摘要

Traditionally, the model of a resource-constrained project-scheduling problem (RCPSP) contains a single objective function of either minimizing project makespan or maximizing project net present value (NPV). In order to be more realistic, in this paper, two multi-objective meta-heuristic algorithms of multi-population and two-phase sub-population genetic algorithms are proposed to find Pareto front solutions that minimize the project makespan and maximize the project NPV of a RCPSP, simultaneously. Based on standard test problems constructed by the RanGen project generator, a comprehensive computational experiment is performed, where response surface methodology is employed to tune the parameters of the algorithms. The metaheuristics are computationally compared, the results are analyzed, and conclusions are given at the end.

  • 出版日期2013-10