Integrating Iterative Crossover Capability in Orthogonal Neighborhoods for Scheduling Resource-Constrained Projects

Abstract

An effective hybrid evolutionary search method is presented which integrates a genetic algorithm with a local search. Whereas its genetic algorithm improves the solutions obtained by its local search, its local search component utilizes a synergy between two neighborhood schemes in diversifying the pool used by the genetic algorithm. Through the integration of these two searches, the crossover operators further enhance the solutions that are initially local optimal for both neighborhood schemes; and the employed local search provides fresh solutions for the pool whenever needed. The joint endeavor of its local search mechanism and its genetic algorithm component has made the method both robust and effective. The local search component examines unvisited regions of search space and consequently diversifies the search; and the genetic algorithm component recombines essential pieces of information existing in several high-quality solutions and intensifies the search. It is through striking such a balance between diversification and intensification that the method exploits the structure of search space and produces superb solutions. The method has been implemented as a procedure for the resource-constrained project scheduling problem. The computational experiments on 2,040 benchmark instances indicate that the procedure is very effective.