A comparative study of GA, PSO and ACO for solving construction site layout optimization

Abstract

The positioning and layout of facilities on a construction site is important to enhance efficiency, productivity and safety. In this paper, three metaheuristics, Genetic Algorithm (GA), Particle Swarm Optimization (PSO), and Ant Colony Optimization (ACO), are proposed to solve the construction site layout problem in which facilities are positioned to locations so the sum of construction cost and interactive cost due to facility layout constraints is minimized. The craziness concept, cross-mutate and scramble mutation techniques are used to increase the diversity of the solutions and to keep the algorithms from being trapped at local optima. The optimal parameters for each algorithm are determined by using the Design of Experiment approach (DOE). Two case studies of facility layout problem derived from literature were used to rigorously compare the performances of the three algorithms, in terms of effectiveness, efficiency, and consistency. ANOVA test was used to compare the performances. The results demonstrate the capability of the modified method in solving facility layout problems effectively, efficiently and consistently. This study contributes to the decision making when determining an appropriate solution for the construction site layout problem.