Robust global and local search approach to resource-constrained project scheduling

• Published in: Canadian journal of civil engineering Vol. 36; no. 3; pp. 375 - 388
• Main Authors: Kim, Jin-Lee, Ellis, Ralph D
• Format: Journal Article
• Language: English
• Published: Ottawa, ON National Research Council of Canada 01.03.2009; NRC Research Press; Canadian Science Publishing NRC Research Press
• Subjects: Algorithms; Applied sciences; Buildings. Public works; calcul de l'évolution; Civil engineering; evolutionary computation; Exact sciences and technology; Experiments; Genetic algorithms; gestion de projet; hybrid methods; Industrial project management; Mathematical optimization; Methods; méthodes hybrids; optimisation; optimization; Project management; Project management. Process of design; resources; ressources; Scheduling; Scheduling (Management); établissement de calendrier; Genetic algorithm; Project management; Evolution; Experimentation; Scheduling; Resource management; Mixed method; Comparative study; Optimization; Recommendation; Construction industry; Result
• ISSN: 0315-1468; 1208-6029
• DOI: 10.1139/L08-135

The resource-constrained project scheduling problem (RCPSP) is one of the most challenging problems in construction scheduling applications, in which optimal solutions are of great value to project planners. This paper presents a new adaptive hybrid genetic algorithm search simulator (AHGASS) for finding an optimal solution to the problem, and provides the strategies and practical procedures to develop the algorithm. Elitist genetic algorithm (EGA) developed is used for the global search, while random walk algorithm for the local search is incorporated into the EGA to overcome the drawbacks of general genetic algorithms, which are computationally intensive and premature convergence to a local solution. Computational experiments are presented to demonstrate the performance and accuracy of AHGASS. The proposed algorithm provides a comparable and competitive performance compared with the existing genetic algorithm (GA) hybrid heuristic methods. The findings demonstrate that AHGASS has significant promise for solving a large-sized RCPSP.