A Multi-Pass Biased Random Sampling Method for Scheduling Multiple Projects

The serial schedule generation scheme used for single project scheduling is extended in this paper,and a multi-pass biased random sampling method is incorporated for solving resource-constrained multi-project scheduling problems.Different priority rules are adopted in the random sampling procedure,and a statistical test shows that they affect the performance of the proposed random sampling method.Among all the priority rules tested,MINSLK is the most suitable one for the random sampling method and is able to reduce the overall project durations effectively.Comparison with other tested heuristics verifies that the proposed random sampling method is significantly superior to other heuristics for scheduling resource-constrained multiple projects.