Vehicle-to-vehicle area division charging matching method based on cloud and fog computing

Abstract

The invention discloses a vehicle-to-vehicle area division charging matching method based on fog and cloud computing, and belongs to the field of vehicle networking electric vehicle charging. The method comprises the following five stages: in the first stage, traffic sub-regions are divided according to the spatial correlation degree of road intersections, the whole traffic network is divided into a plurality of traffic sub-regions, and each fog server is responsible for managing one traffic sub-region; in the second stage, a region weight model is constructed, and weight values are given to all the sub-regions so as to determine the proximity degree between the regions. And in the third stage, on the basis of the regional weight model, a cost profit model is constructed according to factors such as the residual electric quantity, the distance and the price of the electric vehicle. And the fourth stage is to construct a reputation model which can be evaluated by consumers and suppliers of the electric vehicles so as to screen and eliminate the users with too low reputation scores. And in the fifth stage, an improved stable matching model is constructed according to the reputation model and the cost profit model.

Description

Technical Field

[0001] The present invention belongs to the field of Internet of Vehicles, and more specifically, relates to a vehicle-to-vehicle area division charging matching method based on cloud and fog computing. Background Art

[0002] As the global warming problem becomes more and more serious, green and low-carbon transformation has also become an important development direction. Compared with fuel vehicles, electric vehicles use clean energy to achieve zero greenhouse gas emissions and reduce dependence on fossil fuels. Therefore, electric vehicles are considered to be candidates for completely replacing internal combustion engine vehicles. However, there are still some challenges to be solved in order to popularize electric vehicles. First, the large-scale popularization of electric vehicles will lead to the problem of grid overload, because with the rapid growth of the number of electric vehicles, their charging behavior may cause congestion, fluctuations and extreme overload peaks in the distribution network. Secondly, since the growth rate of charging piles is much lower than that of electric vehicles, it is difficult to charge electric vehicles and the queues are long. This low growth rate is difficult to alleviate in a short time, because the high initial installation cost and the choice of suitable sites hinder the popularization of charging stations. In addition, the "range anxiety" problem (worrying that electric vehicles do not have enough battery power to reach the intended destination) is also an important reason affecting the large-scale popularization of electric vehicles, and range anxiety seriously affects the experience of electric vehicle drivers. Although there are some optimization methods for charging and energy management of different electric vehicles based on cloud fog computing, most of these solutions are limited to specific scenarios or urban hot spots. For example, most schemes assume that there are enough energy demanders and providers, but this is too idealistic and has certain limitations. Summary of the invention

[0003] The purpose of the present invention is to solve the problems existing in the prior art and provide a car-to-car regional division charging matching solution based on cloud and fog computing. The solution takes regional division into consideration in car-to-car charging matching and can perform pricing based on the distance between regions. It is not limited to a specific area, so that the car-to-car charging matching method can be applied to large areas and various traffic scenarios.

[0004] To achieve the above object, the technical solution of the present invention is as follows:

[0005] Step 1: Area division stage: In this stage, traffic sub-areas are divided according to the spatial correlation of road intersections. The entire traffic network is divided into several traffic sub-areas, and each fog server is responsible for managing a traffic sub-area.

[0006] Step 2: Assigning regional weights: In this stage, weight values ​​are assigned between each sub-region to determine the degree of proximity between regions;

[0007] Step 3: User cost-profit construction stage: In this stage, a user cost-profit model is constructed based on the remaining power, distance, price and other factors of the electric vehicle on the basis of determining the regional weight;

[0008] Step 4: User reputation model: In this stage, a reputation model that can be evaluated by both electric vehicle consumers and suppliers is constructed to screen out users with low reputation scores;

[0009] Step 5: Car-to-car matching phase: This phase combines the previously proposed model and uses the improved Gale-Shaply matching algorithm to pair consumers and suppliers for car-to-car charging.

[0010] The specific process of the present invention is described as follows:

[0011] Step 1 includes:

[0012] (1) First, the key intersections are determined according to the importance of the intersection nodes, where the importance of the intersection node Imd(n i ) is defined as the road network node degree and Saturation The sum of the road network node degree It refers to the number of edges (or road segments, connecting lines) directly connected to a node in the road network. The larger the value, the greater the intersection n i The higher the importance, the higher the road network saturation. It refers to the ratio between the actual traffic volume of the road in the road network and the designed traffic capacity of the road. The larger the value, the higher the intersection n i The higher the importance.

[0013]

[0014] (2) Then, the preliminary road network division based on the key intersections can make the traffic flow evenly distributed in the road network. We regard each key intersection as a cluster set, and then traverse each non-key intersection, and use the total intersection correlation to calculate the total intersection correlation value between the key intersection and the non-key intersection. Each non-key intersection is selected to join the cluster set of the key intersection with the highest total intersection correlation value until each non-key intersection is assigned. Intersection cycle correlation coefficient Intersection spacing correlation coefficient and intersection density correlation coefficient The three factors are jointly determined, and since their influence on the total correlation of the intersection is different, different weight coefficients are given to them.

[0015]

[0016] (3) Finally, since the cluster sets obtained by the preliminary division may have close connections between them, we use the Newman partition algorithm to merge the closely connected cluster sets. First, the road network after the preliminary division is abstracted into a topological structure graph, in which the cluster sets are nodes and the roads connecting the clusters are edges. Each node is regarded as an independent community. Then the Newman partition algorithm is used to merge the communities. Its basic idea is to proceed in the direction where the modularity increment ΔQ increases the most before and after the community merger. When all nodes are merged into one community, the algorithm terminates. When the modularity Q takes the maximum value, the community partition result reaches the optimal result and can be used as the final result. The final area partition set is set as Area={A1,A2,…,A x}.

[0017]

[0018] ΔQ=e i,j +e j,i -2g i g j

[0019] Step 2 includes:

[0020] First, we traverse each area. When we traverse to a certain area, we determine the number of layers of other areas with this area as the center. Then, based on the number of layers, we use the weight assignment formula to assign values ​​to each area. Finally, the calculated weight values ​​are stored in the weight matrix for easy calling by the cloud center. 1) If an area is directly adjacent to this area (there are no other areas between the two areas), this area is divided into the first layer; 2) If an area is secondarily adjacent to this area (directly adjacent to the area of ​​the first layer), this area is divided into the second layer; 3) And so on, until all other areas are divided into layers with this area as the center; 4) Finally, the weight value centered on this area is set according to the divided level. The final weight value is stored in the weight matrix WA for easy calling by the cloud center.

[0021]

[0022] The specific steps of the hierarchical partitioning algorithm are as follows:

[0023] (1) Enter a region A i All regions are based on the final region set obtained in the previous section: Area = {A1, A2, ..., A x}, so the value of i is [1,x];

[0024] (2) Circular calculation A i With other areas A j The layer relationship of j is [1,x];

[0025] (3) If the binary variable a i,j =1 and A i With A j There is no intermediate area between them, and the area is divided into the first layer, namely c i,j =1. The binary variable is defined as the i With A j Whether it is connected, if it is connected, its value is equal to 1, otherwise it is equal to 0.

[0026] (4) If the binary variable a i,j =1 and A i With A j There is an intermediate area between them, so we discuss it by category.

[0027] 1) If A i With A j There are only m intermediate areas between them, then the area is divided into the m+1th layer, that is, c i,j =m+1.

[0028] 2) If A i With A j If there are m or n intermediate regions between them, then their minimum value is taken, so the area is divided into the min(m,n)+1th layer, that is, c i,j =min(m,n)+1.

[0029] (5) If none of the above conditions are met, the loop continues to determine the next area.

[0030] (6) Finally output A i The layer value C corresponding to the region and all other regions i ={c i,j}

[0031] Step 3 includes:

[0032] (1) First, the peak-valley electricity price function FG(t) is used to divide the electricity price of the day. The basic price of electric vehicle charging is determined by the peak-valley electricity price function.

[0033]

[0034] (2) Then, based on the peak and valley electricity price function, a cost-profit function is constructed. The amount that consumers should pay It consists of three parts: the cost of charging from the supplier Supplier's operating costs And additional costs Supplier's operating costs It consists of three parts: the travel cost required to reach the meeting point The time cost of car-to-car charging and battery degradation costs

[0035]

[0036] (3) Finally, the cost-profit function and regional weights are combined to construct the consumer and supplier utility functions. It mainly includes three parts: the travel cost required for consumers to reach the meeting point The total time cost of consumers in the process of car-to-car charging and the amount that consumers should pay to suppliers The regional problem is also considered, and the weight value A is called from the regional matrix WA according to different regions. i,j , and its weight value A i,j The consumer cost formula is multiplied. The farther apart the regions are, the higher the corresponding cost. The supplier utility function The amount paid by the consumer minus the supplier's initial charging cost and supplier's operating cost plus the platform's incentive amount Jm. Similarly, the incentive amount in different regions is different. The weight value A is called from the regional matrix WA according to different regions. i,j , and its weight value A i,j Multiply it by the supplier's profit formula. The farther apart the regions are, the higher the corresponding profit for the supplier.

[0037]

[0038] Step 4 includes:

[0039] First, we build a word-of-mouth evaluation system, which is jointly evaluated by the platform and users. Consumers and suppliers rate each other, which can ensure the fairness and rationality of the evaluation to a certain extent. Then we build a flexible reputation model, which combines the current rating with the historical rating, so that the reputation score will not be too low due to a malicious evaluation. Finally, we combine the word-of-mouth evaluation system with the flexible reputation model to build a comprehensive consumer reputation model. and supplier comprehensive reputation model

[0040]

[0041] Step 5 includes:

[0042] The specific steps of the vehicle-to-vehicle matching algorithm of the present invention are as follows:

[0043] (1) First, the entire traffic network is divided into several traffic sub-areas according to the area partition model. Each fog server is responsible for managing a traffic sub-area. The traffic sub-area set Area = {A1, A2, …, A x}.

[0044] (2) Then, according to the traffic sub-area obtained by division, Area = {A1, A2, ..., A x}, the layer number relationship between each region is determined by using the regional layer number determination algorithm. And each region is assigned a value based on the regional weight formula to obtain a regional weight matrix WA storing the weight values ​​between each region.

[0045] (3) Then, a cost-benefit function is constructed based on factors such as the energy transaction volume, distance, and speed between consumers and suppliers. And the corresponding regional weight value A is obtained from the regional weight matrix WA. i,j , the regional weight value A i,j Combined with the cost-profit function, construct the consumer utility function and the supplier utility function

[0046] (4) The fog server then calculates the reputation value of the corresponding user based on the comprehensive reputation model. The reputation value must be greater than the minimum reputation value threshold preset by the platform. To participate in the matching.

[0047] (5) Finally, the fog server calculates the utility functions of consumers and suppliers and sorts them to obtain the preference lists of both parties. Then, based on the preference lists, the cloud center uses the improved Gale-Shaply matching algorithm to obtain the final matching pairs of all users (Note: there are cases where users cannot be matched successfully).

[0048] The above technical solution conceived by the present invention has the following beneficial effects compared with the prior art: the solution can divide the urban road network into various sub-areas, and take the regional division into consideration in the vehicle-to-vehicle charging matching cost. That is, the cost of electric vehicle users in the same area is lower, and the longer the distance between electric vehicle users in different regions, the higher the cost; in addition, the proposed solution also takes into account the peak and valley electricity prices, the charging price is high during the peak charging period, and the charging price is low during the valley charging period, which can well alleviate the load pressure of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 This is a framework diagram of the vehicle-to-vehicle charging matching model of the electric vehicle, roadside unit, fog server, and cloud center combined with the present invention.

[0050] Figure 2 It is a flow chart of a vehicle-to-vehicle area division charging matching method based on fog and cloud computing of the present invention.

[0051] Figure 3 It is a schematic diagram of a vehicle-to-vehicle area division charging matching method based on fog and cloud computing of the present invention. DETAILED DESCRIPTION

[0052] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present invention;

[0053] See also Figure 1 The figure shows a three-layer architecture diagram of a vehicle-to-vehicle charging system based on fog computing and cloud computing, which includes RSU (roadside unit), electric vehicle, fog server and cloud center;

[0054] Roadside Unit (RSU): refers to an intelligent device installed on the roadside that can communicate with vehicles in both directions. It is mainly responsible for collecting information about the purchase or sale of energy by electric vehicles and uploading the information to the fog server for processing in real time.

[0055] Electric vehicles: They are responsible for uploading their battery status, remaining power, charging requirements and other information to the RSU, which is broadcast or shared with other vehicles. Electric vehicles are divided into three categories: energy suppliers (electric vehicle users who are willing to sell excess energy), energy consumers (electric vehicle users whose energy is too low and need to buy energy), and ordinary electric vehicles that are not interested. Users choose their roles based on their travel plans and energy status, and decide whether to participate in energy trading.

[0056] Fog server: refers to a server deployed in the fog layer with data storage, processing, communication and other capabilities. It is responsible for processing real-time computing tasks related to charging, such as charging demand, charging power, etc.

[0057] Cloud center: refers to a server deployed in the cloud layer with data storage, processing, communication and other capabilities. It is mainly responsible for information interaction with the fog server and handling global vehicle matching tasks.

[0058] See also Figure 2 The figure shows a process of a vehicle-to-vehicle area division charging matching method based on fog and cloud computing, including the following steps:

[0059] S101: Construct a collaborative matching model among vehicles, roadside units, and fog servers and cloud centers; divide traffic sub-areas based on the spatial correlation of road intersections, and divide the entire traffic network into several traffic sub-areas. Each fog server is responsible for managing a traffic sub-area.

[0060] S102: Based on the regional division, weights are assigned to the divided regions, and the distance between each region and other regions is obtained through the weight values;

[0061] S103: constructing a cost-profit function based on factors such as the remaining power of the electric vehicle, distance, price, etc., and taking the regional weight into account to construct the consumer and supplier utility functions;

[0062] S104: Build a reputation model that can be evaluated by both electric vehicle consumers and suppliers to screen out users with low reputation scores;

[0063] S105: A stable matching algorithm is used to obtain a vehicle-to-vehicle matching result using the utility functions of the consumer and the supplier as preference values.

[0064] See also Figure 3 The figure shows a vehicle-to-vehicle matching step of a vehicle-to-vehicle area division charging matching method based on fog and cloud computing, including:

[0065] In the first round, consumers send invitations to suppliers ranked first in their preference lists. Suppliers s3 and s4 receive multiple invitations, so they make a reverse selection, selecting consumers with higher preference lists and rejecting consumers with lower preference lists. This results in the initial consumer-supplier matching pairs.

[0066] In the second round, the rejected c3 and c5 send invitations to the supplier ranked second in their preference list. Supplier s2 makes a counter-selection, because c3 is better than c1, so it rejects c1 and chooses c3. Since c5 is not in s3's preference list, s3 rejects c5's invitation. And because c5 has only two preference lists, c5 is eliminated. Get the updated consumer-supplier matching pair;

[0067] In the third round, the rejected c1 sends an invitation to the supplier ranked second in the preference list. S1 is idle and has no more invitations. They are successfully matched and the final consumer-supplier matching pair is obtained.

[0068] Each embodiment in this specification is described in a related manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments; in particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0069] The above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention are included in the protection scope of the present invention.

Claims

1. A vehicle-to-vehicle area division charging matching method based on fog and cloud computing, characterized in that: include: Build a collaborative system model of electric vehicles, roadside units, fog servers, and cloud centers; Divide the traffic sub-areas based on the spatial correlation of road intersections, and divide the entire traffic network into several traffic sub-areas. Each fog server is responsible for managing a traffic sub-area. Assign weights to the divided areas, and use the weights to get the distance between each area and other areas; Construct a cost-profit function based on factors such as the remaining power of electric vehicles, distance, and price, and consider regional weights in the cost-profit function to construct consumer and supplier utility functions; Build a reputation model that can be evaluated by both electric vehicle consumers and suppliers to screen out users with low reputation scores; A stable matching algorithm is used to obtain the car-to-car matching results with the utility functions of consumers and suppliers as preference values.

2. A vehicle-to-vehicle area division charging matching method based on fog and cloud computing according to claim 1, characterized in that: Building a collaborative system model of electric vehicles, roadside units, fog servers, and cloud centers includes: RSU roadside units are deployed on the roadside and deployed at a certain distance to achieve full road coverage; fog servers are deployed on base stations, and the coverage areas of each fog server do not overlap; the cloud center is deployed in the traffic dispatch center, responsible for the global vehicle-to-vehicle matching and dispatching.

3. A vehicle-to-vehicle area division charging matching method based on fog and cloud computing according to claim 1, characterized in that: Divide the traffic sub-areas based on the spatial correlation of road intersections. The entire traffic network is divided into several traffic sub-areas. Each fog server is responsible for managing a traffic sub-area. The intersections are set as point set N and the roads are set as edge set L. Then, the key intersections in the urban road network are determined according to the importance of the intersection nodes. Then, the non-critical intersections are traversed, and each non-critical intersection is merged with the critical intersection with the largest intersection correlation value to form a small cluster; the intersection correlation is determined by the intersection period correlation, intersection spacing correlation and intersection density correlation; finally, the modularity increment between each small cluster is calculated, and the two small clusters with the largest modularity increment greater than zero are merged until the movement of clusters in the urban road network no longer leads to an increase in modularity gain, and the final area partition set is set as Area = {A1, A2, ..., A x }.

4. A vehicle-to-vehicle area division charging matching method based on fog and cloud computing according to claim 1, characterized in that: Assign weights to the divided areas, and use the weights to get the distance between each area and other areas: if an area is directly adjacent to this area (there is no other area between the two areas), this area is divided into the first layer; if an area is secondarily adjacent to this area (directly adjacent to the area in the first layer), this area is divided into the second layer; 3) and so on, until all other areas are divided into layers with this area as the center; 4) Finally, set the weight value centered on this area according to the divided level; the final weight value is stored in the weight matrix WA for easy calling by the cloud center; the specific algorithm steps for hierarchical division are as follows: Step 1: Enter a region A i All regions are based on the final region set obtained in the previous section: Area = {A1, A2, ..., A x }, so the value of i is [1,x]; Step 2: Loop calculation A i With other areas A j The layer relationship of j is [1,x]; Step 3: If the binary variable a i,j =1 and A i With A j There is no intermediate area between them, and the area is divided into the first layer, namely c i,j =1; in The binary variable is defined as the judgment A i With A j Whether it is connected, as shown in Formula 7, if it is connected, its value is equal to 1, otherwise it is equal to 0; Step 4: If the binary variable a i,j =1 and A i With A j There is an intermediate area between them, so we will discuss it by category. 1) If A i With A j If there are only m intermediate regions between them, the region is divided into the m+1th layer. That is c i,j =m+1; 2) If A i With A j If there are m or n intermediate regions between them, then their minimum value is taken, so the area is divided into the min(m,n)+1th layer, that is, c i,j =min(m,n)+1; Step 5: If none of the above conditions are met, continue to loop and judge the next area; Step 6: Finally output A i The layer value C corresponding to the region and all other regions i ={c i,j }.

5. A vehicle-to-vehicle area division charging matching method based on fog and cloud computing according to claim 1, characterized in that: The cost-profit function is constructed based on factors such as the remaining power, distance, and price of electric vehicles, and the regional weight is taken into account in the cost-profit function to construct the consumer and supplier utility functions: First, the peak-valley electricity price model is used to divide the daily electricity price. The basic price of electric vehicle charging is determined by the peak-valley electricity price model; then, based on the peak-valley electricity price model, we construct the consumer cost and supplier profit function; finally, the cost-profit function is combined with the regional weight model to construct the consumer and supplier utility function; the consumer utility function It mainly includes three parts: the travel cost required for consumers to reach the meeting point The total time cost of consumers in the process of car-to-car charging and the amount that consumers should pay to suppliers The regional problem is also considered, and the weight value A is called from the regional matrix WA according to different regions. i,j , and its weight value A i,j The consumer cost formula is multiplied; the farther apart the regions are, the higher the corresponding cost is, and the supplier utility function is The amount paid by the consumer minus the supplier's initial charging cost and supplier's operating cost plus the platform's incentive amount Jm. Similarly, the incentive amount in different regions is different; according to different regions, the weight value A is called from the regional matrix WA i,j , and its weight value A i,j Multiply by the supplier benefit formula; the farther apart the regions are, the higher the corresponding benefits to the suppliers.

6. A vehicle-to-vehicle area division charging matching method based on fog and cloud computing according to claim 1, characterized in that: Build a reputation model that can be evaluated by both electric vehicle consumers and suppliers to screen out users with low reputation scores: the word-of-mouth evaluation system we set up is jointly evaluated by the platform and users, and consumers and suppliers rate each other, which can ensure the fairness and rationality of the evaluation to a certain extent; our elastic reputation model combines the current score with the historical score, and will not cause the reputation score to be too low due to a malicious evaluation; finally, combine the word-of-mouth evaluation system with the elastic reputation model to build a comprehensive reputation model; we use fog servers to calculate the reputation score of each user (including consumers and suppliers), which is used to evaluate whether the user can participate in car-to-car charging matching, and the calculated reputation values ​​are uploaded to the cloud center through the fog server for storage.

7. A vehicle-to-vehicle area division charging matching method based on fog and cloud computing according to claim 1, characterized in that: The use of a stable matching algorithm to optimize vehicle-to-vehicle charging includes: Step 1: Output the matching result binary table Initialize; Step 2: Enter the electric vehicle consumer collection and electric vehicle suppliers, Step 3: Select a DV that has not been matched and has not been used by any supplier j The CV that initiated the match request i Enter the matching process; Step 4: CV i The DV that is ranked first in the preference list and has not yet been matched. j Make matching condition judgment; Step 5: If DV j If there is no pairing, then they are matched successfully; if DV j There is a match, but it is different from the current CV i More suitable, then DV j Will abandon the original consumers and CV i Make a match; Except for the above two cases, DV j will reject the matching request; if there is still CV i If no match is found, jump to step 2; Step 6: When each CV i After all the matching processes are completed, the vehicle-to-vehicle matching process is completed; Finally, we get a set of car-to-car matches (consumers and their corresponding suppliers)

Images