Evolutionary Algorithm for Project Scheduling under Irregular Resource Changes

Over the last few decades, project scheduling problems have been solved under a set of resource constraints, which are assumed fixed throughout the project horizon. However, in real-life applications, resources may change over time due to maintenance or because the resources are needed for another project. Therefore, this research introduces a hybrid evolutionary framework, based on two multi-operator evolutionary algorithms, and a heuristic technique, for a multi-mode project scheduling under irregular resources changes. The framework simultaneously considers both algorithms and self-adaptively emphasizes the one which performs comparatively better. The heuristic considers two variants of handling techniques for irregular resources. One is based on inserting buffer activities to characterize resources unavailable, and another is based on a modified serial generation scheme, that determines the best modes of the activities at each time period based on irregular resources. The framework is tested by solving a set of test problems, with the renewable resources considered irregular over the project horizon. The results demonstrate that the multi-method algorithm has some advantages for scheduling a project, under both regular and irregular resources.