| Summary: | 碩士 === 國立交通大學 === 運輸與物流管理學系 === 107 === This research introduces an extension of the general share-a-ride problem or G-SARP, called the G-SARP with electric vehicles (G-SARP-EVs). This problem considers a taxi fleet with mixed plug-in electric vehicles and gasoline vehicles can service passenger and parcel requests simultaneously. In this problem, taxis are allowed to convey more than one passenger at the same time, and there is no restriction on the maximum riding time of a passenger. In addition, the number of parcel requests that can be inserted between the pick-up and drop-off points of a passenger is limited only by vehicle capacity. This problem considers only advance requests that are given prior to the beginning of the planning horizon.
The research develops a multi-layer time-space network to effectively describe the movements of passengers, parcels and taxis in the spatial and temporal dimensions. Each taxi operates on its own layer of the network for tracking the energy consumption and load of taxis. The EVs have the priority to service the passenger and parcel requests. An optimization model is to determine the optimal schedule for the taxi fleet to service the given requests. The objective is to maximize the profit of the taxi company. Also, a meta-heuristic based on simulated annealing is proposed to efficiently solve large-scale instances of the problem.
To examine the performance of the proposed model and the heuristic, this study generates a number of instances with various sizes from the data provided by a logistics service provider in Taiwan. The model is solved by the optimization solver, Gurobi. The results from solving small-size instances show that both Gurobi and SA can obtain the optimal schedules for the taxi fleet to service all the requests in a short time, while the results from solving medium-size and larger-size instances show that SA can obtain a better solution in a shorter time than Gurobi.
|
|---|
