A Markovian approach for optimizing highway life-cycle with genetic algorithms by considering maintenance of roadside appurtenances

Roadside appurtenances, such as guardrails, signs, and luminaries form a major portion of highway features whose upkeep and timely maintenance is very critical to maximize highway life-cycle. Their life-expectancy is largely reduced by poor maintenance. While significant work has been done in pavement and bridge maintenance roadside appurtenances have received limited attention. In this paper we develop a Markovian model for maximizing highway life-cycle by considering optimal maintenance strategies for roadside appurtenances over a given planning horizon. A genetic algorithm is developed to solve the developed optimization model since the resulting Markov tree tends to be clustered and relatively complex. Due to this complexity traditional approaches, such as dynamic programming seem to be computationally inefficient. Separate formulations under deterministic and probabilistic approaches to be undertaken for different maintenance conditions are developed. A probabilistic deterioration function for roadside appurtenances is also developed. An illustrative example to investigate the effectiveness of the deterministic approach is presented. The paper concludes with directions for future research.