10 processes, 1 contaminant, and 1 water regeneration unit. The second, a more complex real industrial case study, is made of 12 processes, 1 contaminant, 4 water regeneration units, and the addition of temperature requirements for each process, which implies the introduction of energy networks alongside water networks. For MOO purposes, several antagonist objective functions are considered according to the case, such as total freshwater flow rate, number of connections, and total energy consumption. The MOO methodology proposed is demonstrated to be very reliable as an a priori method, by providing Pareto-optimal compromise solutions in significant less time compared to other traditional methods for MOO.