A priority-based multi-attribute resource auction method in space-ground integrated network

Abstract

The present invention discloses a priority-based multi-attribute resource auction method in a space-ground integrated network. The present invention designs four models for the auction architecture, including a seller framework, an auctioneer framework, a buyer framework, and a drone framework. Based on auction theory, the present invention proposes a distributed solution. In order to achieve real-time updating of bidding information, a drone model is added, which is more suitable for such a network with frequent changes in time and space. The auction is conducted according to the multiple attributes of the bidding users, and the definition of user priority is added, further ensuring the efficiency and overall effectiveness of the entire resource allocation method. On the one hand, the present invention meets the complex environment of heterogeneous networks, and on the other hand, takes into account real-time communication between services. The present invention more comprehensively considers the priority of users in resource allocation, thereby improving the efficiency of resource allocation.

Description

Technical Field

[0001] The present invention belongs to the field of space-ground networks, and in particular relates to a priority-based multi-attribute resource auction method in a space-ground integrated network. Background Art

[0002] With the growing diversification of user demands, my country is vigorously promoting the construction of a space-ground integrated network. This network system, based on space platforms, acquires, transmits, and processes space information in real time. Satellite networks and drones are key components of this network, offering wide coverage, direct global communication, and rapid information transmission. However, the frequent changes in satellite network topology, the diverse resource types, and the time-varying nature of satellite networks lead to inflexible resource allocation strategies, poor algorithm stability, and problems such as high cost, high resource consumption, and low resource utilization. Therefore, breaking through the current bottlenecks in network resource management and comprehensively scheduling network resources to meet growing business demands is of great research significance.

[0003] While existing work has made some progress in resource allocation for integrated space-ground networks, it still has shortcomings. Given that integrated space-ground networks are a more holistic scenario, the question of how to rationally allocate multi-dimensional resources within the network remains an ongoing issue. Traditional resource allocation methods struggle to meet today's multi-service demands with limited network resources, but existing multi-dimensional resource joint management methods are complex and fail to fully consider the quality of service requirements of different real-time services. Therefore, breaking through the current bottleneck of rational network resource allocation and comprehensively scheduling and allocating network resources to meet growing service demands is key to improving resource utilization.

[0004] Most existing resource allocation methods focus on single resources. Link-oriented resource allocation methods include topology snapshots and dynamic routing. Transmission resource allocation primarily focuses on bandwidth and spectrum resources. These methods don't achieve high resource utilization in space-ground networks. Furthermore, the frequent spatiotemporal changes in integrated space-ground networks also impact resource allocation effectiveness.

[0005] The above solutions do not address the design of mechanisms that can allocate resources from a higher perspective. Auction theory is a classic model for resource allocation mechanism design. It has been extensively studied in economics and widely applied in other fields, particularly in resource competition. Mechanism design considers four key properties: budget balance, computational efficiency, individual rationality, and incentive compatibility. However, resource auction mechanisms in integrated ground-to-space networks are still limited by the centralized control of traditional terrestrial networks, and user utility cannot be effectively met. Summary of the Invention

[0006] The present invention aims to overcome these shortcomings by providing a priority-based, multi-attribute resource auction method for a space-ground integrated network. Three different priority calculation methods are proposed within the core algorithm module. Resource auctions are then conducted based on multiple attributes, including bid price, completion time, and coordination. This method also optimizes the resource auction model and improves resource utilization.

[0007] In order to achieve the above object, the present invention comprises the following steps:

[0008] S1, extracting services and resources from the integrated space-ground network;

[0009] S2, building a resource auction framework;

[0010] S3, analyze and model the resource auction framework to obtain the seller framework, auctioneer framework, buyer framework and drone framework;

[0011] S4, based on auction theory, proposed a distributed solution;

[0012] S5, combining different attributes of auction users, defines a multi-attribute auction model;

[0013] S6, formulates a variety of different methods based on the user priority during the auction process.

[0014] In S2, the resource auction framework is as follows:

[0015] Assume that the buyer set N = {1,2,…,n}, and the drone represents the role of the auctioneer. As a trusted third party, it determines the allocation rules, payment rules, and priority rules. The resource capacity of each virtual resource pool has an upper limit, and the set of resource capacities is represented by R = {R1,R2,…,R m};

[0016] Assume that the size of resources occupied by the buyer is B = {b1, b2, ..., b u}, bidding matrix P = {p nm ,1≤n≤N,1≤m≤M}, where p nm represents the bid price of user service n for virtual resource pool m. Each virtual resource pool announces its fixed price f m , if the buying price p nm Below the selling price f m , then the auction between user n and virtual resource pool m will fail. Otherwise, using u nm =1 to record this pair of transactions.

[0017] After the buyer and seller have determined their bids, the cost for user n is determined to be The payment for virtual resource pool m is The auctioneer's income is When a user successfully bids, determine the user's satisfaction sat nm , which means the satisfaction value when user n purchases resources from virtual resource pool m. The virtual resource pool provides various resources to user services and needs to bear maintenance costs and other costs. Then MT={mt1,mt2,…,mt m} represents the cost of the virtual resource pool;

[0018] The calculation formula of the user's utility function is as follows:

[0019]

[0020] The calculation formula of the utility function of the virtual resource pool is as follows:

[0021]

[0022] The auctioneer’s utility function is calculated as follows:

[0023]

[0024] In S3, the resource auction framework includes the seller framework, core framework, buyer framework, and drone framework;

[0025] The core framework includes the auction information transmission layer, the core algorithm layer and the decision collection layer. The auction information transmission layer and the drone framework allow the virtual resource pool to know in real time whether the buyer's needs are valid, and transmit the information to the bid processing module for processing. The drone framework includes the receiving and sending module and the bid information library, which are used to maintain real-time communication between the core frameworks.

[0026] In S4, the distributed solution is to form an auction game between the business flow and the virtual resource pool, and the user business bids for each resource required in the virtual resource pool.

[0027] In S5, the multi-attribute auction model includes a bidding price model, the completion time of user bidding, and whether there is collaboration between users.

[0028] In S6, priority is used as a multi-attribute of user bidding. Auction failure is proportional to the priority of the user's bid and the possibility of exiting the auction. Priority is proportional to the user's chance of winning the auction.

[0029] Statistical Bid Attribute Set Based on the attributes of the bids received and refined according to priority, The winning function formula is defined as follows:

[0030]

[0031] According to the results of successful bidding, three methods of adjusting priorities are formulated, and the number of resource auctions tpi (u i ) and the number of auctions won by the bidding user vic(u i ), the priority calculation formula is as follows:

[0032]

[0033] Priority is defined based on the consecutive failures of the bidder. The longer the user's bid fails, the higher the priority. For this purpose, a failure threshold fts is defined. The priority calculation formula is as follows:

[0034]

[0035] Define winning / losing streak fi i , which is determined by the ratio of the winning function and is calculated as follows:

[0036]

[0037] The priority is calculated in a similar way to the above two methods, as shown in the formula:

[0038]

[0039] in, For users to quote their bid for each resource, ft i The completion time of the user's bid, Co i is the user's collaborative situation, cfc(u i ) is the continuous failure coefficient of the bidding user, F(P′ i ) is the winning function, is the sum of all suitability values ​​for the in-person auction.

[0040] Compared with the existing technology, the present invention designs four models for the auction architecture, including a seller framework, an auctioneer framework, a buyer framework, and a drone framework. Based on auction theory, the present invention proposes a distributed solution. In order to achieve real-time updates of bidding information, a drone model is added, which is more suitable for such networks with frequent changes in time and space. The auction is conducted according to the multiple attributes of the bidding users, and the definition of user priority is added, further ensuring the efficiency and overall effectiveness of the entire resource allocation method. On the one hand, the present invention meets the complex environment of heterogeneous networks, and on the other hand, it takes into account real-time communication between services. The present invention more comprehensively considers the priority of users in resource allocation, thereby improving the efficiency of resource allocation.

[0041] Furthermore, the present invention can characterize the problem of resource auctions, and the UAV framework maintains real-time communication between core frameworks, including the receiving and sending modules and the bidding information library, which improves user satisfaction and system response time. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a system architecture diagram of the present invention;

[0043] Figure 2 This is a resource auction framework diagram of the present invention;

[0044] Figure 3 It is the multi-attribute graph of the present invention. DETAILED DESCRIPTION

[0045] The present invention will be further described below with reference to the accompanying drawings.

[0046] The present invention discloses a priority-based multi-attribute resource auction method in a space-ground integrated network. Figure 1 On this basis, a resource auction framework diagram was constructed. Figure 2 , Figure 2 The core algorithm module in the core framework determines the allocation rules, payment rules and priority rules. This auction method is based on the user's multi-attribute method. Figure 3 , specifically including the following steps:

[0047] Step 1: There are many types of services and resources in the integrated space-ground network. In order to optimize resource allocation and user satisfaction, the services and resources in the integrated space-ground network are extracted and the services that need to participate in resource auctions and the required resources in the network environment are classified.

[0048] Step 2: Build a resource auction framework, such as Figure 1 As shown in Figure 1, the auction system architecture in the integrated space-ground network has three layers: the buyer layer, the seller layer, and the auctioneer layer. The basic functions are as follows:

[0049] The buyer layer consists of service modules, including various voice and video services, observation services for emergency scenarios, and satellite-to-ground communications. Its function is to send and receive messages between the buyer and the auctioneer. This module includes a buyer bid collection module and a buyer notification module: one for sending bid requests to the auctioneer and the other for receiving the auctioneer's decision information. Finally, the service module determines the bid price and executes the bid results.

[0050] Seller layer: It is composed of air-based, space-based and ground-based virtual resource pools. As the owner of the resources, it provides the auctioneer with an inquiry price.

[0051] Auctioneer layer: The auctioneer framework is the core part of the entire network model, consisting of the core framework and the drone framework.

[0052] Assume that the set n = {1,2,…,n} represents the buyer set. The auctioneer, as a trusted third party, determines the allocation rules, payment rules, and priority rules. The resource capacity of each virtual resource pool has an upper limit. The set of resource capacities is represented by R = {R1,R2,…,R m ,}. Let B={b1,b2,…,b u} represents the size of resources occupied by the buyer, P = {p nm ,1≤n≤N,1≤m≤M} represents the bidding matrix, where p nm represents the bid price of business n for virtual resource pool m. Each virtual resource pool announces its fixed price f m , if the buying price p nm Below the selling price f m , then the auction between user n and virtual resource pool m will fail. Otherwise, using u nm =1 to record this pair of transactions; after the buyer and seller determine the bid, the cost of user in is determined to be The payment for virtual resource pool m is The auctioneer's income is When a user successfully bids, the user's satisfaction is expressed as sat nm It means the satisfaction value of user n when purchasing resources from virtual resource pool m. The virtual resource pool provides various resources to user services and needs to bear the maintenance costs and other costs. Then MT = {mt1,mt2,…,mt m} represents the cost of the virtual resource pool, and the resource demand vector of each user is defined as Resource usage Represents the amount of resources required by the multi-dimensional resources in the virtual resource pool to complete a user's bid per unit time, the business processing volume V k , which represents the number of user services that can be executed in a unit time for each virtual resource pool k.

[0053] The calculation formula of the user's utility function is as follows:

[0054]

[0055] The calculation formula of the utility function of the virtual resource pool is as follows:

[0056]

[0057] The auctioneer’s utility function is calculated as follows:

[0058]

[0059] Step 3: Analyze and model the seller framework, auctioneer framework, buyer framework, and drone framework. The core framework is the core part of the auctioneer framework and consists of three parts: the auction information transmission layer, the core algorithm layer, and the decision collection layer. The auction information transmission layer and the drone framework allow the virtual resource pool to know in real time whether the buyer's demand is valid and transmit the information to the bid processing module for processing. The main function of the drone framework is to maintain real-time communication between the core frameworks, including the sending and receiving modules and the bid information library.

[0060] Step 4: Due to the problems of traditional centralized network methods, a distributed solution for resource allocation is proposed. This distributed solution is based on the application of auction theory. By forming an auction game between service flows and virtual resource pools, user services bid for each resource they need in the virtual resource pool. Based on the bid information submitted by the user services, the price of each resource in each virtual resource pool is determined and announced to the users along with the calculated service processing volume. Each user maximizes its profit based on the price received from the virtual resource pool and updates its bid for the next round.

[0061] Step 5: Define a multi-attribute auction model by combining various attributes of auction users, such as Figure 3 As shown, the attribute set includes bid price, bid completion time, user collaboration status and priority.

[0062] The bidding price is defined as follows: each bidding user provides a bidding price for resource l in virtual resource pool k: Each virtual resource pool calculates the price of each resource: The resource price matrix is ​​P = (p l,k ), business processing volume per unit time: The resource price and business processing volume per unit time of each virtual resource pool will be announced to users, and their final business processing volume will be calculated: Each virtual resource pool further uses resource prices to update its business processing capacity by maximizing its overall utility:

[0063] Priorities are defined to maintain relative fairness throughout the auction process. Since users bid repeatedly to compete for resources, the system can track users' past auction records, identify users who are likely to withdraw from the auction, and make fair decisions to maintain user trust in the auction. To this end, users are assigned priorities based on their historical records during resource allocation.

[0064] In addition to priority and price, we should also consider the completion time of user bids and whether there is collaboration among users.

[0065] In step 6, the user priority in the auction process is formulated in three ways: obtaining a bidding coefficient, a continuous failure coefficient, and a winning coefficient based on multiple attributes.

[0066] According to the results of successful bidding, three methods of adjusting priority can be formulated. Obtaining the bidding coefficient is the most intuitive way to calculate the priority. The number of resource auctions tpi (u i ) and the number of auctions won by the bidding user vic(u i ), the priority calculation formula is as follows:

[0067]

[0068] This coefficient provides the auctioneer with the percentage of auctions each bidder has won since the auctioneer entered the auction market. Using this priority, a bidder who has won all auctions will have priority while a bidder who has never won an auction will have a priority close to 1.

[0069] Priority is defined based on the number of consecutive bidder failures. The longer a user's bids fail, the higher their priority. To this end, a failure threshold fts is defined. After losing a certain number of consecutive auctions, as long as fts is reached, the bidder is just as likely to leave the auction as the losing bidder. The priority calculation formula is as follows:

[0070]

[0071] cfc(u i ) is the continuous failure coefficient of the bidding user. The bidder who just succeeded in bidding has priority s i = 0, and the bidding users who participated in the FTS auction but did not win have priority s i = 1. Compared with the OBF method, we can say that the CFC method has no long-term memory because a bidding user will see his priority reduced to 0 after winning the auction, regardless of his past result records.

[0072] The multi-attribute winning coefficient method treats bidders who almost win the auction and bidders who offer the worst bids in the same way. On the one hand, this situation may encourage some bidders to offer dummy bids that have no chance of winning in order to increase their priority and get a higher chance of winning the auction. On the other hand, some bidders who almost win the auction may become frustrated when they realize that the attribute values ​​of the bids have no effect on the priority assessment. To avoid this, the attributes of the bids offered are considered in the priority definition. Define the winning / losing streak f i , which is determined by the ratio of the winning function and is calculated as follows:

[0073]

[0074] The priority is calculated in a similar way to the above two methods, as shown in the formula:

[0075]

[0076] in, For users to quote their bid for each resource, ft i The completion time of the user's bid, Co i is the user's collaborative situation, cfc(u i ) is the continuous failure coefficient of the bidding user, F(P′ i ) is the winning function, is the sum of all suitability values ​​for the in-person auction.

[0077] If two bidding users win the auction for the same amount, the bidding user that provides higher quality attributes will receive the highest priority.

Claims

1. A priority-based multi-attribute resource auction method in a space-ground integrated network, characterized in that: The following steps are involved: S1, extracting services and resources from the integrated space-ground network; S2, build a resource auction framework as follows: Assume that the buyer set N = {1,2,…,n}, the auctioneer acts as a trusted third party to determine the allocation rules, payment rules and priority rules. The resource capacity of each virtual resource pool has an upper limit, and the set of resource capacities is represented by R = {R1,R2,…,R m }; Assume that the size of resources occupied by the buyer is B = {b1, b2, ..., b u }, bidding matrix P = {p nm ,1≤n≤N,1≤m≤M}, where p nm represents the bid price of user service n for virtual resource pool m. Each virtual resource pool announces its fixed price f m , if the buying price p nm Below the selling price f m , then the auction between user n and virtual resource pool m will fail. Otherwise, using u nm =1 to record this pair of transactions; S3, the resource auction framework, consists of a core framework and a drone framework. The core framework includes an auction information transmission layer, a core algorithm layer, and a decision collection layer. The auction information transmission layer and the drone framework allow the virtual resource pool to know in real time whether the buyer's demand is valid and transmit this information to the bid processing module. The drone framework includes a receiving and sending module and a bid information library to maintain real-time communication between the core frameworks. S4, based on auction theory, proposes a distributed solution. This solution forms an auction game between service flows and virtual resource pools, where user services bid for each resource they need in the virtual resource pool. S5, combining different attributes of auction users, defines a multi-attribute auction model; the multi-attribute auction model includes the bid price model, the completion time of user bids, and whether there is coordination between users; Bid prices are defined. Each bidding user provides a bid price for resources in the virtual resource pool. Each virtual resource pool calculates the price of each resource and combines all resource prices into a resource price matrix. The business processing volume per unit time is calculated based on the bid price, resource price, and resource demand. The resource price and business processing volume per unit time of each virtual resource pool are announced to users to calculate their final business processing volume. Each virtual resource pool further uses the resource price to update its business processing volume by maximizing its overall utility. S6, according to the user priority in the auction process, develop a variety of different methods; priority is a multi-attribute of the user's bid, the auction failure is proportional to the user's bid priority and the possibility of exiting the auction, and the priority is proportional to the user's chance of winning the auction; a variety of different methods are used to obtain the bid coefficient, the continuous failure coefficient and the winning coefficient based on multiple attributes; statistical bidding attribute set Based on the attributes of the bids received and refined according to priority, The winning function formula is defined as follows: According to the results of successful bidding, three methods of adjusting priorities are formulated, and the number of resource auctions tpi (u i ) and the number of auctions won by the bidding user vic(u i ), the priority calculation formula is as follows: Priority is defined based on the consecutive failures of the bidder. The longer the user's bid fails, the higher the priority. For this purpose, a failure threshold fts is defined. The priority calculation formula is as follows: Define the winning / losing streak f i , which is determined by the ratio of the winning function and is calculated as follows: The priority is calculated in a similar way to the above two methods, as shown in the formula: in, For users to quote their bid for each resource, ft i The completion time of the user's bid, Co i is the user's collaborative situation, cfc(u i ) is the continuous failure coefficient of the bidding user, F(P′ i ) is the winning function, is the sum of all suitability values ​​for the in-person auction.

2. The priority-based multi-attribute resource auction method in a space-ground integrated network according to claim 1, characterized in that: S2 also includes the cost of user n after the buyer and seller have determined the bid. The payment for virtual resource pool m is The auctioneer's income is When a user successfully bids, determine the user's satisfaction sat nm , which means the satisfaction value when user n purchases resources from virtual resource pool m. The virtual resource pool provides various resources to user services and needs to bear maintenance costs and other costs. Then MT={mt1,mt2,…,mt m } represents the cost of the virtual resource pool; The calculation formula of the user's utility function is as follows: The calculation formula of the utility function of the virtual resource pool is as follows: The auctioneer’s utility function is calculated as follows:

Images