stage genetic algorithm

Advanced planning and scheduling is a production management process in which the resources and production capacities are optimally assigned to the customers’ demands. This approach can particularly be applicable in the complex environments. A common assumption in the advanced planning and scheduling problems is that the processing time of a given product is constant and independent of its position in the production sequence. However, in the real-world situations, an operator’s skill may continuously be improved when the production time is passing which is known as the learning effect phenomenon. In this article, with regard to the learning effect, an extended multi-product optimization framework for the advanced planning and scheduling problem of a typical flexible production environment is developed to provide a more ability to address the actual situations. Due to the high computational complexity of the proposed model, a multi-stage genetic solution algorithm is also presented. Numerical results confirm that the proposed algorithm can obtain the optimum/near optimum solutions in much less computational times compared to the exact solutions.

In this study، we present a MRCPSP/max (Multi-mode Resource-Constrained Project Scheduling Problem with Minimum and Maximum time lags) model with minimization tardiness costs and maximization earliness rewards of activities as objective. The proposed model is nearby to real-world problems and has wide applications in various projects. This problem is not available in the literature exactly and we developed it for the first time. In order to solve this problem، we developed a two-stage genetic algorithm. In the first stage، the main problem is simplified، through applying a genetic algorithm، in which each activity has only one executive mode. In the second phase، with developing another genetic algorithm، the best answer of the problem is achieved. Each phase has its own codification، fitness function، crossover operator and mutation operator. Finally، the computational results obtained from the algorithms of this research، which was written in MATLAB programming language، was compared with the results existing in the project scheduling problems library (PSPLIB). The findings show that، our algorithm improved some of the best solutions، recorded in the PSPLIB.