Front. Energy Res. Frontiers in Energy Research Front. Energy Res. 2296-598X Frontiers Media S.A. 1379612 10.3389/fenrg.2024.1379612 Energy Research Original Research RETRACTED: Energy e-commerce user portrait and multi-agent cooperative game price mechanism design Yang et al. 10.3389/fenrg.2024.1379612 Yang Biao Liu Suwei Gao Hongda * State Grid Energy Research Institute Co., LTD., Beijing, China

Edited by: Zening Li, Taiyuan University of Technology, China

Reviewed by: Lirong Deng, Shanghai University of Electric Power, China

Zhenjia Lin, Hong Kong Polytechnic University, Hong Kong, SAR China

 *Correspondence: Hongda Gao, ghd2612@163.com 22 05 2024 12 11 2025 2024 12 1379612 31 01 2024 21 03 2024 Copyright © 2024 Yang, Liu and Gao. 2024 Yang, Liu and Gao

This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

With the development of big data, our lives have gradually entered the information age, which has changed and reshaped the behavior of enterprises and consumers. In this paper, a user portrait clustering model based on big data is proposed to implement business model design for specific groups after clustering, target potential user groups for active marketing, and promote actual purchase behavior. In this paper, cost, risk, and contribution factors are introduced to improve the basic cooperative game allocation method. The improved model considers the operating cost of the main body, the level of risk, and the contribution proportion of the actual energy supply. In order to verify the effectiveness and applicability of the benefit distribution strategy based on the cooperative game proposed by the project, the research results provide a certain reference for precision marketing in relevant industries and enterprises.

 customer habit business modes market forecasting game theory big data State Grid Corporation of China10.13039/501100010880 section-at-acceptance Sustainable Energy Systems 1 Introduction

The definition of “new retail” in the realm of e-commerce aims to build an offline channel and seamlessly combine it into the existing online retail channel (Wang et al., 2020a). The industry emphasizes leveraging technological advancements to optimize the retail experience and meet evolving consumer demands (Mahadevan, 2004). Information technology encompassing big data, virtual reality, and artificial intelligence is poised to revolutionize the landscape and trajectory of the new retail business. Scholars emphasize that the essence of new retail is integrating online and offline channels, satisfying the complex needs of customers, striking equilibrium between personalization and privacy, and optimizing the supply chain network (Zhang, 2021). Studies indicate that online channels can significantly boost sales, reduce costs, optimize inventory management, and enhance return on investment (ROI) for individual industries. Furthermore, the harmonious combination of offline and online retail channels is beneficial for enterprises, fostering collaboration rather than competition (Amit and Zott, 2001).

In order to meet customers’ needs, enterprises need to focus on providing unique value and creating an inviting shopping environment. By offering products or services that stand out and cater to specific customer desires, businesses can differentiate themselves from competitors and leave a lasting impression on customers. This emphasis on unique value and an appealing shopping environment plays a pivotal role in attracting and retaining a loyal customer base, ultimately contributing to the long-range success of the companies.

The emergence of new retail faces three primary challenges: determining suitable locations for offline stores, competing on price with established online retailers, and achieving consumer recognition across both channels. The crux of these challenges lies in establishing a retail platform that can facilitate seamless integration between multiple online and offline channels. Furthermore, supporting this integration with cutting-edge technologies and innovations, particularly emphasizing that information technology is paramount to the success of new retail, is essential (Wang and Chang, 2021). The value of AI solutions has emerged as a critical area of interest. Additionally, scholars have delved into the impact of supply chain management within the context of new retail, recognizing its pivotal role in ensuring streamlined operations, optimized inventory management, and responsive adaptation to evolving consumer demands (Lindgadt et al., 2009). The effective application of big data and AI is pivotal in shaping and advancing innovative business models and formats (Jiang et al., 2021). The ongoing advancement of Internet technology serves as the technical backbone of the growth of new retail, enabling e-commerce and offline stores to evolve in tandem rather than as siloed entities. This convergence ushers in a novel zero-operation paradigm, where the new retail landscape thrives within a constantly evolving normal. The transformation is marked by the profound fusion of online and offline channels with logistics infrastructures, coupled with the extensive utilization of big data to augment retail operations (Osterwalder et al., 2005). The progression of new retail necessitates an in-depth examination of the essence of retail, striving to comprehensively meet consumer demands through the amalgamation of cutting-edge technologies, sophisticated logistics, and varied platforms. Academics have emphasized the critical role of science and technology, notably the deployment of big data, in driving the advancement of new retail (Gan et al., 2023).

The Shapley value was proposed by Olga Bondaleva and Lloyd Shapley (1960s) as a solution to the only expected payment based on the players’ marginal contribution in the case of multi-player league games (Wang et al., 2019). It is a mathematical method used to solve the problem of the distribution of players’ benefits during the game. It describes the necessary and sufficient conditions of the no emptiness of the core of a cooperative game in the form of a characteristic function and can form a Pareto improved optimal within the alliance. The Shapley value avoids equalitarianism in the distribution of the benefits of alliance members and is fairer and more rational than any distribution method solely reliant on the input value of resources, the efficiency of resource allocation, and the combination of the two. At the same time, it mobilizes the enthusiasm of the cooperative member enterprises and reflects the game process of the alliance members.

In recent years, many countries have been vigorously promoting enterprise digital transformation efforts (Li et al., 2022). The traditional development model of energy e-commerce has gradually failed to meet the development needs of energy enterprises, and it is urgent to upgrade and transform. The energy Internet form cultivated by “Internet + source” has brought new opportunities for traditional energy e-commerce (Wang et al., 2017). Energy enterprises can make full use of Internet thinking, accelerate the promotion of the status and role of energy e-commerce in enterprises, gradually form a new generation of energy e-commerce development models, and provide enterprises with a sharing platform for factor reorganization, integration, and innovation.

However, during the process of setting up the business model, the most important factor is to design a reasonable profit distribution model. Multiple entities participating in business innovation could be regarded as a game. Only when the participants benefit from both competition and cooperation can the business model continue to develop (Chen et al., 2022). Many scholars have investigated the revenue allocation problems of energy e-commerce (Williams and Tagami, 2002; Wang et al., 2020b). Solidarity value and Shapley value methods are common methods of income distribution in cooperative games. It is a solution set defined in the theory of multi-player cooperative games. The solidarity value assumes that different allies have the same possibility of joining any alliance. Although different weights are considered for different alliances, the marginal contributions of alliance partners are averaged. The profits shared by alliance participants under the solidarity value method are similar to the average for their marginal contributions to all the alliances (Cubukcu, 2019). The integration of a discount factor has enabled the expansion of Shapley values in cooperative games to include interval-valued (IV) cooperative games. Through the introduction of IV discounted Shapley values, researchers have provided a simplified methodology for calculating these values in a specific category of IV cooperative games. This innovative approach facilitates a more thorough examination of coalitional games, accounting for the discounted contributions of players across various time periods. This development represents a significant advancement in the field of game theory, offering new insights into the dynamics of cooperation and competition among players in complex systems (Fei et al., 2018).

According to the findings of the above research, the basic cooperative game distribution method has its own limitations, which include a single consideration factor, a simple model, and the assumption that all income stakeholders have equal status in the distribution of income. Additionally, factors such as the responsibility of stakeholders, the cost of participation, and the level of risk are not taken into consideration. To address these issues, this paper proposes the incorporation of cost, risk, and contribution factors to enhance the basic cooperative game allocation method. The improved model considers the operating cost of the main body, the level of risk, and the proportion of actual energy supply contributed by each stakeholder.

The main contribution of this paper includes the following aspects:

(1) The formation mechanism and typical business model of energy e-commerce are systematically studied. The formation mechanism of the energy e-commerce business model is analyzed from the perspectives of energy e-commerce cooperation subjects, target users, key businesses, profit models, cooperation networks, trading methods, and benefit distribution methods. Then, based on the root factors that affect users’ consumption habits, the typical business model of energy e-commerce is constructed from multiple perspectives by integrating the formation mechanism of the business model.

(2) A multi-dimensional selection model of typical business models of energy e-commerce based on user consumption preferences is constructed. The project uses the user portrait method (UPM), and the K-means clustering method offers an effective means to categorize both C- and B-end large users. Specifically, the C-end users can be classified as rational, high-end, interested, and guided users, while the B-end large users can be segmented into high-value, low-value, and ordinary large users. These methods provide valuable insights for targeted marketing strategies and personalized user experiences, and differentiated business models are designed for different user types. The research provides a decision-making aid for the orderly development of energy e-commerce.

(3) The benefit calculation and benefit distribution model of customers’ participation in energy e-commerce cooperation are established. This paper designs the business operation mode of the energy e-commerce charging business under the cooperation of e-commerce and users and redistributes the benefits of both through case analysis, solving the problem of benefit distribution under the cooperative game of multi-parties in energy e-commerce.

2 Building the user portrait model based on consumption data

The clustering analysis method is derived from the target key factor, focusing on grouping similar factor combinations. This technique is primarily used for sales data analysis, aiming to develop a classification model that identifies customers’ interests and consumption tendencies. Subsequently, the data are categorized into predefined groups, enabling predictions about future consumer behavior. The K-means algorithm model for consumers is exemplified as follows:

E = E 1 , E 2 , E 3 , . . . , E 24 , where a random sample is selected as the initial clustering center. Then, the shortest distance between each sample and the designated cluster center is calculated. The principle is that a larger distance indicates a higher probability of being selected as a new cluster center. Among the commonly used similarity measurement techniques, including Euclidean distance, cosine similarity, Mahalanobis distance, and information entropy, cosine similarity is chosen as the metric. The formula for calculating cosine similarity is provided as follows: cos ⁡ θ = m i p ′ m j p ′ + m i v ′ m j v ′ m i p ′ 2 + m i v ′ 2 m j p ′ 2 + m j v ′ 2 .

The above step is repeated until the Kth cluster center C k is selected to form the initial cluster set, and these three types of sets are defined as Q ζ , according to E i ∈ Q ζ , denoted as E i ζ . In addition, the sample mean of each set is used as the new clustering center Q ζ n e w , i.e., Q ζ n e w = 1 n ζ ∑ k = 1 n ζ E i ζ .

(1) Then, each sample point is assigned to the cluster represented by the closest cluster center point.

(2) The centroid of all sample points within each cluster serves as the center point of the cluster.

(3) Steps (1) and (2) are repeated, re-measuring the similarity of each sample and repeating the cycle until the termination condition is met. The sum of squares of error is defined as the termination function, i.e.,

 T e n d I = ∑ ζ = 1 3 ∑ k = 1 n j E i ζ − S ζ I 2 , where n j is the number of samples contained in the cluster set j . When the difference between two adjacent error function values satisfies the given precision, clustering ends.

 3 Business model construction scheme for C-end customers

Drawing from the user profiling method, the user profile is refined from a detailed perspective, allowing for further segmentation of user types to offer a tailored selection of energy e-commerce business models.

3.1 User portrait modeling scheme design

In order to effectively observe, record, and quantify consumer purchasing behavior and gain insights into consumer psychology, this study uses a questionnaire format to gather behavioral data. To ensure the questionnaire’s scientific design, the expert opinion method was used to gather input from e-commerce professionals, and the questionnaire content was iteratively revised. The final questionnaire design encompasses consumers’ basic information, transaction details, preference information, and specific product-related data.

 3.2 Results analysis based on the k-means clustering method

The choice of the K-value plays a crucial role in determining the level of clustering. Excessive user categorization can necessitate a larger number of user labels, potentially complicating subsequent tasks. Hence, selecting an appropriate K-value to classify users is essential, striking a balance between adequate data segmentation and avoiding undue complexity. To determine the suitable K-values, we must calculate the sum of squares of errors (SSE) within clusters. Figure 1 depicts the relationship between the sum of squares of errors within clusters and the number of clusters, as K varies from 0 to 20. This graph shows that as the K-value increases, clustering becomes more detailed, which is advantageous for a more refined segmentation of user groups. The horizontal axis represents the K-value, while the vertical axis indicates the corresponding SSE value.

Curve of the trend of sum of squares of errors within clusters.

It can be observed that as the K value increases, the sum of squares of errors within the clusters exhibits a decreasing trend. Initially, when the K-value is less than 4, the decrease in SSE is rapid. However, once the K-value reaches 4, the rate of decrease in SSE slows down significantly. Based on these observations, it can be inferred that the optimal K-value for this scenario is 4.

 3.3 Presentation of user portrait results

By setting the K-value to 4, we can identify four central points, which are based on price, emotion, demand, and attachment. By analyzing the coordinates of each central point and marking the crowd characteristic information, we can create user portraits. A label system is then constructed to classify the group portraits. The results of this process are depicted in Figure 2.

User portrait result display.

These user portraits provide a comprehensive understanding of the different user groups, allowing for targeted marketing strategies and product development. The label system also enables easy identification and tracking of user behavior patterns, further enhancing the effectiveness of marketing efforts.

The position of a point on the x-axis indicates the extent to which a user’s purchase behavior is influenced by the price of a product. Points further to the right suggest that users are less price-sensitive, prioritize personal interests, and value satisfaction over cost. Meanwhile, the y-axis represents the level of attention users pay to the added value of a product. Higher y-values indicate that users increasingly focus on additional benefits beyond the core functionality of the product.

Based on Figure 2, which illustrates the clustering outcomes and behavioral traits of various consumer profiles, a descriptive analysis is conducted for each consumer category.

 4 Business model and benefit distribution model 4.1 Modeling of the income distribution approach by cooperative game

In cooperative game theory, the Shapley and Banzhaf value methods are two well-established solution concepts. This section delineates allocation strategies based on these two methodologies.

4.1.1 Allocation strategy using the Shapely value method

The Shapley value represents a distinctive expected payment solution that hinges on the marginal contributions of players within an N-player cooperative game. It serves as a mathematical tool used to address the issue of benefit distribution among players throughout the course of the game. This approach articulates the prerequisites and sufficiency for the non-vacuity of the core of a cooperative game via the medium of a characteristic function and has the capacity to generate a Pareto-improved optimal solution within the coalition. The specific model is outlined as follows: V i = ∑ S ⊆ N \ i S ! N − S − 1 ! N ! v S ∪ i − v S , ∀ i ∈ N , where V i is the profit shared by the subject i and N \ i represents the set composed of all subjects except the subject i in the set N .

 4.1.2 Solidarity value method allocation strategy

Solidarity value and Shapley value methods are common methods of income distribution in cooperative games. It is a solution set defined in N-person cooperative game theory. Solidarity values assume that different allies have the same possibility of joining any alliance. Although different weights are considered for different alliances, the marginal contributions of alliance partners are averaged. Under the solidarity value method, the profit shared by alliance participants is equal to the average of their marginal contribution to all alliances, and its specific model is shown as follows: x i = ∑ S ⊆ N \ i 1 2 N − 1 v S ∪ i − v S   ∀ i ∈ N .

4.1.3 Discriminant conditions

V i ≥ X i . In the formula, the profits obtained by participating in cooperation are greater than those obtained by the subject acting alone: ∑ i ⊂ I V i = V , where V is the total profit, that is, the income of all parties involved in the cooperation is the total income. V ∅ = 0 .

When you do not participate in alliance cooperation, it is assumed that there is no benefit.

 4.2 Income distribution method based on the modified cooperative game

Based on the above research, it can be found that the basic cooperative game distribution method has its own shortcomings; that is, the consideration factor is single, the model is easy to understand, the income stakeholders are thought to have equal status in the income distribution, and influence factors such as the responsibility of stakeholders, the cost of participants, and the level of risk are not considered. In the following, the improved method applies the size of the operating cost of the main body, the level of risk, and the contribution proportion of real energy supplements.

4.2.1 Introduction of the cost factor correction algorithm

Because of the different investment characteristics, the investment costs of different investment entities are different. Therefore, the cost factor is introduced to improve the basic cooperation game. V i ′ = C i C − 1 n ⋅ V + V i , where n is the number of alliance participants, V is the total revenue of the alliance, C is the total cost, and V i ′ is the improved allocation model based on investment.

 4.2.2 Introduction of a value correction algorithm for risk factors

Based on the original assumption of a cooperative game, the risks borne by all parties are assumed to be equal, represented as R ¯ = 1 n ; however, in the actual situation, the risks borne by different entities are different. Therefore, it is necessary to introduce risk factors to revise the basic cooperative game value method.

Let R i be the actual risk to be borne; that is, the enhanced model, which is based on the variance between the actual and average risk, can be represented as follows: V i ″ = R i − R ¯ ⋅ V + V i , where ∑ i = 1 n R i = 1 , V i ″ is an improved allocation model based on the operational risk.

In order to further accurately evaluate the actual operating risk value of different entities, an index system for evaluating the operational risk of the two entities is first constructed, which includes the following three secondary indexes: physical index, economic index, and social index. Among them, the physical index includes construction, output, equipment failure, and equipment maintenance risks. The risk is measured from the perspective of life cycle theory. The economic index includes the net present value index, investment return rate, and dynamic investment payback period, and its risk is measured from the angle of efficiency and value. Sociality includes ecological environmental protection and political risk, which are measured in terms of environmental conservation and policy. The specific indicators are shown in Table 1.

Risk index system.

Secondary index Three-level index
Physical index Construction risk
Output-operating risk
Equipment failure risk
Equipment maintenance risk
Economic index Net present value index
Return on investment
Dynamic payback period
Social index Ecological and environmental protection risks
Political risk
4.2.3 The value correction algorithm for a contribution factor is introduced

Because of the different investment characteristics, different investment entities meet different user needs after investment. Therefore, it is necessary to introduce a contribution factor to improve the basic cooperative game method. V i ‴ = D i D − 1 n ⋅ V + V i , where D is the total demand of users, D i the user load demand contributed by subject i , and V i ‴ is the improved distribution model based on cooperative contribution.

 4.2.4 Comprehensive correction algorithm

In summary, the income distribution algorithm of the improved cooperative game comprehensive revision algorithm is as follows: V i * = λ V i ′ + θ V i ″ + μ V i ‴ , λ + μ + θ = 1 .

In the formula, the value of the weight coefficient of λ , θ , and μ is a fuzzy judgment that is more influenced by subjective factors, and different results will have a certain impact. Therefore, based on the actual situation, the determination method combining subjective and objective using AHP and entropy weight methods is adopted to assign its weight. The discriminant conditions of the improved cooperative game comprehensive correction algorithm refer to the discriminant conditions before the improvement. The specific steps are as follows:

Step 1 : Based on the basic model (4) or (5) and combined with the actual data in the operation process, the initial cooperation game value is obtained.

Step 2 : Based on the actual data of daily operating costs and combined with model (9), the cost-based revision value is obtained.

Step 3 : On the basis of the completion of evaluation indicators, the fuzzy comprehensive evaluation method is used to derive the fundamental weights for physical, economic, and social indicators. Second, on the basis of determining the weights of the secondary indexes, the weight coefficients of the three indexes, including construction risk and operation risk, are obtained using the fuzzy comprehensive evaluation method. Finally, combined with the actual data and the basic weight value, the actual operating risks of the integrated energy distribution network and the main network of each park are obtained. Combined with model (10), the cooperative correction value based on risk management is obtained.

Step 4 : Based on the actual contribution of the agent to meet the load demand of the user, combined with model (11), the cooperative revision value based on the cooperative contribution is obtained.

Step 5 : First, the analytic hierarchy process and entropy weight method are combined to find the weight of cost, business risk, and cooperation contribution. Second, combined with model (12) and the values obtained from steps 2 to 4, a comprehensive and improved cooperative game correction is obtained.

5 Energy e-commerce business model benefit calculation and distribution case analysis

If you want to ensure the applicability of the benefit distribution strategy by cooperative game presented by the project, this section takes the typical cooperative business model as the basis and analyzes the charging business in an electric vehicle business system in J Province.

5.1 Business model design

This chapter integrates and combines the business model of an electric vehicle charging business of an e-commerce company in J Province, as shown in Tables 2–6.

Typical business models of energy e-commerce in actual cases.

Major component Selection mode
Cooperative subject User
Target user C-end electric vehicle users
Critical business Electric vehicle charging
Profit model Charging income
Precision user management benefits (demand response)
Benefit distribution mode Multi-agent cooperative game

Among them, the benefits of refined user management are mainly obtained through user participation in demand response, and the relevant models are as follows:

5.1.1 Price package design

The price package is adjusted according to the customer’s sensitivity to price adjustments during different periods, and peak and off-peak time-based charging rates are implemented. Users proactively adjust their charging schedules upon receiving signals about changes in charging prices.

Based on the basic principle of designing the electricity price of peak–valley time-sharing charging by the economic principle, the demand price elasticity matrix model can be obtained: E i = e 11 i e 12 i . . . e 1 j i e 21 i e 22 i . . . e 2 j i . . . . . . . . . . . . e j 1 i e j 2 i . . . e j j i , where E i is the elastic matrix of the load demand and e j j i is the elastic coefficient of the load.

Among them, the load elastic coefficient model is e z j i = Δ Q z i / Q z i Δ P c z i / P c z i z = j Δ Q j i / Q j i Δ P c z i / P c z i z ≠ j .

In Eq. 15, when z = j , e z j i represents the self-elastic coefficient and when z ≠ j , e z j i represents the cross-elastic coefficient. Q z i and Q j i are the original loads at the time periods z and j , respectively. Δ Q z i and Δ Q j i are the user load variations at the time periods z and j , respectively.

The solution for the self-elasticity coefficient can be derived from the equilibrium of supply and elastic demand, while the specific processes are referenced by Yue (2016); the solution of the cross-elasticity coefficient is according to the multi-period price response process of users, and the specific process is referenced (Zhang et al., 2016).

Using the demand elasticity matrix, the calculation model of customer charging amount after the implementation of peak–valley time-sharing charging price is presented as follows (Victor Mayer Schonberg, 2013): Q 1 i * Q 2 i * . . . Q t i * = E ⋅ Q 1 i 0 . . . 0 0 Q 2 i . . . 0 . . . 0 . . . . . . 0 0 . . . Q t i ⋅ Δ P c 1 i / P c 1 i Δ P c 2 i / P c 2 i . . . Δ P c n i / P c t i + Q 1 i Q 2 i . . . Q t i , where Q t i and Q t i * refer to the user’s charging load demand before and after the implementation of t time-sharing charging, respectively. P c t i and Δ P c t i are the original price of the period load and the amount of price change, respectively.

 5.1.2 Incentive package design

Incentive package refers to the pre-incentive agreement established by energy e-commerce, which is composed of user load reduction amount and corresponding unit compensation price and signed with users (Zhong, 2014). Users have the option to establish a pre-incentive agreement with the energy supplier, enabling them to adjust their charging demand in accordance with the agreement within a specified time frame, including potential interruptions or increases in demand.

The stepped load reduction unit compensation price model is set as follows: P i , t , m I B D R = P i , t , 1 I B D R , Q i , ⁡ min ≤ Q i , t ≤ Q i , 1 P i , t , 2 I B D R , Q i , 1 < Q i , t ≤ Q i , 2 . . . P i , t , n I B D R , Q i , ⁡ max < Q i , t , where P i , t , m I B D R is the unit compensation price in the period m of t and Q i , t is the load that can be reduced in the period t . Q i , ⁡ min and Q i , ⁡ max are the minimum and maximum load reductions, respectively.

 5.2 Actual data of benefit measurement

Based on the electric vehicle charging demand of users covered by the e-commerce in J Province and charging price by the e-commerce, the elastic matrix of users’ charging demand is obtained through investigation, and the benefit calculation and benefit distribution of users’ participation in energy e-commerce cooperation are studied. The actual data are shown in Tables 3–6. The charging demand of users is described in Figure 3.

Charging demand of users.

Charging price list of electric vehicles of energy e-commerce.

Category Time segment Intra-time Segment price
Electrovalence Valley interval (23:00–7:00) 0.3 yuan/kW·h
Medium period (11:00–14:00) 0.575 yuan/kW·h
(18:00–23:00)
Peak hour (7:00–11:00) 0.85 yuan/kW·h
(14:00–18:00)

Elasticity of user charging demand.

Electric load
Self-elasticity Cross elasticity
Peak–peak −0.262 Peak–medium 0.1334
Medium–medium −0.116 Peak–valley 0.1065
- - Medium–peak 0.2685
- - Valley–peak 0.7894
- - Valley–medium 0.115
Valley–valley −0.0221 Medium–valley 0.0923

Parameters of the energy e-commerce design incentive package.

Load type Maximum load increase or decrease/MW·h Load increase or decrease/MW Load reduction compensation price/(yuan/kwh)
Electricity 6 (4.5,6] 0.18
(3,4.5] 0.12
(1.5, 3] 0.06
(0, 1.5] 0.03
(0, 1] 0.03

Changes in peak and valley time-sharing charging price and charging load.

Period Category Variation
23:00–6:00 Spot electricity price 0.29 yuan/kW·h
Electricity price increases and decreases −0.01 yuan/kW·h
(low load period) Increase in charging load in each period +1.331 MW
7:00–11:00 Spot electricity price 0.853 yuan/kW·h
15:00–19:00 Electricity price increases and decreases +0.003 yuan/kW·h
(peak load period) Increase in charging load in each period −1.699 MW
12:00–14:00 Spot electricity price 0.556 yuan/kW·h
20:00–22:00 Electricity price increases and decreases −0.019 yuan/kW·h
(medium load period) Increase in charging load in each period +1.071 MW
5.3 Calculation of business model benefit distribution 5.3.1 Setting multiple scenarios

Aiming to analyze the benefits generated by the cooperation alliance of users and e-commerce, four kinds of cases have been built as follows: case 1 is that e-commerce does not reach cooperation with customers; case 2 is that e-commerce reaches collaborating with users through price-based demand response; case 3 is that e-commerce reaches collaborating with users through incentive-based demand response; and case 4 is that agreement is reached with users when both incentives and prices are implemented.

 5.3.2 Demand response results of users participating in e-commerce cooperation alliances

Table 7 presents the variation in user charging load in each period under the implementation of the incentive package.

Changes in charging load after the implementation of incentive package.

Time period 22:00–2:00 (MW) 6:00–9:00 10:00
11:00–12:00 (MW) 13:00–19:00 (MW)
Interruption of electrical load in each period +6 −3 −1.5
5.3.3 Benefit calculation results under multiple scenarios

Table 8 shows the results of basic benefit distribution under different cases.

Economic benefits of all parties in the alliance under different cases.

Case Case 1 Case 2 Case 3 Case 4
Participant Category Benefit/10,000 yuan
Electricity supplier charging business income Total 51.348 52.105 52.46 54.572
User revenue Incentive package income 0 0 1.836 1.836
Cost reduction benefit 0 1.922 2.645 4.072
Total 0 1.922 4.481 5.908
5.3.4 Analysis of benefit distribution results

In scenario 4, a comprehensive cooperation agreement is reached between the user and energy e-commerce platform to achieve optimal economic benefits. To reflect the significant role of user participation in the design of e-commerce packages in enhancing the overall benefits of e-commerce, this section uses the fundamental Shapley value approach, the basic solidarity value method, and the Shapley value method based on the enhancement of the contribution factor to conduct a reallocation of benefits. Table 9 presents the returns accrued to both parties under the basic Shapley value method.

Returns of the two parties under the basic cooperative game method.

Main body Basic Shapley value method revenue/10,000 yuan Base solidarity value method
Revenue/10,000 yuan
55.914 43.077
E-commerce user 4.566 30.24

Based on the examination of Table 9, from the viewpoint of distribution outcomes, the fundamental solidarity value method does not adhere to the integrity principle and is, hence, inapplicable to the benefit distribution results. Under the basic Shapley value method, e-commerce earnings augment by 13,420 yuan, whereas the corresponding user earnings diminish by 13,420 yuan. This does not mirror the constructive role of user participation in energy consumption packages in augmenting the earnings of e-commerce charging businesses, necessitating further enhancement through the amalgamation of the improvement model. Table 10 presents the earning outcomes of the three parties under the enhanced Shapley value method.

Results of tripartite returns under the improved Shapley value method.

Main body Contribution rate Improved Shapley value method, revenue/10,000 yuan Increase or decrease compared to the original income, revenue/10,000 yuan
E-commerce user 0.484 53.604 −0.968
0.516 35.327 0.968

According to the analysis in Table 10, the results of benefit allocation are consistent with role positioning. Second, the user revenue increased by 0.968 million yuan. In addition, although the revenue of e-commerce decreased by 0.968 million yuan, its net income increased by 22,256 million yuan compared with not cooperating with users.

 6 Conclusion

Initially, this paper surveys the concept and developmental characteristics of energy e-commerce. Building upon this, the formation mechanism of the typical business model of energy e-commerce is proposed from the perspectives of collaborative entities, target users, key businesses, profit model, and benefit distribution. Typical operational business models such as intermediary, leading, and cooperative are constructed. Subsequently, with the aid of user portrait methods and k-means clustering methods, C-end users can be categorized into rational, high-end, interested, and guided users, while B-end large users can be classified as high-value, low-value, and ordinary large users, with differentiated business models designed for different user types. Finally, based on the cooperative business model, in conjunction with the electric vehicle charging demands of users covered by the e-commerce in Province J and the charging prices implemented by the e-commerce, the benefit calculation and benefit distribution of users participating in the energy e-commerce cooperation are studied. With charging business as the core, this paper designs the business operation mode of energy e-commerce charging business under the cooperation of e-commerce and users and redistributes the benefits of both through case analysis, verifying that the distribution model proposed by the project can be applied to the benefit distribution of multi-entity cooperation in energy e-commerce.

From the perspective of suggestions, the power grid enterprises should fully grasp the current trend of “Internet +” and “intelligent +,” explore the business model and profit model of market-oriented operation, integrate cutting-edge information technology and energy e-commerce businesses closely, form innovative products and technical achievements with market competitiveness, better meet the real and urgent needs of users, and further enhance the ability of innovation and development.

 Data availability statement

The original contributions presented in the study are included in the article/Supplementary Material; further inquiries can be directed to the corresponding author.

 Author contributions

BY: conceptualization, data curation, formal analysis, investigation, validation, writing–original draft, and writing–review and editing. SL: conceptualization, data curation, formal analysis, investigation, validation, writing–original draft, and writing–review and editing. HG: conceptualization, data curation, formal analysis, investigation, methodology, validation, writing–original draft, and writing–review and editing.

 Funding

The author(s) declare that financial support was received for the research, authorship, and/or publication of this article. This paper is supported by the Young Talents Project of State Grid Energy Research Institute Co., Ltd. (SGNY202114010, Research on the Construction and Evaluation of Digital New Infrastructure Operation Model Based on the Complex Network). The funder was not involved in the study design, collection, analysis, interpretation of data, the writing of this article, or the decision to submit it for publication.

 Conflict of interest

Authors BY, SL and HG were employed by State Grid Energy Research Institute Co., Ltd.

 Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors, and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

 References Amit R. Zott C. (2001). Value creation in e‐business. Strategic Manag. J. 22 (6‐7), 493–520. 10.1002/smj.187 Chen W. Wang J. Yu G. Chen J. Hu Y. (2022). Research on day-ahead transactions between multi-microgrid based on cooperative game model. Appl. Energy 316, 119106. 10.1016/j.apenergy.2022.119106 Cubukcu K. M. (2019). The problem of fair division of surplus development rights in redevelopment of urban areas: can the Shapley value help? Land Use Policy 91, 104320. 10.1016/j.landusepol.2019.104320 Fei W. Li D. F. Ye Y. F. (2018). An approach to computing interval-valued discounted Shapley values for a class of cooperative games under interval data. Int. J. General Syst. 47 (8), 794–808. 10.1080/03081079.2018.1523903 Gan C. Yu J. Zhao W. Fan Y. (2023). Big data industry development and carbon dioxide emissions: a quasi-natural experiment. J. Clean. Prod. 422, 138590. 10.1016/j.jclepro.2023.138590 Jiang J. Yao W. Xueyan L. (2021). Precise demand forecast analysis of new retail target products based on combination model. Open J. Bus. Manag. 09 (03), 1312–1324. 10.4236/ojbm.2021.93071 Li R. Rao J. Wan L. (2022). The digital economy, enterprise digital transformation, and enterprise innovation. Manag. Decis. Econ. 43 (7), 2875–2886. 10.1002/mde.3569 Lindgadt Z. Reeves M. Stalk G. (2009). Rusiness model innovation: when the game gets tough, change the game. Boston Consult. Group (9), 1–8. Mahadevan B. A. Framework for business model innovation. IMRC 2004 Conference, Bangalore, India, 2004. Osterwalder A. Pigneur Y. Tucci C. L. (2005). Clarifying business models: origins, present, and future of the concept. Commun. Assoc. Inf. Syst. 16 (1), 1. 10.17705/1cais.01601 Victor Mayer Schonberg (2013). Big data era: great changes in life, work and thinking. Hangzhou, China: Zhejiang People's Publishing House. Wang K. Yu J. Yu Y. Qian Y. Zeng D. Guo S. (2017). A survey on energy internet: architecture, approach, and emerging technologies. IEEE Syst. J. 12 (3), 2403–2416. 10.1109/jsyst.2016.2639820 Wang X. Ng C. T. (2020a). New retail versus traditional retail in e-commerce: channel establishment, price competition, and consumer recognition. Ann. Oper. Res. 291, 921–937. 10.1007/s10479-018-2994-9 Wang Y. Chang J. (2021). Future development trend of “New Retail” and e-commerce based on big data. J. Phys. Conf. Ser. 1852 (3), 032029. 10.1088/1742-6596/1852/3/032029 Wang Y. Cheng D. Liu X. (2019). Matrix expression of Shapley values and its application to distributed resource allocation. Sci. China Inf. Sci. 62 (2), 22201. 10.1007/s11432-018-9414-5 Wang Y. Fan R. Shen L. Jin M. (2020b). Decisions and coordination of green e-commerce supply chain considering green manufacturer's fairness concerns. Int. J. Prod. Res. 58 (24), 7471–7489. 10.1080/00207543.2020.1765040 Williams E. Tagami T. (2002). Energy use in sales and distribution via e‐commerce and conventional retail: a case study of the Japanese book sector. J. Industrial Ecol. 6 (2), 99–114. 10.1162/108819802763471816 Yue W. (2016). The application research on tourism enterprise implementing precision marketing in the era of high-speed rail. Proceedings of 2015 2nd International Conference on Industrial Economics System and Industrial Security Engineering. Springer Singapore Zhang H. Zhang L. Cheng X. Chen W. (2016). A novel precision marketing model based on telecom big data analysis for luxury cars. International Symposium on Communications and Information Technologies. IEEE Qingdao, China Zhang Y. (2021). The study on the new retail’s emergence and development in China. Acad. J. Eng. Technol. Sci. 4 (3). Zhong W. (2014). Regulation of personal data privacy in the era of big data. Beijing, China: Social Sciences Academic Press.