Title Development of the Task Assignment Models for Crowdsourcing Transportation
Year 2019
Degree Master
School National Yang Ming Chiao Tung University Department of Transportation and Logistics Management
Author Liu, Shih-We
Summary
        With the situation of high car holding rate and low utilization rate, and insufficient urban parking space, crowdsourcing transportation is more important. Most of the research in spatial crowdsourcing considered the assignment problem which spatial tasks contain only one situation, and it rarely assigned the tasks that contain two situation, i.e., transportation task. Most of the taxi carrier’s assignment model in Taiwan do not consider origin and destination at the same time. If the company use the automated assignment mode can improve the lack that no one is willing to serve the remote order, and improve the fairness of every driver’s revenue and the efficiency of transportation resource. Therefore, the study will stand in the perspective of operator of crowdsourcing transportation platform to use the mathematical approach to establish the automatic assignment model that considers the origin and destination of task as well as fairness of the assignment at the same time.
       The study will base on the driver's past average rating from customer, past average task difficulty, past average income, worker unassigned times, and task unassigned times to determine the assigns priority of each potential match. This study set the platform to assign the drivers in every five minute, and solved the small-scale and large-scale examples. The result shows that it is efficient to assign crowd drivers to perform the transportation tasks. The sensitivity analyses show the effect of whether the model considers unassigned times in assignment process. With the popularity of sharing economy, the study establishes the assignment model that considers the origin and destination of task as well as fairness of the assignment at the same time. The operator of crowdsourcing transportation platform could use the proposed model as reference for automatic assigning crowdsourcing transportation task.
Hashtags
View count: 151