Genetic Algorithm Applied for Optimization of Pavement Maintenance under Overload Traffic: Case Study Indonesia National Highway

Abstract

National Road Network which consists of a traditional road structure and modern roads, require planned maintenance and should be in accordance with the needs. The limited choice of available national road network and the deviation of the overloading encourage the government to be more responsive to carry out maintenance management. The institution in charge of road maintenance is often constrained by the limited budget available. A two-objective optimization model considers maximum roughness and minimum maintenance cost for used road network with overload. The study was conducted on the entire national road network in West Java which are paved with flexible pavement. In the proposed approach, data mining model are used for predicting the roughness index over a given period of time. Routine and periodic maintenance are chosen in this study. Multi-objective optimization model was developed based on Genetic Algorithms. Budget constraints and overloading are the two constraints in the developed model. Based on the R-Tools result, the Pareto optimal solutions of the two objective functions are obtained. From the optimal solutions represented by roughness index and cost, an agency more easily obtain the information of the maintenance planning. Results of the developed model has been implemented through the selection of maintenance on the road network scenarios with different levels of overload.