Hybrid PSO and GA approach for optimizing surveyed asphalt pavement inspection units in massive network

Abstract

Abstract This study proposes an optimal arrangement of surveyed pavement inspection units (SIUs) for cost reduction, minimization of inspection errors and accuracy improvement of pavement network analysis. Inspection process requires surveying billions of distresses characteristics for different sections of a specific area. A comprehensive database is generally required for such pavement management system (PMS). A major concern with this type of data is lack of powerful methods for an effective analysis in a network level. A number of inspection units are surveyed with various sampling patterns for minimizing the cost and time of inspection. Analysis of large numbers of sections and inspection units is time consuming and needs high computation efforts. To address this issue, this paper focuses on developing efficient methods for decreasing complexity of the system. Accordingly, various combinations of the hybrid genetic algorithm (GA) and particle swarm optimization (PSO) are used for analyzing a typical pavement network. The numerical results confirm the ability of the proposed approach to optimize the arrangement of SIUs in network inspection error (NIE), computation time (CPU Time), number of SIUs (NSIUs), and convergence diagram for network, project and section management levels. The hybrid approaches result in an optimal solution in a short time with high accuracy for each section in a massive network. As a result, the inspection process can be performed with minimal costs.