Infrastructure management offices plan and complete several thousand small construction projects annually. Effective planning is vital if the public and private sectors are to maintain valuable infrastructure investments at the least cost to the taxpayer or shareholder. This paper presents the results an application of Genetic Algorithms (GA) in multi-project resource allocation to minimize the total cost of work order execution on realistically sized infrastructure management problems. In addition to direct crew costs indirect costs for set-up, idle time, and travel are included in this model. Results of test cases demonstrate the effectiveness of the approach when compared to several standard heuristics.
E. William East