An optimization Model Combining Operator-Based Relocation With User-Based Relocation for Electric Carsharing Systems

Abstract

This paper proposes an optimization model combining operator-based relocation with user-based relocation for electric carsharing systems, aiming to improve the low usage of vehicles and the imbalance between supply and demand. First, the back-propagation neural network prediction model of vehicle quantity demand and the distribution fitting model of energy demand are constructed. In the operator-based scheme, optimization models of inter-regional and intra-regional relocation are constructed. In the user-based scheme, a multi-objective optimization model considering the operator profit and user experience is proposed. Taking the actual operation data of an electric carsharing company in Shanghai as an example, the mixed scheduling strategy of operator-based relocation and user-based relocation is validated. Results show that: (i) users are very sensitive to price changes, and station distance and price incentive will change users’ pickup and return behavior; (ii) compared with systems without vehicle relocation, the mixed scheduling strategy can reduce the using failure rate by 6.68% and increase net profit by 42.6% in the best situation; and (iii) the using-failure rate alert line, as a switching mechanism between user-based and operator-based relocation, can play a regulatory role, and with the loosening of the alert line, the frequency of operator-based relocation intervention increases, and the using-failure rate decreases further.