考虑混合车位的自动驾驶电动共享汽车系统配置与运营联合优化

Joint Optimization of Mixed Parking Spaces and Dynamic Pricing for a Shared Autonomous Electric Vehicle System

自动驾驶电动共享汽车作为未来可持续发展的交通方式,结合配建充电车位与非充电车位(混合车位)的站点,是调和城市出行、停车和充电供需矛盾的有效途径。考虑混合车位配置、基于Logit弹性需求研究动态定价和车辆自主调度的联合优化问题。基于单商品流时空电池网络,构建了以车位数、出行定价、时空电池弧流量为决策变量,以运营利润最大化为目标的混合整数非线性规划模型。由于该模型难以得出高质量解,以外内切割线近似方法逼近Logit函数,之后通过一种基于二次方的整数变量分解与车位约束松弛方法,共同将原模型重构为混合整数线性规划,并利用GUROBI优化引擎求解。对比了重构前后的配置和运营联合决策优化结果,表明所提出方法可以有效提升求解效率。成都市案例结果表明:与全部使用充电车位相比,混合车位配置在同等订单满足率的前提下,充电车位提高10.59%的利用率并降低28.87%的成本,缩减4.15%的搬迁成本;与固定出行定价相比,动态定价策略在提升整体运营利润的基础上,有效缩减16.03%的非必要调度;动态定价区间高低与车辆搬迁数、OD实际需求量成反比;高车位成本将使非充电车位的租赁数急剧下降;充电速率下降则减少运营收益可达12.35%,且充电与非充电车位数分别增幅14.75%和12.34%。

As a future sustainable transportation mode, shared autonomous electric vehicles combined with the construction of charging and noncharging parking spaces(mixed parking spaces) at stations effectively reconcile the contradictions between urban travel, parking, and charging supply and demand. This paper considered the joint optimization of mixed parking space allocation, dynamic pricing based on the logit-based elastic demand, and vehicle scheduling. Based on the single-commodity space-time-battery network, a mixed-integer nonlinear programming model was constructed with the number of parking spaces, travel pricing, and space-time-battery flow as the decision variables and the maximization of operational profit as the objective. Due to the difficulty of obtaining high-quality solutions for this model, this paper fitted the logit function using an outer-inner approximation method. Then, using a quadratic-based integer variable decomposition and a parking constraint relaxation method, the original model was reconstructed as a mixed-integer linear programming model and solved using the GUROBI engine. A comparison of the configuration and operational joint optimization results before and after the reconstruction shows that the proposed method effectively improves the solution efficiency. The results of the Chengdu case study indicate that compared to using only charging spaces, mixed parking space allocation increased charging space utilization by 10.59% and decreased charging space rent by 28.87% while reducing relocation costs by 4.15% under the same order fulfillment rate. Compared with fixed pricing for trips, dynamic pricing effectively reduces fleet size by 16.03% while increasing operational profits. The range of dynamic pricing is inversely proportional to the number of vehicle relocations and the actual demand for origin-destination. Higher parking costs led to a sharp decrease in the leasing of noncharging parking spaces. A decrease in the charging rate can reduce the operational revenue by 12.35%, which contributed to an increase of 14.75% in charging parking spaces and 12.34% in noncharging parking spaces.