Fund Allocation in Complex Rehabilitation Programs

Tarek Hegazy, Roozbeh Rashedi

Civil Infrastructure assets require continuous renewal actions to sustain their operability and safety. Allocating limited renewal funds amongst numerous building components, however, represents a large-scale optimization problem and earlier efforts utilized genetic algorithms (GAs) to optimize medium size problems yet exhibit steep performance degradation as problem size increases. In this research, after experimenting with various approaches of segmenting a large problem into multiple smaller sub-problems, clustered segmentation proved to be the most promising. The paper discusses the underlying life cycle analysis model, the various segmentation methods, and the optimization results using the improved GAs + clustered segmentation, which proved to be able to optimize asset renewals for 50,000 components with no noticeable performance degradation. The proposed method is simple and logical, and can be used on variety of asset types to improve infrastructure fund allocation. Future extension of this research is then highlighted.

Full Text