DYNAMIC MATCHING SYSTEM FOR DRIVER DISPATCHING WITH APPLICATION TO ON-DEMAND SERVICE PLATFORMS

摘要

#$%^&*AU2020100669A420200604.pdf#####ABSTRACT With the development of technological advances in mobile networks and Global Position System, on-demand services have emerged on a massive scale. We propose a measurement-based algorithm, which is capable to deal with online case in large scale and to approximate the optimum at a ratio o for cost minimization. The request-driver measurement is based on four factors: passenger's waiting time calculated by the drop-off time of its last request plus the drive from the drop-off location to the pick-up location of the request; the time to home after serving the request; the requests density to evaluate the benefit to serve the current order; the importance of the order, which may reflect the revenue or the willing to pay. The request is assigned to the driver with the lowest measurement. The system will also set a threshold and reject the requests whose measurement function is above the threshold for all drivers.DRAWINGS The measurement function with d is start smaller than the current driver New renueqt r Yes No Assign the new Assign r to d,i.e. d*order to do by d default. d* = do Find a (next) driver d d is the last driver? No Yes Assign r to d*, add it to the end of d*'s to-do list end Figure 1. Schematically flowchart of the proposed MB algorithm