Evolutionary algorithms for construction site layout planning

Abstract

The arrangement of temporary facilities within a construction site is essential for successfully undertaking a project, as it enhances productivity and ensures both safety and environmental protection. The Construction Site Layout Planning (CSLP) problem is a challenging discrete combinatorial optimization problem involving multiple objectives and has been tackled using various methods, from linear programming to heuristic and meta-heuristic techniques. Evolutionary algorithms have patently been preferred for solving the CSLP problem due to their ability to provide efficient (near-optimal) solutions in reasonable computational time. The present work aims to comparatively evaluate the effectiveness of five well-known evolutionary algorithms in terms of these performance indicators based on a number of case studies of different structure and characteristics. The model implementation is structured in an Excel environment to facilitate the problem setting and calculations while the optimization algorithms have been implemented in the Matlab software. The examined case studies include simple, single-objective formulations (i.e., minimizing the total traveling distances among facilities) and multi-objective formulations that consider, in addition, preferences or constrains in facility setting to account for operational, safety, and environmental considerations. The evaluation results indicate that all methods perform reasonably well from a practical point of view, however, those based on harmony search, simulated annealing, and particle swarm optimization appear to be more flexible in attaining better solution quality and lower computational time.