A hybrid genetic algorithm for resource-constrained multi-project scheduling problem with resource transfer time

The RCMPSPTT (resource-constrained multi-project scheduling problem with resource transfer time) problem usually exists in distributed collaborative manufacturing systems, in which scarce resources are shared by different projects dispersed in distributed physical places. Resources are needed to be transferred among different projects with non-neglectable time. In this paper, a hybrid genetic algorithm is proposed for the considered problem. Besides standard operators, EPS (Elite population based dual Population Structure) and VNS (Variable Neighborhood Search) operators are introduced for both diversification and intensification consideration to improve effectiveness. The EPS keeps the elite solutions found during the search and they are updated using a similarity strategy. The VNS generates new solutions by a proposed local search strategy. Experiments show that 26.1% has been improved on solutions by DGAVNS compared with an existing priority rule based heuristic algorithm.