Research On Resource Constrained Multi Project Scheduling Problem

The resource-constrained multi-project scheduling problem(RCMPSP) is one of the widely used combinatorial optimization problems. The problem studies how to schedule the activities of two or more projects to achieve some optimal managerial objectives under precedence and resource constraints. RCMPSP not only has rich theoretical research value, but also has tremendous engineering value which plays an active significance in improving enterprise resource, shortening the project period and reducing the project management costs.Most of research literatures on classic RCMPSP only use single schedule generation scheme, either serial schedule generation scheme or parallel schedule generation scheme. However, it might not be possible to know what kind of schedule generation scheme is suitable for specific instance of the problem. Therefore, a self-adapting genetic algorithm is proposed. An additional gene is added to determine which schedule generation scheme is actually used. Through genetic manipulation, the suitable one was survived by cross and mutation and other genetic operations, and the other one was eliminated. This allows the genetic algorithm to adapt itself to the problem instance and ensures the optimum of the scheduling results, and the matching genetic operation is designed.Classic RCMPSP is based on an assumption that resources are transferred from one job to another one without any expense of time or cost, but this assumption doesn't accord with the actual situation. So resource transfer time was defined in this paper, and a mathematical model of the multi-project scheduling problem which target is that the weighted sum of the duration of the multi-project and each project is shortest is proposed. Unlike classic RCMPSP, new model must asynchronously consider the constraint of transfer time and the number of transfer resource. The three-stage heuristic rules are used as decode into modified GA.Finally, in order to validate the necessity of considering the transfer time and the effictiveness of the proposed algorithms, a case with 5-projects problem is proposed to compares with other algorithms and finds out that the test results of the proposed algorithms are better than the other algorithms. So, the effectiveness of this algorithm is confirmed and in the same time the importance of resource transfer times is verified.