In academia, the problem of CS planning has been extensively studied [
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15,
16]. In Ref. [
2], research has been conducted on CSs’ profit from the operator’s point of view. A heuristic Removing and Merging Possible Locations (RMPL) will screen nodes in the network and ignore the nodes with few profits to obtain: The multi-objective evolutionary algorithm used in Ref. [
3] aims to minimize the investment and operation costs of the distribution system while maximizing the annually captured traffic flow. This research mainly focuses on the location of CSs: Reference [
4] gives a charging infrastructure planning with the goal to minimize the fleet’s daily charging operation time. Compared with general research, this work focuses on how to manage the fleet. Luo et al. [
5] assume that charging services are provided by multiple competitors. Bayesian game is used in this situation so that each of them could get an optimal placement policy. What is more, they found that the CS placement is highly consistent with the heatmap of the traffic flow. Reference [
6] focuses on how to balance the charging demand and power network stability. A spatiotemporal model of the charging demand is proposed and a heuristic algorithm involving the grid constraints (HAG) is designed. The utilization of chargers and the carrying capacity of the power network are all improved. Reference [
7] presents a Mixed-Integer Non-Linear (MINLP) optimization approach for the optimal placing and sizing of the fast CSs. What is significant is that the robustness and efficacy of the proposed method is studied. Bi et al. [
8] study the reasonable layout of CSs by distributing chargers. The nodes, where CSs are located, are selected by the betweenness centrality (BC) in a complex network; In Ref. [
9], four methods, i.e., the iterative MILP, greedy approach, effective MILP and chemical reaction optimization, are proposed to minimize construction cost of CSs, and they are evaluated from multiple perspectives. Reference [
10] aims to minimize drivers’ charging cost. This work takes EV drivers’ strategic and competitive charging behaviors into account and makes the problem closer to reality. Reference [
11] takes the environmental factors and service radius of the CSs into account to minimize total cost of CSs. They set the distance between two CSs with EVs’ reaching distance and choose the location of charging stations with the Voronoi diagram. In Ref. [
12], the yearly cost of CSs considering the battery capacity constraint instead of the service radius, which makes the modeling more rigorous and realistic. In Ref. [
13], CSs are divided into fast charging and slow charging. At the first stage, we select the best CS locations to minimize EV transportation energy losses. Then, the optimum numbers of slow and fast charging facilities are obtained to minimize the cost and meet demand. In Ref. [
14], the LCC criterion is used to assess the project and a modified differential evolution algorithm is adopted to solve the problem. What is creative is that the research object of this work is the battery-swap station rather than the CS. Reference [
15] uses regression equations to predict the parking demand variables. They aim to minimize the EV users’ station access costs with considering the parking demand of a vehicle. In Ref. [
16], uncertainties existing in the development of future EV technology are properly modeled to ensure the robustness of the planning scheme. It is worth mentioning that this work not only focuses on the present, but also exhibits robustness for all the considered scenarios in the future stage. Reference [
17] presents CS Dimensioning and Placement (CSDP) framework for provisioning fast charging infrastructure at minimum cost to accommodate the charging demand of the incremental integration of EVs. The solution can efficiently expand the CS network to accommodate future EV charging and conventional load demands. Reference [
18] finds the locations for fast and slow charging stations by analyzing drivers’ daily schedules. A centralised charging station database (CSDB) is employed to reduce the waiting times at charging stations. They find that combining some centralized fast charging stations with many distributed slow charging stations will reduce the charging time. Reference [
19] proposes a two-stage CS planning method for sharing EV: In the first stage, the charging stations are sited and sized based on the SEV charging demand estimation. In the second stage, the unsatisfied charging demands are assessed so as to update the charging station capacity accordingly. In Ref. [
20], a CS planning model is established to maximize the fuzzy quality of service (FQoS), considering queueing behavior, blocking reliability, and multiple charging options. The results demonstrate that the consideration of FQoS faciliates the finding of the more robust capacity plan. The implementation of the proposed model will be useful for designing a charging station without enough EV arrival and charging service data.
Although above studies have analyzed and optimized the allocation of CSs from various aspects, it should be noted that difference of land price in regions (e.g., spatial land price) is an important factor, which will bring a lot of changes to the problem. References [
3,
4,
6,
8,
9,
12,
15,
18,
20] have not taken spatial land price into considering, this makes them ignore the great impact of spatial land price. Therefore, these works lack some realism. Research such as [
3,
6,
8] notice the construction cost of CS, but they only associate it with the number of chargers. It is obvious that the location would affect the construction cost. References [
2,
5,
7,
10,
11,
13,
14,
16,
17,
19] consider spatial land price. In these studies, the construction cost of CS is related to chargers’ number and location. But they just took the construction cost of CS as a parameter; the specific impacts brought by spatial land price and its gap are still left out. The location of CSs and allocation of charging resources are two crucial aspects for EVs’ transportation. Reasonable location and charging resources allocation of CSs can maximize the effectiveness of the limited budget. For example, setting CSs in places with large population and traffic flow may be effective, but it will also bring high construction costs, mainly due to land prices. On the other hand, relatively remote CSs cannot meet the charging demand of EVs and cause unnecessary driving time. Spatial land price will make the value of nodes different and the gap of spatial land price will bring great impacts on the whole system.
In view of the above, construction cost of CSs and importance of nodes are added to the allocation of charging resources. An optimization algorithm is designed for charging resources allocation. The influence brought by spatial land price is studied from multiple aspects and the cause of changes is analyzed and discussed deeply. The major contributions are as follows: