Capacity Planning And Battery Charging Policy For Electric Vehicle Battery Swap Station

The electric vehicle(EV)is widely considered as a solution to the problems of increasing carbon emissions and dependence on fossil fuels.However,the adoption of EVs remains slow due to limited driving range,long charging time and high upfront cost.To overcome these challenges,the industry has proposed an innovative business model called battery swapping.In this model,a fully charged battery for an EV can be replaced within less than 3 minutes using an ingenious automatic system in battery swap station(BSS),which essentially solves the problem of long charging time.The BSS faces crucial trade-off decisions regarding balancing the initial capacity investments and later operational costs as well as balancing the operational cost between high and low efficiency supply modes of the inventory.Also,when a battery swapping service provider manage two stations concurrently,she must decide how to charge the batteries and how to transshipment the demand to provide a high service level.In this context,this thesis investigates the following three problems:First,this study proposes a finite horizon Markov decision process(MDP)model to explore the problem of determining the optimal charging policy for a BSS.In particular,we consider normal charging mode to supply the electricity for the batteries.We derive a series of structural properties and characterize the optimal charging policy.An efficient monotone backward induction algorithm is provided to solve the finite horizon MDP problem.Second,this paper investigates the problem of the joint decision of capacity planning and battery charging policy for a BSS.We develop a two-stage stochastic program to solve this problem,and the model particularly incorporates a MDP.With stochastic demands and time-varying charging costs,the optimal numbers of batteries and the fast and normal charging devices are decided in the first stage,and the optimal charging policy between fast and normal charging modes is selected in the second stage to minimize the total cost.We thus obtain the optimal capacity planning and “two-threshold” charging policy.Some important managerial insights are derived from our extensive computational experiments,in which some parameters are based on real-world data.Third,we build a stochastic dynamic program for the problem of joint managing two battery swap stations considering demand transfer.We adopt the concept of multimudularity to analyze the model,and obtain some structural properties.Based on these properties,we characterize the optimal charging policy and demand transfer policy.