Solving the MRCPSP/max with the objective of minimizing tardiness/earliness cost of activities with double genetic algorithms

Abstract

This paper presents a meta-heuristic method for a resource-constrained project scheduling problem. In this problem, the activities are given with multiple executive modes and there are minimum as well as maximum time lags between different activities. The objective here is to determine a mode and a start (or finish) time, for each activity such that total tardiness/earliness cost of activities could be minimized. The meta-heuristic approach was developed, as a two-phased genetic algorithm. The process of solving the proposed problem includes two stages (phases). In the first stage, through applying a genetic algorithm, the main problem is simplified so that each activity has only one executive mode. In the second phase, through developing another genetic algorithm, the best answer to the problem is achieved. Finally, the computational results obtained from the solution algorithms of this research were compared with the results existing in the Project Scheduling Problem Library (PSPLIB). These algorithms were provided in the MATLAB programming language. The findings show that our algorithm improved some of the best recorded solutions in the PSPLIB.