CN114338685A - Edge server resource allocation method based on credit-price relationship - Google Patents

Abstract

The invention requests to protect an edge server resource allocation method based on credit-price relationship, which comprises the following main steps: s1, acquiring the resource number, the credit and the asking price of the edge server, and acquiring the bid price and the required credit of the user; s2, sorting the edge server and the user according to the credit rating model; s3, matching the user and the edge server according to the sorting result, the price constraint and the credit constraint; s4, judging whether the edge server and the user can carry out price dynamic update, if so, updating the price and then continuing matching; s5, screening the matching result, and screening repeated user requests; s6, calculating a transaction price and transaction utility according to the final successful matching result; and S7, updating the credit of the edge server according to the credit evaluation model. The invention provides a credit degree-price relation and provides a dynamic price updating mechanism based on the credit degree-price relation, thereby integrally improving the reliability of resource allocation and the resource utilization rate.

Description

Edge server resource allocation method based on credit-price relationship

technical field

The invention belongs to the field of mobile edge computing, in particular to an edge server resource allocation method based on a credit-price relationship.

Background technique

The proliferation of users, driven by ubiquitous wireless communication networks, has led to an ever-increasing amount of information and data, but most current users are limited in size, resulting in limited computing power and battery life. The powerful computing and storage capabilities of the cloud server enable the tasks generated by the users to be effectively processed and bring more effective services to the end users. However, with the advancement of 5G and the development of Internet and computer technology, users are showing an explosive growth trend. The number of users connected to remote cloud servers has also increased dramatically, and the demand for high-quality services has put enormous pressure on cloud servers. In addition, since the cloud server is deployed in the core network center very far away from the user, the completion of the data task generated by the user will lead to greater delay and more energy consumption, and the long-distance transmission delay will greatly reduce the operating efficiency of the system . Therefore, the traditional centralized service model based on cloud computing has been difficult to meet the current business needs.

Mobile edge computing (MEC) is an emerging computing paradigm. Edge servers process and store data at the edge of the network close to user terminal devices. It is a computing model that provides reliable and stable services to users nearby. Since edge servers are deployed closer to users, they can directly provide more efficient services to ensure lower latency and avoid uploading all data to the cloud when the server processes data locally, thereby reducing bandwidth pressure and being widely distributed The edge server reduces the energy consumption of the cloud server to a certain extent. Therefore, mobile edge computing can be introduced to process data tasks generated by users, and more effectively handle the task requirements generated by users at a location closer to the user. However, although the application of mobile edge computing can achieve better results, it also faces some challenges, such as limited edge server resources, and inconsistent interests of users and edge servers, resulting in profit-driven resources cannot be well allocated and use. Auctions are currently a popular form of transaction that efficiently allocates sellers’ resources to buyers at competitive prices. However, there are inevitably some dishonest participants in the cloud market, or due to network dynamics, resource heterogeneity, and untrustworthy edge servers, which lead to unreasonable resource allocation for edge servers when faced with a large number of task demand requests or resource allocation failed. Therefore, how to design a resource allocation mechanism that can improve the resource utilization of edge servers, ensure the reliability of the entire edge computing resource allocation process, and maximize the utility of edge servers and users while meeting user needs is still a very important issue. Questions worth investigating.

Faced with the limited resources of edge servers, but the resources cannot be fully utilized, the resource allocation process will be interrupted due to network dynamics, resource heterogeneity, and server untrustworthiness, and the interests of both parties will be inconsistent. An effective and reliable resource is needed. Allocate methods to handle highly concurrent edge tasks and heterogeneous edge resources.

Since the tasks offloaded to the edge server are all generated due to the insufficient storage and computing power of the mobile terminal, and in the case of limited resources, each requesting user hopes to obtain reliable edge server resources to more effectively Process tasks. If in the process of resource matching, the credit factor of the edge server is considered, and the resource request amount of all users requesting a server resource and the resource ownership of the edge server are compared to make a matching selection, which can not only improve the reliability of allocation It can also improve resource utilization and successful transaction volume of the entire cloud market.

After retrieval, the publication number is CN107995660B, a joint task scheduling and resource allocation method supporting D2D-edge server offloading, characterized in that: the method includes the following steps: S1: modeling user joint overhead; S2: modeling user task execution required Delay; S3: Modeling the energy consumption required for the execution of user tasks; S4: Modeling user task scheduling and resource allocation constraints; S5: Under the conditions of task scheduling and resource allocation, determine user tasks based on the minimization of user joint overhead Scheduling and resource allocation strategy; the S1 is specifically: modeling the joint overhead of users according to the formula is the sum of the overheads of all users in the network performing tasks, where is the overhead required for the i-th user to perform tasks, 1≤i≤N, N is the number of users to perform tasks in the network; modeled as where, t represents the delay required for the ith user to perform tasks, e represents the energy consumption required for the ith user ii to perform tasks, and represents the delay overhead of the ith user The weighting coefficient of , which represents the weighting coefficient of the i-th user's energy consumption overhead.

The differences between the invention patent whose publication number is CN107995660B and the present invention are as follows:

1. What this invention considers is task offloading, and what this invention considers is resource allocation. Although the overall concept is not very different, the actual content is very different due to different research directions and research methods;

2. The invention aims to minimize joint overhead such as delay and energy consumption, while this study focuses on economic benefits, takes credit and price factors as the goal, and considers the reliability of resource allocation and resource waste;

3. In the present invention, the situation of unreliable nodes is overcome, and the situation of inaction after the edge server obtains profit is avoided as much as possible;

4. The present invention proposes a price dynamic update mechanism, which overcomes the waste of resources as much as possible by improving the algorithm.

SUMMARY OF THE INVENTION

The present invention aims to solve the above problems of the prior art. An edge server resource allocation method based on the credit-price relationship is proposed. The technical scheme of the present invention is as follows:

An edge server resource allocation method based on a credit-price relationship, comprising the following steps:

S1, obtain the number of edge server resources, asking price and credit, obtain the user's bid and the required credit; S2, sort the edge servers and users according to the credit evaluation model; S3, according to the sorting result, price constraints and credit constraints, etc. Match the resources owned by the user and the edge server; S4, determine whether the edge server and the user can dynamically update the price, and if so, update the price and continue matching; S5, filter the matching results, and filter duplicate user requests; S6, Calculate the transaction price and transaction utility according to the final successful matching result; S7, update the credit degree of the edge server according to the credit degree evaluation model.

Further, in the S1, obtain the number of edge server resources, the asking price and the credit rating, and obtain the user's bid and the required credit rating, specifically including:

否则直接获取边缘服务器的要价与用户的出价。Obtain the information of the edge server and the user. If it is the first time to participate in the matching, initialize the price between the edge server and the user according to the credit-price relationship mechanism in the credit evaluation model, and calculate the initial asking price of the edge server i. Ask _i and user j are paired Initial request price for edge server i

Otherwise, directly obtain the asking price of the edge server and the user's bid.

Further, the steps of initializing the asking price of the edge server and the bidding of the user according to the credit-price relationship mechanism are as follows:

(1) First, initialize the asking price AskS _i of the edge server to bid with the user according to the following formula

AskS _i = Cre _i *10 (1)

AskS_i表示边缘服务器i的初始要价，Cre_i表示边缘服务器i的初始信用度；

表示用户j对边缘服务器i的初始竞价，

表示用户j对边缘服务器i的信用度要求；where Cre _i ∈ [0,1],

AskS _i represents the initial asking price of edge server i, and Cre _i represents the initial credit of edge server i;

represents the initial bid by user j for edge server i,

Represents the credit requirement of user j to edge server i;

(2) According to the following formula, randomly generate the final edge server asking price and user bidding;

Ask _i =AskS _i +Random(-a,a) (3)

表示用户j对边缘服务器i的最终竞价，而Random(-a,a)表示随机在[-a,a]的区间中生成一个随机数，同理Random(0,b)表示在[0,b]区间生成一个随机数；Among them, Ask _i represents the initial final price of edge server i,

Represents the final bid by user j for edge server i, and Random(-a,a) represents a random number generated in the interval of [-a,a]. Similarly, Random(0,b) represents in [0,b] ] interval to generate a random number;

得初始竞价

然后根据Random(0,b)进行随机价格波动；最终得到的

表示本次匹配时用户j对边缘服务器i的最终竞价。In each subsequent iteration, according to the credit-price relationship, for the edge server, an initial price AskS _i is first generated according to the current credit value of the edge server; then, random price fluctuations are performed according to Random(-a, a), and finally The generated Ask _i is the final asking price of the edge server in this match; for the user, it is similar to the initial asking price; first, a user j's credit requirement value for the edge server i is randomly generated

initial bid

Then perform random price fluctuations according to Random(0,b); the final obtained

Indicates the final bid by user j for edge server i in this match.

也表示在进行匹配时，边缘服务器i的信用度Cre_i需要大于用户j对边缘服务器i的信用度要求

在最开始初始化时，设定每个边缘服务器的信用度值为0.5，而对于用户j对边缘服务器i的信用度要求

为了更符合实际生活中的随机情况，则是根据如下公式进行设置。即随机生成0到1中的一个小数即为

的初始值。Further, the said

It also means that when matching, the credibility Cre _i of the edge server i needs to be greater than the credibility requirement of the user j for the edge server i

At the initial initialization, the credit value of each edge server is set to 0.5, and user j requires the credit of edge server i

In order to be more in line with the random situation in real life, it is set according to the following formula. That is, a decimal number from 0 to 1 is randomly generated, which is

the initial value of .

对用户的请求进行降序排序。Rank_i，

定义如式所示。Further, the method for allocating edge server resources based on the credit degree-price relationship is characterized in that the step S2 sorts the edge servers and users according to the credit degree evaluation model, which specifically includes: according to the credit degree evaluation model, sorting the edge servers and users according to the credit degree evaluation model. Users are prioritized. For edge servers, sort them in ascending order by judging the size of Rank _i . For users, according to

Sort user requests in descending order. Rank _i ,

The definition is as shown in the formula.

Rank _i =Ask _i /Cre _i (6)

表示用户j对边缘服务器i的信用度要求。Among them, Cre _i represents the initial credit of edge server i,

Represents the credit requirement of user j to edge server i.

Further, the step S3 performs matching according to the sorting result, the credit degree required by the user and the resources possessed by the edge server, which specifically includes: matching the user with the edge server according to price constraints and credit constraints, etc., if the resources of the edge server are insufficient during matching. , then it cannot be matched. Since each user may bid on a different edge server, a user may appear multiple times, and an edge server may appear multiple times.

Further, in the step S4, it is determined whether the edge server and the user can dynamically update the price, and if it can be updated, the price will be updated and then continue to match, which specifically includes: on the basis of the proposed credit-price relationship, propose a dynamic price update mechanism. ; The formula of the dynamic price update mechanism is as follows:

Ask' _i =Ask _i -a' (8)

为更新后用户请求价格，b'为用户竞价更新值。对于用户，在区间

内进行价格向上迭代，迭代至

为价格不再迭代边界；对于边缘服务器，在区间[Ask'_i-a,Ask'_i+a]内进行价格向下迭代，迭代至Ask'_i-a为价格不再迭代边界。where Ask' _i is the updated asking price, and a' is the updated value of the edge server asking price;

Requested price for the updated user, b' is the updated value for the user's bid. For users, in the interval

price upward iteration within

No more iterative boundaries for prices; for edge servers, iterate downwards for prices within the interval [Ask' _i -a,Ask' _i +a], and iterate to Ask' _i -a for no more iteration boundaries for prices.

Further, in the step S5, the matching results are screened, and repeated user requests are screened, which specifically includes: after the matching is completed, the matching results need to be screened; because a user can only request one edge server at the same time, but one edge server At the same time, the server can process requests from multiple users with sufficient resources; because it is necessary to filter out the repeated requests from users, the filtering rule is to discard user requests with lower utility.

Further, in the step S6, after the final matching is completed, the transaction price and utility are calculated according to the prices of both parties, and the calculation formula is as shown in the formula.

V = V _j +V _i (14)

代表的是边缘服务器i跟用户j匹配成功时所取得交易价格，Vi^j表示边缘服务器i与用户j匹配成功时边缘服务器的效用；V_i表示为本次匹配边缘服务器i的总效用，因为边缘服务器i存在提供服务给多个用户，所以边缘服务器i总效用V_i应该为V_i ^j累加；而V_j则表示用户j本次匹配的总效用，V则是单次匹配总效用。in

Represents the transaction price obtained when edge server i is successfully matched with user j, Vi ^j represents the utility of edge server when edge server i and user j are successfully matched; _Vi represents the total utility of matching edge server i this time, because edge Server i exists to provide services to multiple users, so the total utility Vi of edge server _i should be accumulated by _Vi ^j ; and V _j represents the total utility of user j for this match, and V is the total utility of a single match.

Further, in the step S7, the edge server credit is updated by using the credit update formula in the credit assessment model, and the credit update formula is shown in the formula:

是为了削减上次信用度值对本次更新的影响；Num_ki代表的是i边缘服务器本次拍卖中的总交易量，即总匹配数量；Pri_ki代表的是i边缘服务器本次拍卖中的总交易价格，NumS_k代表的是本轮拍卖所有边缘服务器的总交易量，PriS_k代表的是本轮拍卖所有边缘服务器的总交易价格。Among them, Cre _ki represents the credit value of the server after the end of the kth auction, and it is necessary to update the credit value of the edge server according to the matching result; α is a variable factor, which takes the value of

It is to reduce the impact of the last credit value on this update; Num _ki represents the total transaction volume of the i-edge server in this auction, that is, the total number of matches; Pri _ki represents the total amount of the i-edge server in this auction. Transaction price, NumS _k represents the total transaction volume of all edge servers in this round of auction, and PriS _k represents the total transaction price of all edge servers in this round of auction.

The advantages and beneficial effects of the present invention are as follows:

1. The present invention introduces the demand attribute of credit degree by combining the double auction algorithm process, and considers the factors of credit degree and price, and proposes a credit degree evaluation model based on the relationship between credit degree and price, so that the credit degree can improve resource allocation by affecting resource matching. reliability. The proposed credit rating model based on the credit rating-price relationship is specifically described in claims 2-5 and claim 10. The innovative advantages are embodied in: in the existing invention, when using credit rating in auctions, all There is no definition of the relationship between creditworthiness and price. When initializing the price, if the credit and price are randomly generated, the edge server with good credit will get a lower asking price, which is obviously unreasonable, and it is impossible to clearly draw the credit value after each update. The impact of price in a match. Therefore, the present invention proposes a credit evaluation model based on the relationship between credit and price, and explains the relevant formulas (1)(2)(3)(4)(5)(6)(15) in the claims. illustrate.

2. On the basis of 1, the present invention proposes a dynamic price update mechanism, and improves the double auction algorithm by using dynamic price update. All resources are matched as much as possible to reduce the waste of edge server resources. The proposed dynamic price update mechanism is specifically described in claim 8, and the innovative advantages are embodied in: since there will be remaining resources and users who have not matched the resources after each match, in order to utilize these resources, the present invention Based on the credit-price relationship proposed in 1, the price range in the relationship is taken as the price no longer iterative boundary, and the price is updated before matching, which avoids waste of resources as much as possible. Therefore, the present invention proposes a dynamic price update mechanism, and the related formulas (8) and (9) proposed are explained in the claims.

Description of drawings

Fig. 1 is the mobile edge computing model provided by the preferred embodiment of the present invention;

Figure 2 is a flow chart of the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be described clearly and in detail below with reference to the accompanying drawings in the embodiments of the present invention. The described embodiments are only some of the embodiments of the invention.

The technical scheme that the present invention solves the above-mentioned technical problems is:

A resource allocation strategy based on a credit-price relationship provided by the present invention includes the following steps:

否则直接获取边缘服务器的要价与用户的出价。S1, obtain the information of the edge server and the user. If it is the first time to participate in matching, initialize the price between the edge server and the user according to the credit-price relationship mechanism in the credit evaluation model, and calculate the initial asking price of edge server i, Ask _i , and the initial request price of user j for edge server i.

Otherwise, directly obtain the asking price of the edge server and the user's bid.

对用户的请求进行降序排序。Rank_i，

定义如式所示。S2, rank edge servers and users according to the credit evaluation model. For edge servers, sort them in ascending order by judging the size of Rank _i . For users, according to

Sort user requests in descending order. Rank _i ,

The definition is as shown in the formula.

Rank _i =Ask _i /Cre _i

表示用户j对边缘服务器i的信用度要求。Among them, Cre _i represents the initial credit of edge server i,

Represents the credit requirement of user j to edge server i.

S3, the user's request is matched with the edge server, and if the resources of the edge server are insufficient, the matching cannot be performed. When matching, each user may bid on a different edge server. Therefore, a user may appear multiple times, and an edge server may appear multiple times.

S4, if there are still edge servers with resources after the matching is completed, the price of the edge servers with remaining resources and the users who have not completed the matching will be updated, and the matching will be performed again after the update is completed. Until the edge server has no remaining resources or the user cannot update the price with the edge server.

S5, after the matching is completed, the matching results need to be filtered. Because a user can only request one edge server at the same time, but an edge server can process requests from multiple users at the same time with sufficient resources. Because it is necessary to filter out repeated requests by users, the filtering rule is to discard user requests with lower utility.

S6, after the final matching is completed, the transaction price and utility are calculated according to the prices of both parties. Its calculation formula is shown in the formula.

V=V _j +V _i

代表的是边缘服务器i跟用户j匹配成功时所取得交易价格，Vi^j表示边缘服务器i与用户j匹配成功时边缘服务器的效用。V_i表示为本次匹配边缘服务器i的总效用，因为边缘服务器i存在提供服务给多个用户，所以边缘服务器i总效用V_i应该为V_i ^j累加。而V_j则表示用户j本次匹配的总效用，V则是单次匹配总效用。in

Represents the transaction price obtained when edge server i is successfully matched with user j, and Vi ^j represents the utility of the edge server when edge server i and user j are successfully matched. V _i represents the total utility of matching edge server i this time, because edge server i exists to provide services to multiple users, so the total utility V _i of edge server i should be accumulated for V _i ^j . And V _j represents the total utility of user j for this match, and V is the total utility of a single match.

S7, after calculating the transaction price and utility, update the credit of the edge server according to the credit update formula in the credit evaluation model, and the updated credit value will be used as the initial value of the credit for the next matching of the edge server.

S8, the matching is completed, and the entire auction process is completed.

In this embodiment, the steps of initializing the asking price of the edge server and the bidding of the user according to the credit-price relationship mechanism in the step S1 are as follows:

(1) First, initialize the asking price AskS _i of the edge server to bid with the user according to the following formula

AskS _i = Cre _i *10

AskS_i表示边缘服务器i的初始要价，Cre_i表示边缘服务器i的初始信用度。

表示用户j对边缘服务器i的初始竞价，

表示用户j对边缘服务器i的信用度要求。进行匹配时，边缘服务器i的信用度Cre_i需要大于用户j对边缘服务器i的信用度要求

这样使得用户在对边缘服务器信用度有要求的前提下，不会匹配到较低信用度的边缘服务器。where Cre _i ∈ [0,1],

AskS _i represents the initial asking price of edge server i, and Cre _i represents the initial credit of edge server i.

represents the initial bid by user j for edge server i,

Represents the credit requirement of user j to edge server i. When matching, the credibility Cre _i of the edge server i needs to be greater than the credibility requirement of the user j for the edge server i.

In this way, users will not be matched with edge servers with lower credit ratings under the premise of having requirements for edge server credibility.

为了更符合实际生活中的随机情况，则是根据如下公式进行设置。即随机生成0到1中的一个小数即为

的初始值。In the initial initialization, the credit value of each edge server is set to 0.5. For user j's credit requirements for edge server i

In order to be more in line with the random situation in real life, it is set according to the following formula. That is, a decimal number from 0 to 1 is randomly generated, which is

the initial value of .

(2) After initializing the initial price between the edge server and the user, a fixed value is generated, which does not conform to the actual situation. In order to be more in line with the competitive auction situation in real life, a price fluctuation range is set according to the price range in the experimental environment, and the final edge server asking price and user bidding are randomly generated according to the following formula.

Ask _i =AskS _i +Random(-a,a)

表示用户j对边缘服务器i的最终竞价。而Random(-a,a)表示随机在[-a,a]的区间中生成一个随机数。当Random(-a,a)取-a时，这时计算出的Ask_i表示边缘服务器能接受的最低价格，当Random(-a,a)取a时，这时计算出的Ask_i表示边缘服务器的最高价格边界。同理Random(0,b)表示在[0,b]区间生成一个随机数。Random(0,b)取b时，这时计算出的

表示用户愿意接受的最高竞价。通过在上述公式的基础上，加入随机数，使得边缘服务器在信用度的基础上可上下动态浮动价格，不固定价格。也使得用户出价随机变化，符合实际情况。Among them, Ask _i represents the initial final price of edge server i,

represents the final bid of user j for edge server i. And Random(-a,a) means to randomly generate a random number in the interval of [-a,a]. When Random(-a,a) takes -a, the calculated Ask _i represents the lowest price that the edge server can accept; when Random(-a,a) takes a, the calculated Ask _i represents the edge The maximum price boundary for the server. Similarly Random(0,b) means to generate a random number in the interval [0,b]. When Random(0,b) takes b, the calculated

Indicates the highest bid the user is willing to accept. By adding random numbers on the basis of the above formula, the edge server can dynamically float up and down the price on the basis of credit, and the price is not fixed. It also makes the user's bid to change randomly, which is in line with the actual situation.

然后根据Random(0,b)进行随机价格波动。最终得到的

表示本次匹配时用户j对边缘服务器i的最终竞价。Therefore, in each subsequent iteration, according to the credit-price relationship, for the edge server, an initial price AskS _i is first generated according to the current credit value of the edge server; then, random price fluctuations are performed according to Random(-a,a). , and the resulting Ask _i is the final asking price of the edge server during this match. For users, it is similar to the initial asking price. First, randomly generate a credit requirement value of user j for edge server i

Then make random price fluctuations according to Random(0,b). end up

Indicates the final bidding of user j for edge server i in this match.

In this embodiment, in the step S4, after each round of matching is completed, there may be remaining resources, in order to utilize these resources. On the basis of the proposed credit-price relationship, a dynamic price update mechanism is proposed. The formula of the dynamic price update mechanism is shown in the formula.

Ask' _i =Ask _i -a'

为更新后用户请求价格，b'为用户竞价更新值。对于用户，在区间

内进行价格向上迭代，迭代至

为价格不再迭代边界。对于边缘服务器，在区间[Ask'_i-a,Ask'_i+a]内进行价格向下迭代，迭代至Ask'_i-a为价格不再迭代边界。where Ask' _i is the updated asking price, and a' is the updated value of the edge server asking price.

Requested price for the updated user, b' is the updated value for the user's bid. For users, in the interval

price upward iteration within

Boundaries are no longer iterated for prices. For the edge server, iterate down the price within the interval [Ask' _i -a,Ask' _i +a], and iterate until Ask' _i -a is the price and no longer iterate the boundary.

In order to reduce the time complexity of the present invention, the value of a' is a'=a/2, and the value of b' is b'=b/2.

In this embodiment, in the step S7, the edge server credit is updated by using the credit update formula in the credit assessment model, and the credit update formula is as shown in the formula.

是为了削减上次信用度值对本次更新的影响，Num_ki代表的是i边缘服务器本次拍卖中的总交易量，即总匹配数量。Pri_ki代表的是i边缘服务器本次拍卖中的总交易价格，NumS_k代表的是本轮拍卖所有边缘服务器的总交易量，PriS_k代表的是本轮拍卖所有边缘服务器的总交易价格。Among them, Cre _ki represents the credit value of the server after the end of the k-th auction, and the credit of the edge server needs to be updated according to the matching result. α is a variable factor whose value is

In order to reduce the impact of the last credit value on this update, Num _ki represents the total transaction volume of the i-edge server in this auction, that is, the total number of matches. Pri _ki represents the total transaction price of i edge servers in this auction, NumS _k represents the total transaction volume of all edge servers in this auction, and PriS _k represents the total transaction price of all edge servers in this auction.

It should also be noted that the terms "comprising", "comprising" or any other variation thereof are intended to encompass a non-exclusive inclusion such that a process, method, article or device comprising a series of elements includes not only those elements, but also Other elements not expressly listed or inherent to such a process, method, article of manufacture or apparatus are also included. Without further limitation, an element qualified by the phrase "comprising a..." does not preclude the presence of additional identical elements in the process, method, article of manufacture or device that includes the element.

The above embodiments should be understood as only for illustrating the present invention and not for limiting the protection scope of the present invention. After reading the contents of the description of the present invention, the skilled person can make various changes or modifications to the present invention, and these equivalent changes and modifications also fall within the scope defined by the claims of the present invention.

Claims (10)

1. An edge server resource allocation method based on credit-price relationship is characterized by comprising the following steps:

s1, acquiring the resource number, asking price and credit degree of the edge server, and acquiring the bid price and the required credit degree of the user; s2, sorting the edge server and the user according to the credit rating model; s3, matching the resources owned by the user and the edge server according to the sorting result, the price constraint, the credit constraint and the like; s4, judging whether the edge server and the user can carry out price dynamic update, if so, updating the price and then continuing matching; s5, screening the matching result, and screening repeated user requests; s6, calculating a transaction price and transaction utility according to the final successful matching result; and S7, updating the credit of the edge server according to the credit evaluation model.

2. The method for allocating edge server resources based on a credit-price relationship according to claim 1, wherein the step S1 of obtaining the number of edge server resources, the asking price and the credit, and obtaining the user' S bid price and the required credit comprises:

acquiring information of the edge server and the user, if the information is matched for the first time, initializing the prices of the edge server and the user according to a credit-price relation mechanism in a credit evaluation model, and calculating an initial asking price Ask of the edge server i_iWith initial request price of user j to edge server i

Otherwise, directly acquiring the ask price of the edge server and the bid price of the user.

3. The method for edge server resource allocation based on credit-price relationship according to claim 2, wherein the step of initializing ask price of the edge server and user's bid according to the credit-price relationship mechanism is as follows:

(1) first, the ask AskS of the edge server is initialized according to the following formula_iBidding with a user

AskS_i＝Cre_i*10 (1)

Wherein Cre_i∈[0,1]，

AskS_iIndicates the initial ask, Cre, for the edge server i_iRepresenting the initial credit of the edge server i;

indicating an initial bid by user j for edge server i,

representing the credit requirement of the user j on the edge server i;

(2) randomly generating a final marginal server ask and a user bid according to the following formula;

Ask_i＝AskS_i+Random(-a,a) (3)

wherein, Ask_iRepresenting the initial final price of the edge server i,

represents the final bid of user j on edge server i, and Random (-a, a) represents the Random number of [ -a, a []Generates a Random number in the interval of (1), and similarly, Random (0, b) is represented as [0, b ]]Generating a random number in the interval;

at each iteration later, according to the credit-price relationship, for the edge server, an initial price AskS is firstly generated according to the current credit value of the edge server_i(ii) a Then, Random price fluctuation is carried out according to Random (-a, a), and finally the generated Ask_iThe asking price of the edge server at the final matching time is obtained; for the user, the price is similar to the initial asking price; firstly, randomly generating a credit degree requirement value of a user j to an edge server i

Get initial bid

Then Random price fluctuation is carried out according to Random (0, b); obtained finally

Representing the final bid of user j on edge server i at this match.

4. The method of claim 3, wherein the method comprises allocating resources to the edge server based on a credit-price relationship

Also indicates the credit Cre of the edge server i when matching is performed_iNeeds to be greater than the credit requirement of user j for edge server i

At initial initialization, a credit value of 0.5 is set for each edge server, and a credit requirement is set for user j for edge server i

In order to better meet the random situation in real life, the setting is carried out according to the following formula. That is, randomly generating a decimal number from 0 to 1

Is started.

5. The method for allocating resource of an edge server based on a credit-price relationship according to claim 4, wherein the step S2 is to rank the edge server and the user according to a credit evaluation model, and specifically comprises: and carrying out priority ordering on the edge server and the user according to the credit rating model. Aiming at the edge server, by judging Rank_iSorting the data in ascending order according to size, and aiming at the user, sorting the data according to size

Sorting the requests of the users in descending order, Rank_i，

The definition is shown in the formula.

Rank_i＝Ask_i/Cre_i (6)

Wherein Cre_iIndicating the initial credit of the edge server i,

representing the credit requirement of user j for edge server i.

6. The method for allocating resource of edge server based on credit-price relationship as claimed in claim 5, wherein said step S3 matches the credit required by the user and the resource owned by the edge server according to the sorting result, specifically comprising: and matching the user with the edge server according to price constraint, credit constraint and the like, wherein the matching cannot be performed if the resources of the edge server are insufficient during matching. Since each user may bid on a different edge server, one user may appear multiple times and one edge server may appear multiple times.

7. The method for allocating resource to an edge server based on a credit-price relationship according to claim 6, wherein the step S4 is performed to determine whether the edge server and the user can perform a price dynamic update, and if so, the method continues to match after updating the price, and specifically includes: on the basis of the proposed credit-price relationship, a dynamic price updating mechanism is proposed; the dynamic price updating mechanism formula is shown as the following formula:

Ask'_i＝Ask_i-a' (8)

wherein

Updating the value of the asking price of the edge server for the updated asking price of the edge server;

to request a price for the updated user, b' bids the updated value for the user. For the user, in the section

Iterating the price upwards

No more iteration boundaries for price; for edge Server, in the interval [ Ask'_i-a,Ask'_i+a]Inner progress price iteration down to Ask'_iA is the price no-more-iteration boundary.

8. The method for allocating resource to an edge server based on a credit-price relationship according to claim 7, wherein the step S5 is performed to filter matching results and filter repeated user requests, and specifically includes: after matching is completed, the matching result needs to be screened; because only one edge server can be requested by one user at the same time, but one edge server can process the requests of a plurality of users under the condition of sufficient resources at the same time; because the request repeatedly sent by the user needs to be screened out, the screening rule is to discard the user request with lower effectiveness.

9. The method for allocating resource to an edge server based on credit-price relationship of claim 8, wherein in step S6, after the final matching is completed, the transaction price and the utility are calculated according to the prices of both parties, and the calculation formula is shown as the formula.

V＝V_j+V_i (14)

Wherein

Representing the transaction price, Vi, achieved when the edge server i successfully matches the user j^jRepresenting the utility of the edge server when the edge server i is successfully matched with the user j; v_iExpressed as the total utility of the matching edge server i at this time, the edge server i exists to provide service to a plurality of users, so the total utility V of the edge server i_iShould be V_i ^jAccumulating; and V_jThe total utility of the current matching of the user j is shown, and V is the total utility of the single matching.

10. The method for allocating edge server resources based on credit-price relationship of claim 9, wherein in step S7, the edge server credit is updated by using a credit update formula in the credit evaluation model, wherein the credit update formula is as follows:

wherein Cre_kiRepresenting the credit value of the server after the k-th auction is finished, and updating the credit of the edge server according to the matching result; alpha is a variable factor having a value of

The influence of the credit value on the updating at this time is reduced; num_kiThe representative is the total transaction amount, namely the total matching amount, in the auction of the edge server; pri_kiRepresentative is the total transaction price, NumS, in this auction for the i-edge server_kRepresentative is the total transaction volume, PrIS, of all edge servers in the auction round_kThe representative is the total transaction price of all edge servers in the current round of auction.

Images