| Summary: | In order to solve project schedule delays caused by resource conflicts and uncertainty of the environment, work duration, and logical relationship, this paper proposed CCPSP (critical chain project scheduling problem) optimal scheduling model which integrates CCM (critical chain method) and multi-level GERT (graphical evaluation and review technique) network. DEACA (differential evolution ant colony algorithm) is designed to solve the problem, in which DE (differential evolution algorithm) is used to deal with the cross problem and mutation problem, and ACA (ant colony algorithm) is used to deal with each stage of the optimization process. The neighborhood structure is introduced to reduce the invalid search. By storing the non-dominated Pareto optimal solution in external files, the Pareto optimal frontier is obtained. This paper makes full use of the advantages of the two algorithms to achieve fast convergence and a global search for optimal solutions. At the same time of seeking the optimization algorithm, considering the multi-resource constraints, environmental uncertainties and other factors of the project, a buffer is set up, and a “minimum duration - maximum robustness” project schedule is formulated. Finally, through the analysis and verification of the project example, it shows that the project schedule made by this method is feasible.
|
|---|