Study of Optimal Solution For Routing Problems Using Metaheuristic Algorithms

Abstract

Due to many real-world applications including courier services, garbage pickup, newlinewaste collection, and public transportation planning, solving VRP efficiently has gained newlinesignificant attention in both academic and industrial domains. VRP is a well-known newlinecombinatorial optimization problem that plays a crucial role in the field of logistics and newlinetransportation which involves determining the optimal set of routes for a fleet of newlinevehicles to deliver goods or services to a set of customers, subject to various constraints newlinesuch as vehicle capacity, route length, and time windows. It is NP-hard problem whose newlinecomputational complexity increases rapidly with the number of customers and newlineconstraints, making exact solutions impractical for large instances. newlineA deep insight of the literature reveals the need to develop efficient solution strategies newlinefor complex VRP variants using intelligent optimization techniques. Traditional newlinealgorithms often fail to produce feasible solutions within reasonable computational newlinetimes for large-scale problems. Therefore, metaheuristic algorithms like Genetic newlineAlgorithm (GA), Particle Swarm Optimization (PSO) including their hybridization with newlinevarious clustering techniques have been employed to address this challenge effectively. newlineThis research purposes three logistic models to address the vehicle routing problems newlinenamely, (i) capacitated vehicle routing problem, (ii) vehicle routing problem with time newlinewindow and (iii) vehicle routing problem with uncertain customer demands. The first newlinemodel generates classical, distance centric and economically viable route plans. In newlinesecond model, the environment friendly route is generated via reducing fuel newlineconsumption along route plans. In third model route plans are designed to full fill newlineuncertain demands of customers in economic way. The three models are formulated newlineand are solved by metaheuristic algorithms. Owing to the computational time newlinecomplexity of NP- hard problems for generating time bound results, Genetic algorithm newline(GA) is designed to solve the models. The perfor