UTILIZATION OF EVOLUTIONARY ALGORITHMS TO IMPROVE ENGINEERING DECISION-MAKING

摘要

While genetic algorithms have been explored academically, and in at least two known cases, commercialized, they still seem to be an underutilized technique for solving engineering management problems. One of the reasons for this underutilization is the general lack of awareness and understanding about genetic algorithms. Another reason is that genetic algorithms have generally been viewed as being too costly in terms of computing resources to implement for complex problems. This paper examines an example management problem and an example genetic algorithm that was developed to solve the problem in order to illustrate the applicability of such algorithms to similar or more complex management problems. In an experiment described in this paper, a Java-based genetic algorithm package was used to implement an evolutionary process for finding optimal solutions to a selected aggregate planning problem. The performance and results of the process were compared to solutions to the same problem found using manual and non-evolutionary automated techniques. The results found using the experimental evolutionary process surpassed those obtained by the other techniques, both in terms of speed and optimization.