An Alternative Approach using Genetic Algorithm Based Heuristics for Capacitated Maximal Covering Location Allocation Problem

Abstract

The Maximal Covering Location Problem (MCLP) has the objective of maximizing the total demand volume covered by a facility within a maximum allowable travel distance, S. In this paper, three sets of data 30-node, 324-node and 818-node networks with no existing facilities are analyzed as the capacitated model of MCLP (CMCLP) and first solved by the commercial optimization software, CPLEX. As the software produces results which violate the capacity constraint when the constraints are tight, an alternative approach of Genetic Algorithm based heuristics is used to solve the problem for more competitive results. Combination of the best number of facilities to open and the random node order assignment is used to maximize the percentage of total demand covered. The approach is found to solve all the capacitated MCLP in shorter time and in more promising result compared to CPLEX. Finally, the result of the approach is presented and applied to analyze a “real-world” example on a selected area, Telok Panglima Garang, Selangor which is currently served by 5 public health care facilities. This is to explore the implications of the model as the area already has existing facilities and to provide insight on future decisions that can be made on the expansion and development of current facilities.