Extending Microsoft ® Project for Real-World Job-Shop Scheduling with Genetic Algorithm

Abstract — Although production scheduling has attracted the research interest of production economics communities for decades, there still remains a gap between academic research and real-world problems. Genetic Algorithms (GA) constitute a technique that has already been applied to a variety of combinatorial problems. We will explain the application of a GA approach to bridge this gap for job-shop scheduling prob-lems, for example to minimize makespan of a production pro-gram or to increase the due-date reliability of jobs. The pre-sented approach focuses on integrating a scheduling algorithm, based on GA, into a commercial software product, namely Microsoft Project 2003. We extended Microsoft Project in a range of aspects: A new graphical user interface is introduced to support users by a guided wizard describing the problem for which an optimal schedule is sought. The GA was developed to search for the solution with the maximum results for a given set of production logistical objectives. The developed GA algo-rithm and operators are tested on real-world data from a one-of-a-kind manufacturing department of a major company. It includes different aggregation operators for combining objec-tives. Furthermore, the efficiency of the algorithm was com-pared to benchmark tests available in literature