An efficient algorithm for taxi system optimization

Taxi service is an important mode of modern public transportation. However, operated by a large number of self-controlled and profit-driven taxi drivers, taxi systems are quite in efficient and difficult to analyze and regulate. While there has been some work on designing algorithms for improving taxi system efficiency, the state of the art algorithm, unfortunately, cannot scale up efficiently. To address the inadequacy, we propose a novel algorithm--FLORA--in this paper. Using convex polytope representation conversion techniques, FLORA provides a fully compact representation of taxi drivers' strategy space, and avoids enumerating any type of schedules. Experimental results show orders of magnitude improvement of FLORA in terms of the complexity.