| Summary: | 碩士 === 國立成功大學 === 工業與資訊管理學系 === 107 === With the increasing popularity of on-line shopping and cloud kitchens, the same-day delivery demands have rapidly grown and caused much city logistics challenges. More vehicles for shipping huge amount of origin-destination (OD) deliveries inside a city lead to more traffic jams, accidents, energy consumption and air pollutions. We proposed a Hub-and-Spoke (HS) framework to design the logistics network for these OD shipments, where small autonomous vehicles are used to conduct the first-and-last mile delivery that connect customers to a hub, and then trucks tranship those shipments between hubs. Such an HS logistics network helps reduce the number of vehicles required for inidividual OD shipmeents. Given the estimated amount of OD shipments in each time period during one day, we seek optimal schedules and routes for transshipment trucks between hubs. Moreover, by treating the autonomous vehicles as shared vehicles, we can also reposition these autonomous vehicles between hubs in different time periods to increase the utilization of autonomous vehicles and trucks. We have proposed Integer Programs on a time-space network to calculate exact optimal solutions, but it is too time consuming. To solve more practical cases of larger sizes, we propose a greedy algorithm and a Genetic Algorithm to find good feasible truck routes in short time. Then we propose a framework that iteratively solves a smaller time-space network of few periods in a rolling horizon fashion, which produces the best results in our computational experiments.
|
|---|
