Optimal Scheduling Problem for Express Delivery Services Using E-scooters

• Main Authors: Sun, Jui-Jing, 孫瑞璟
• Other Authors: Lu, Chung-Cheng
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/xa4xx7

碩士 === 國立交通大學 === 運輸與物流管理學系 === 107 === In response to B2C and C2C business models, courier service providers adopt scooters for express delivery service in metropolitan areas, which increases convenience and timeliness for delivery. Besides, along with the growth of environmental awareness, the government actively promote low-carbon transport, which makes courier service providers gradually replace petroleum motorcycles with electric scooters to achieve green logistics. However, courier services providers will need to consider vehicle mileage range and the location of charging stations when planning the routing of a fleet of e-scooters. Therefore, the study addresses the optimization problem of scheduling a fleet of e-scooters that are assigned for paired pickup and delivery services. The study proposes an optimization model which is developed based on the time-space network. The objective is to minimize service providers’ total operating cost subject to a set of operating constraints for e-scooters. The model is formulated as an integer multiple-commodity network flow problem, which is characterized as NP-hard. Hence, the study develops a decomposition-based heuristic, with the assistance of the mathematical problem solver, Gurobi, to efficiently solve the problem with practical sizes. Test instances are generated based on the data provided by a Taiwan courier service provider, in order to evaluate the efficiency and effectiveness of the proposed model and the heuristic algorithm. The result shows that the heuristic algorithm takes within 30 minutes to complete the solution process. In addition, the gaps between heuristic solutions and lower bounds that are obtained by Gurobi are less than 8%. As a result, it is shown that the proposed model and heuristic algorithm are effective planning tools for scheduling a fleet of e-scooters that are appointed for paired pickup and delivery services.