Optimization Modeling of Urban Freight Transportation Network by Using a Metaheuristic Approach, Genetic Local Search Procedure

Abstract

Creating a transportation network that reduces the cost of urban freight is highly challenging. Freight route is determined by the amount of congestion generated by the vehicle type. Furthermore, it must compete with other users to locate the optimal path for their traffic on the city's restricted road network infrastructure system. Freight network routing begins with identifying and determining the optimum route. The freight network is optimized by introducing a heavy traffic restriction mechanism in metropolitan areas, and an attempt is made to propose a set of routes as a set of freight traffic modes. The primary goal of freight routing in this research is making a freight network model optimization to find a set of freight routing by optimizing (efficiency) trip prices due to limited road infrastructure and the difficulties of constructing road infrastructure in metropolitan areas. Therefore, selecting a group of routes as the journey of the commodities is a realistic alternative to minimizing costs. The problem of route selection is one of combinatorial optimization. The challenge is to narrow the pool of action options to a set of recommended actions. Due to vehicle characteristics and traffic flow, route selection carefully considers vehicle behavior. A two-level mathematical model that was created by formulating route options served as the framework for the research. The combination of chosen routes is maximized using a genetic algorithm such as a Genetic Local Search. The model is examined via its application to a fictitious network. The result converges to the target value of 246,311.9 IDR. It shows that the model satisfies the convergence condition of producing in 0.76 seconds. As a result, a model with a genetic local search technique that can search more effectively in the city's freight network's ideal path is created. The GLS combinatorial optimization model shows us it could find the best set of urban freight networks with the best performance. This is consistent with Yamada et al. GLS methods perform well in solving combinatorial problems.