Non-dominated Set of Public Service System Designs using Simulated Annealing Approach

Abstract

Public service system designing is often an important subject of operations researchers' interest due to many applicable solving and modelling approaches. This paper focuses on special subclass of location problems, solving of which leads to construction of a special small set of non-dominated solutions. The necessity of searching for a Pareto front or its approximation arises whenever several quality criteria of the design need to be followed at the same time. The scientific content of this contribution consists in introducing a separate heuristic method producing a good approximation of the Pareto front. The suggested algorithm is based on simulated annealing probabilistic metaheuristic approach. The quality of obtained set of non-dominated solutions is compared to the exact Pareto front making use of real-world benchmarks. The main goal of reported research is to enrich the state-of-the-art solving tools by an efficient heuristic tool for bi-criteria public service system designing.