An algorithmic network resource allocation method and system, electronic device and storage medium

Abstract

The application discloses a kind of algorithm network resource allocation method, system, electronic equipment and storage medium, method includes: according to the overall coverage of user side algorithm network resource and effective utilization rate determines user side initial candidate list;According to the overall coverage of resource provider algorithm network resource and effective utilization rate determines resource provider initial candidate list;According to user side initial candidate list, determine user side satisfaction score, generate user side final candidate list based on user side satisfaction score and corresponding resource provider number identification;According to resource provider candidate list, determine resource provider satisfaction score, generate resource provider final candidate list based on resource provider satisfaction score and corresponding user side number identification;Based on user side and resource provider final candidate list, using two sides satisfaction driven stable matching mechanism completes algorithm network resource allocation.The application can simultaneously consider accuracy, fairness and satisfaction in multidimensional, uncertain resource allocation.

Description

Technical Field

[0001] This invention belongs to the field of computing network resource allocation, and specifically relates to a computing network resource allocation method, system, electronic device and storage medium. Background Technology

[0002] With the rapid development of computing networks, the complexity of resource allocation has increased significantly. Traditional resource allocation methods often ignore the dynamic fluctuations in user demand and resource supply, typically employing evaluation methods based on single numerical values, resulting in insufficient accuracy and adaptability in the allocation results. Meanwhile, although stable matching methods can guarantee the stability of the matching, their one-sided optimality may lead to unfair allocation results, making it difficult to meet diverse user needs and dynamically changing resource environments. Against this backdrop, how to balance accuracy, fairness, and user satisfaction in allocation has become a critical issue that urgently needs to be addressed in the field of computing network resource allocation.

[0003] Existing computing network resource allocation technologies have the following main shortcomings:

[0004] 1. The methods for describing resource capabilities are insufficient to reflect uncertainty. Existing methods for allocating computing network resources generally use single, fixed values ​​to describe the capabilities of computing power, network, storage, and other resources. This fails to reflect the fluctuation characteristics of resource performance under different load states, operating environments, and time dimensions, leading to discrepancies between resource assessment results and actual available capabilities, thus affecting the accuracy and rationality of resource allocation results.

[0005] 2. The resource matching process lacks feasible domain constraint mechanisms and multi-dimensional comprehensive evaluation mechanisms. Existing resource allocation methods typically verify the allocation results after matching is completed, but fail to introduce explicit feasibility constraints during the matching phase. This results in a large number of resource combinations that cannot be satisfied at the physical capability or performance index level participating in the matching process, increasing system computational complexity and reducing overall scheduling efficiency. Furthermore, the resource matching process often focuses on a single-dimensional matching index, making it difficult to comprehensively evaluate the matching quality from both the degree of resource demand satisfaction and the degree of resource utilization. This can easily lead to low resource utilization or the inability to effectively meet the needs of some users.

[0006] 3. Stable matching methods suffer from unilateral optima and insufficient overall efficiency. Existing network resource allocation methods based on stable matching theory often use fixed candidate lists. The matching results are usually more favorable to the matching initiator, making it difficult to dynamically reflect changes in resource utilization efficiency and the degree of user demand satisfaction. This can easily lead to an imbalance of interests between the resource side and the user side, affecting the overall system operating efficiency and user satisfaction. Summary of the Invention

[0007] Purpose of the invention: The purpose of this invention is to provide a computing network resource allocation method to solve the problem that existing computing network resource allocation methods are unable to simultaneously take into account accuracy, fairness and satisfaction in multi-dimensional and uncertain resource allocation.

[0008] Technical solution: The present invention provides a method for allocating computing network resources, comprising:

[0009] Based on the user's performance indicators, determine the overall coverage and effective utilization of the user's computing network resources, and form an initial candidate list of users with the corresponding user ID identifier.

[0010] Based on the performance indicators of the resource providers, the overall coverage rate and effective utilization rate of the computing network resources of the resource providers are determined, and an initial candidate list of resource providers is formed by combining the corresponding resource provider IDs.

[0011] The user satisfaction score is determined based on the initial candidate list of users, and the final candidate list of users is generated based on the user satisfaction score and the corresponding resource provider ID.

[0012] Based on the initial candidate list of resource providers, a satisfaction score for the resource provider is determined, and a final candidate list of resource providers is generated based on the satisfaction score and the corresponding user ID.

[0013] Based on the final candidate list of users and the final candidate list of resource providers, a stable matching mechanism driven by two-sided satisfaction is adopted to complete the allocation of computing network resources. The stable matching mechanism driven by two-sided satisfaction means that the resource provider proposes a match, the user accepts or rejects the match, and the rejected resource provider continues to propose a match in the next round until one of the two parties has completed the match, and the resource allocation ends.

[0014] Furthermore, based on the user's performance indicators, the overall coverage and effective utilization rate of the user's computing network resources are determined, including:

[0015] If the product of the resource provider's network resource indicator and the performance fluctuation rate of the resource provider's network resource indicator is greater than or equal to the upper limit of the user's network resource indicator expectation, or if the resource provider's network resource indicator is less than or equal to the lower limit of the user's network resource indicator expectation, then the overall coverage rate of the user's network resources is 0; otherwise, the overall coverage rate of the user's network resources is determined based on the upper limit of the user's network resource indicator expectation, the lower limit of the user's network resource indicator expectation, the resource provider's network resource indicator, and the performance fluctuation rate of the resource provider's network resource indicator.

[0016] If the product of the resource provider's network resource index and the performance fluctuation rate of the resource provider's network resource index is greater than or equal to the upper limit of the user's network resource index expectation, or if the resource provider's network resource index is less than or equal to the lower limit of the user's network resource index expectation, then the effective utilization rate of the user's network resources is 0; otherwise, the effective utilization rate of the user's network resources is determined based on the upper limit of the user's network resource index expectation, the lower limit of the user's network resource index expectation, the resource provider's network resource index, and the performance fluctuation rate of the resource provider's network resource index.

[0017] Furthermore, based on the resource provider's performance indicators, the overall coverage and effective utilization rate of the computing network resources of the resource provider are determined, including:

[0018] If the product of the resource provider's network resource index and the performance fluctuation rate of the resource provider's network resource index is greater than or equal to the upper limit of the user's network resource index expectation, or if the resource provider's network resource index is less than or equal to the lower limit of the user's network resource index expectation, then the overall coverage rate of the resource provider's network resources is 0; otherwise, the overall coverage rate of the resource provider's network resources is determined based on the upper limit of the user's network resource index expectation, the lower limit of the user's network resource index expectation, the resource provider's network resource index, and the performance fluctuation rate of the resource provider's network resource index.

[0019] If the product of the resource provider's network resource index and the performance fluctuation rate of the resource provider's network resource index is greater than or equal to the upper limit of the user's network resource index expectation, or if the resource provider's network resource index is less than or equal to the lower limit of the user's network resource index expectation, then the effective utilization rate of the resource provider's network resources is 0; otherwise, the effective utilization rate of the resource provider's network resources is determined based on the upper limit of the user's network resource index expectation, the lower limit of the user's network resource index expectation, the resource provider's network resource index, and the performance fluctuation rate of the resource provider's network resource index.

[0020] Furthermore, the overall coverage rate and effective utilization rate of computing network resources of the users in the initial candidate list of users are both non-zero values; the overall coverage rate and effective utilization rate of computing network resources of the resource providers in the initial candidate list of resource providers are both non-zero values.

[0021] Furthermore, based on the initial candidate list of users, a user satisfaction score is determined, including:

[0022] The user satisfaction score is calculated by weighting and summing the products of the overall coverage and effective utilization of network resources for each indicator, with the weights representing the preferences for each indicator.

[0023] Furthermore, a resource provider satisfaction score is determined based on the initial candidate list of resource providers, including:

[0024] The resource provider's effective utilization rate of computing network resources for each indicator is weighted and summed to calculate the resource provider's satisfaction score, with the weights representing the preference for each indicator.

[0025] Furthermore, based on the final candidate list of users and the final candidate list of resource providers, a stable matching mechanism driven by two-sided satisfaction is adopted to complete the allocation of computing network resources, specifically including:

[0026] After the final candidate lists for users and resource providers are completed, each unmatched resource provider initiates a matching request to the user with the highest current satisfaction rating in its own final candidate list.

[0027] After each user receives all matching requests, it selects the resource provider with the highest satisfaction rating from its final candidate list and temporarily accepts the match; the remaining unselected resource providers will be rejected, and the unselected resource providers will remove the rejected users from their final candidate list and obtain an updated final candidate list of resource providers.

[0028] The user and the resource provider enter a temporary matching state. The rejected resource provider continues to send matching requests to the user with the highest current satisfaction rating in the updated final candidate list of resource providers in the next round.

[0029] Once one of the participating parties has completed the matching process, the resource allocation ends, and the final matching status between the user and the resource provider is obtained.

[0030] Based on the same inventive concept, another aspect of the present invention provides a computer network resource allocation system, comprising:

[0031] The user-side initial candidate list construction unit is used to determine the overall coverage and effective utilization rate of the user-side computing network resources based on the user-side performance indicators, and to form the user-side initial candidate list with the corresponding user-side ID identifier.

[0032] The initial candidate list construction unit for resource providers is used to determine the overall coverage and effective utilization rate of computing network resources of resource providers based on the performance indicators of resource providers, and to form the initial candidate list of resource providers with the corresponding resource provider ID identifier;

[0033] The user-side final candidate list construction unit is used to determine the user-side satisfaction score based on the user-side initial candidate list, and generate the user-side final candidate list based on the user-side satisfaction score and the corresponding resource provider ID identifier.

[0034] The resource provider final candidate list construction unit is used to determine the resource provider satisfaction score based on the initial resource provider candidate list, and generate the resource provider final candidate list based on the resource provider satisfaction score and the corresponding user ID identifier;

[0035] The resource allocation unit is used to complete the allocation of computing network resources based on the final candidate list of users and the final candidate list of resource providers, using a stable matching mechanism driven by two-sided satisfaction. The stable matching mechanism driven by two-sided satisfaction means that the resource provider proposes a match, the user accepts or rejects the match, and the rejected resource provider continues to propose a match in the next round until one of the two parties has completed the match, at which point the resource allocation ends.

[0036] Based on the same inventive concept, in another aspect of the present invention, an electronic device includes:

[0037] Memory, used to store computer programs;

[0038] A processor is used to execute the computer program to implement the network resource allocation method.

[0039] Based on the same inventive concept, another aspect of the present invention provides a non-volatile storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the aforementioned computer network resource allocation method.

[0040] Beneficial effects: Compared with the prior art, the significant technical effects of this invention are as follows: An initial candidate list for users is determined by the overall coverage and effective utilization rate of computing network resources on the user side, and an initial candidate list for resource providers is determined by the overall coverage and effective utilization rate of computing network resources on the resource provider side; initial screening of users and resource providers is performed before the matching stage to avoid infeasible resource combinations from participating in the matching; and a two-sided satisfaction index is constructed from two dimensions: user-side demand satisfaction and resource-side utilization efficiency, namely, user-side satisfaction score and resource provider satisfaction score. A matching candidate list is dynamically generated based on the two-sided satisfaction index, and finally, computing network resource allocation is completed through a stable matching mechanism driven by two-sided satisfaction. Attached Figure Description

[0041] Figure 1 This is a flowchart of the method of the present invention;

[0042] Figure 2 This is a flowchart of the process for determining network resource indicators and interval values;

[0043] Figure 3 This is a flowchart of the user's initial candidate list construction method;

[0044] Figure 4 This is a flowchart illustrating the method for constructing the initial candidate list of resource providers;

[0045] Figure 5 This is a flowchart of the user's final candidate list construction process;

[0046] Figure 6 This is a flowchart illustrating the process of constructing the final candidate list of resource providers;

[0047] Figure 7 This is a flowchart of the network resource allocation process. Detailed Implementation

[0048] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0049] like Figure 1 As shown, the network resource allocation method includes:

[0050] Based on the user's performance indicators, determine the overall coverage and effective utilization of the user's computing network resources, and form an initial candidate list of users with the corresponding user ID identifier.

[0051] Based on the performance indicators of the resource providers, the overall coverage rate and effective utilization rate of the computing network resources of the resource providers are determined, and an initial candidate list of resource providers is formed by combining the corresponding resource provider IDs.

[0052] The user satisfaction score is determined based on the initial candidate list of users, and the final candidate list of users is generated based on the user satisfaction score and the corresponding resource provider ID.

[0053] Based on the initial candidate list of resource providers, a satisfaction score for the resource provider is determined, and a final candidate list of resource providers is generated based on the satisfaction score and the corresponding user ID.

[0054] Based on the final candidate list of users and the final candidate list of resource providers, a stable matching mechanism driven by two-sided satisfaction is adopted to complete the allocation of computing network resources. The stable matching mechanism driven by two-sided satisfaction means that the resource provider proposes a match, the user accepts or rejects the match, and the rejected resource provider continues to propose a match in the next round until one of the two parties has completed the match, and the resource allocation ends.

[0055] Specifically, the following steps are included:

[0056] First, before determining the overall coverage and effective utilization rate of computing network resources for both users and resource providers, we determine the computing network resource indicators for both resource providers and users, and define the range values ​​for each type of computing network resource indicator. By evaluating the range values ​​of computing network resource indicators, we can comprehensively capture the fluctuation range and uncertainty of resource attributes, thereby improving the accuracy and reliability of resource allocation.

[0057] Step 1: Determine the computing network resource indicators for both the resource provider and the user, and determine the range values ​​for each type of computing network resource indicator;

[0058] Specifically, the network resource indicators for both resource providers and users are categorized, and range values ​​for each category are determined. These network resource indicators include computing power performance indicators, network performance indicators, and storage performance indicators. The range values ​​for both resource provider and user performance indicators are further divided. By defining these range values, the fluctuation range of network resource indicators in actual operation is characterized, forming comprehensive resource indicator range values, providing fundamental data support for subsequent resource matching. Figure 2 As shown, the process includes determining network resource indicators, determining the performance indicator ranges for resource providers and users, and determining a comprehensive resource indicator range assessment for both resource providers and users. Specifically, it includes:

[0059] It is known that the size of the resource provider is The user base is , For the collection of resource providers, For the number of resource providers, For user-side collection, The number of users and the types of network resource indicators are: , and They represent the first The value of the m-th network resource indicator for the resource provider, and the value of the u-th user's... The values ​​of the computing network resource indicators are determined. In this embodiment, there are three types of computing network resource indicators, including computing power performance indicators, network performance indicators, and storage performance indicators.

[0060] Step 11: Determine computing network resource indicators. Based on the user's computing network resource requirements and the characteristics of the resource provider's computing network resources, computing network resource indicators can be categorized as follows, but this application is not limited to these three:

[0061] (1) Computing performance indicators

[0062] For resource providers, computing performance metrics are used to measure the computing performance of equipment. These metrics include: CPU floating-point arithmetic capability, CPU clock speed, number of CPU cores, number of CPUs, GPU double-precision floating-point arithmetic capability, GPU clock speed, number of GPU cores, and number of GPUs. For users, computing performance metrics are used to measure their computing performance requirements. To simplify the user experience, users can directly submit their expected ranges for CPU and GPU computing performance.

[0063] (2) Network performance indicators

[0064] For resource providers, network performance metrics are used to measure the network performance of devices. These metrics include network bandwidth, jitter, latency, and packet loss rate. For users, network performance metrics are used to measure the user's expected performance of the current network. These metrics also include network bandwidth, jitter, latency, and packet loss rate.

[0065] (3) Storage performance indicators

[0066] For resource providers, storage performance metrics are used to measure the storage performance of devices. These metrics include total disk storage, disk utilization, disk read speed, and disk write speed. For users, storage performance metrics are used to measure the user's expected performance for the current storage. These metrics also include total disk storage, disk utilization, disk read speed, and disk write speed.

[0067] Step 12: Determine the range values ​​for various computing network resource indicators. The range values ​​for computing network resource indicators include the performance indicator range values ​​for resource providers and users. The resource provider performance indicator range values ​​include: computing power performance range, network performance range, and storage performance range. The user performance indicator range values ​​include: computing power performance range, network performance range, and storage performance range. First, in constructing the computing power performance range values, the computing power performance of the resource provider is calculated, and computing power performance range values ​​are constructed to clarify the range of computing power performance expected by the user's equipment. Second, in constructing the network performance and storage performance range values, based on equipment characteristics and user needs, corresponding network performance and storage performance range values ​​are constructed for both the resource provider's equipment and the user's equipment. Finally, the comprehensive resource indicator range values ​​are constructed. The specific method is as follows:

[0068] The formula for calculating the CPU computing power performance index for resource providers is as follows:

[0069] (1)

[0070] in, Indicates the resource provider CPU computing power performance Indicates the CPU's floating-point arithmetic capability. Indicates CPU clock speed. Indicates the number of cores in a single CPU. Indicates the number of CPUs.

[0071] The formula for calculating the GPU computing power performance metrics for resource providers is as follows:

[0072] (2)

[0073] in, This indicates the GPU computing power performance of resource provider p. This indicates the GPU's double-precision floating-point arithmetic capability. Indicates the GPU clock speed. Indicates the number of cores in a single GPU. This indicates the number of GPUs.

[0074] The lower limit of the computing power performance range of the resource provider is the product of the computing power performance of the resource provider and the rate of change of the computing power performance index, and the upper limit is the computing power performance of the resource provider.

[0075] Specifically: Considering that CPU and GPU computing performance are affected by environmental factors, load, and other indicators, the performance ranges for CPU and GPU computing power of the resource provider are as follows: , ,in and These represent the fluctuation rates of CPU computing power performance metrics and GPU computing power performance metrics, respectively. .

[0076] The lower limit of the user's computing power performance range is the lower limit of the user's expected computing power performance, and the upper limit is the upper limit of the user's expected computing power performance.

[0077] Meanwhile, the constructed user-side CPU computing power performance and GPU computing power performance index ranges are as follows: , .in, This represents the lower limit of the user's expected CPU computing power performance. This indicates the upper limit of the user's expected CPU computing power performance. This represents the lower limit of the user's expected GPU computing power performance. This indicates the upper limit of the user's expected GPU computing power performance.

[0078] The lower limit of the resource provider's network performance range is the product of the resource provider's network performance and the rate of change of the network performance index, and the upper limit is the resource provider's network performance.

[0079] The lower limit of the storage performance range of the resource provider is the product of the storage performance of the resource provider and the rate of change of the storage performance index, and the upper limit is the storage performance of the resource provider.

[0080] Specifically, the range values ​​for the network performance and storage performance of the resource provider are constructed as shown in Table 1 below.

[0081] Table 1. Construction of range values ​​for network and storage performance of resource providers

[0082]

[0083] In Table 1, Indicates the resource provider network bandwidth, This represents the rate of change in network bandwidth performance, and ; Indicates the resource provider The shaking, This represents the rate of change in jitter performance, and ; Indicates the resource provider The time delay, This represents the rate of change of latency performance, and ; Indicates the resource provider packet loss rate This represents the rate of change in packet loss performance, and ; Indicates the resource provider Total disk storage, This represents the rate of change in the performance of the total disk storage capacity, and ; Indicates the resource provider Disk usage This indicates the rate of change in disk usage performance, and ; Indicates the resource provider Disk read speed, This indicates the rate of change in disk read speed performance, and ; Indicates the resource provider Disk write speed, This represents the rate of change in disk write speed performance, and .

[0084] The lower limit of the user's network performance range is the lower limit of the user's expected network performance, and the upper limit is the upper limit of the user's expected network performance.

[0085] The lower limit of the user's storage performance range is the lower limit of the user's expected storage performance, and the upper limit is the upper limit of the user's expected storage performance.

[0086] Specifically, the range values ​​for network performance and storage performance on the user side are shown in Table 2 below.

[0087] Table 2. Constructing the range values ​​for network performance and storage performance on the user side.

[0088]

[0089] In Table 2, This represents the lower limit of the user's expected network bandwidth. This indicates the user's expected upper limit of network bandwidth. This represents the lower bound of the user's expected jitter. This represents the upper limit of the user's expected jitter. This represents the lower bound of the user's expected latency. This represents the upper limit of the user's expected latency. This represents the lower bound of the user's expected packet loss rate. This represents the upper limit of the user's expected packet loss rate. This represents the lower limit of the user's expected total disk storage capacity. This indicates the user's expected upper limit on the total disk storage capacity. This represents the lower limit of the user's expected disk utilization. This indicates the user's expected upper limit for disk usage. This indicates the lower limit of the user's expected disk read speed. This indicates the user's expected upper limit for disk read speed. This indicates the lower limit of the user's expected disk write speed. This indicates the user's expected upper limit for disk write speed.

[0090] Finally, the comprehensive resource indicator range values ​​obtained from the resource providers are constructed as shown in Table 3 below.

[0091] Table 3. Construction of Comprehensive Resource Indicator Range Values ​​for Resource Providers

[0092]

[0093] The range of comprehensive resource indicators for users is shown in Table 4 below.

[0094] Table 4. Construction of the comprehensive resource index range values ​​for the user side.

[0095]

[0096] The proposed method for classifying and modeling network resource indicators, combining the characteristics of user network resource requirements with the characteristics of resource provider network resource capabilities, classifies network resource indicators. These indicators include at least computing power performance indicators, network performance indicators, and storage performance indicators. Corresponding interval value models are then constructed for each type of resource indicator, resulting in interval values ​​for computing power performance indicators, network performance indicators, and storage performance indicators. This constructs a comprehensive range of resource indicator intervals to characterize the available range of resource capabilities, thereby improving the reliability and stability of resource assessment.

[0097] After determining the computing network resource indicators and range values, the similarity of computing network resource matching and the effectiveness of its use are further determined. By introducing two metrics, "overall coverage of computing network resources" and "effective utilization rate of computing network resources", the similarity of resource matching and the effectiveness of resource use for resource providers and users are determined respectively.

[0098] Step 2: Based on the user's performance indicators, determine the overall coverage and effective utilization rate of the user's computing network resources, and form an initial candidate list of users with the corresponding user ID identifiers;

[0099] Specifically, based on the user's computing power performance indicators, network performance indicators, and storage performance indicators, the similarity and effectiveness of the user's computing network resource matching are determined. If the input resource is from the user, the similarity of the user's computing network resource matching is defined and calculated through the overall coverage rate of the user's computing network resources; whereby the overall coverage rate of computing network resources is used to measure the attribute coverage ratio of the demand being met, aiming to reflect the accuracy of resource matching, and thus characterize the similarity of computing network resource matching.

[0100] Based on the calculation of the overall coverage of computing network resources on the user side, the effectiveness of the use of computing network resources on the user side is defined and calculated through the effective utilization rate of computing network resources on the user side. The effective utilization rate of computing network resources is used to measure the actual utilization efficiency of resources, aiming to reflect the sufficiency of resource utilization and characterize the effectiveness of computing network resource use.

[0101] Given that there are m types of network resource indicators, including CPU computing power performance indicators ( GPU computing power performance indicators ( ), network bandwidth ( ), shaking ( ), delay ( Packet loss rate Total disk storage ( ), disk usage ( ), disk read speed ( Disk write speed )wait.

[0102] like Figure 3 As shown, the specific steps include:

[0103] Step 21: Determine the overall coverage of the user's computing network resources:

[0104] If the product of the resource provider's network resource metric and the performance fluctuation rate of the resource provider's network resource metric is greater than or equal to the upper limit of the user's network resource metric expectation, or if the resource provider's network resource metric is less than or equal to the lower limit of the user's network resource metric expectation, then the overall coverage rate of the user's network resources is 0. Otherwise, the overall coverage rate of the user's network resources is calculated based on the upper limit of the user's network resource metric expectation, the lower limit of the user's network resource metric expectation, the resource provider's network resource metric, and the performance fluctuation rate of the resource provider's network resource metric. The formula for calculating the overall coverage rate of the user's network resources is as follows:

[0105] (3)

[0106] in, This represents the upper limit of the expected network resource indicators for the m-th user. This represents the lower bound of the expected network resource indicators for the m-th user. This represents the network resource index of the m-th resource provider. This represents the rate of change in the performance of the network resource indicators of the m-th resource provider. These indicators have been defined previously.

[0107] Taking CPU computing power performance metrics as an example, the known range of CPU computing power performance metrics from resource providers is... and The user's CPU computing power performance index range is [value missing]. The calculation of the overall coverage of network resources on the user side includes the following six scenarios, as shown below.

[0108] Scenario 1: The user's resource demand is far less than the resource provider's resource supply, i.e. The overall coverage rate of network resources on the user side is 0.

[0109] Scenario 2: The user's resource demand far exceeds the resource provider's resource supply, i.e. The overall coverage rate of network resources on the user side is 0.

[0110] Scenario 3: The user's maximum resource requirement is met, but the minimum resource requirement is not met. and The overall coverage rate of user-side computing network resources is ;

[0111] For example: If the user needs [10,19], the resource provider can provide [12,23]. The user's coverage rate is: [(19-12)+(19-10)] / [(23-10)+(19-10)]=16 / 22.

[0112] Scenario 4: The user's resource requirements are fully met, that is... and The overall coverage rate of user-side computing network resources is ;

[0113] Scenario 5: The user's minimum resource requirement is met, but the maximum resource requirement is not met. and The overall coverage rate of user-side computing network resources is ;

[0114] Scenario 6: The user's partial intermediate resource requirements are met, i.e. and The overall coverage rate of user-side computing network resources is .

[0115] Taking GPU computing power performance metrics as an example, the known range of GPU computing power performance metrics from resource providers is... and The user's GPU computing power performance index range is [value missing]. The calculation of the overall coverage of network resources on the user side includes the following six scenarios, as shown below.

[0116] Scenario 1: The user's resource demand is far less than the resource provider's resource supply, i.e. The overall coverage rate of network resources on the user side is 0.

[0117] Scenario 2: The user's resource demand far exceeds the resource provider's resource supply, i.e. The overall coverage rate of network resources on the user side is 0.

[0118] Scenario 3: The user's maximum resource requirement is met, but the minimum resource requirement is not met. and The overall coverage rate of user-side computing network resources is ;

[0119] Scenario 4: The user's resource requirements are fully met, that is... and The overall coverage rate of user-side computing network resources is ;

[0120] Scenario 5: The user's minimum resource requirement is met, but the maximum resource requirement is not met. and The overall coverage rate of user-side computing network resources is ;

[0121] Scenario 6: The user's partial intermediate resource requirements are met, i.e. and The overall coverage rate of user-side computing network resources is .

[0122] The remaining network and storage performance metrics coverage for users are calculated in the same way (the same way as the CPU and GPU computing power performance metrics mentioned above), and will not be elaborated further.

[0123] Step 22: Determine the effective utilization rate of user-side computing network resources:

[0124] If the product of the resource provider's network resource index and the performance fluctuation rate of the resource provider's network resource index is greater than or equal to the upper limit of the user's network resource index, or if the resource provider's network resource index is less than or equal to the lower limit of the user's network resource index, then the effective utilization rate of the user's network resources is 0; otherwise, the effective utilization rate of the user's network resources is calculated based on the upper limit of the user's network resource index, the lower limit of the user's network resource index, the resource provider's network resource index, and the performance fluctuation rate of the resource provider's network resource index.

[0125] The actual utilization efficiency of computing network resources by the user is calculated. The formula for calculating the effective utilization rate of computing network resources by the user is as follows:

[0126] (4)

[0127] in, This represents the upper limit of the expected network resource indicators for the m-th user. This represents the lower bound of the expected network resource indicators for the m-th user. This represents the network resource index of the m-th resource provider. This represents the rate of change in the performance of the network resource indicators of the m-th resource provider. These indicators have been defined previously.

[0128] Taking CPU computing power performance metrics as an example, the known range of CPU computing power performance metrics from resource providers is... and The user's CPU computing power performance index range is [value missing]. The calculation of the effective utilization rate of network resources by the user includes the following six scenarios, as shown below.

[0129] Scenario 1: The user's resource demand is far less than the resource provider's resource supply, i.e. The effective utilization rate of network resources on the user side is 0.

[0130] Scenario 2: The user's resource demand far exceeds the resource provider's resource supply, i.e. The effective utilization rate of network resources on the user side is 0.

[0131] Scenario 3: The user's maximum resource requirement is met, but the minimum resource requirement is not met. and The effective utilization rate of network resources on the user side is ;

[0132] Scenario 4: The user's resource requirements are fully met, that is... and The effective utilization rate of network resources on the user side is ;

[0133] Scenario 5: The user's minimum resource requirement is met, but the maximum resource requirement is not met. and The effective utilization rate of network resources on the user side is ;

[0134] Scenario 6: The user's partial intermediate resource requirements are met, i.e. and The effective utilization rate of network resources on the user side is .

[0135] Taking GPU computing power performance metrics as an example, the known range of GPU computing power performance metrics from resource providers is... and The user's GPU computing power performance index range is [value missing]. The calculation of the effective utilization rate of network resources by the user includes the following six scenarios, as shown below.

[0136] Scenario 1: The user's resource demand is far less than the resource provider's resource supply, i.e. The effective utilization rate of network resources on the user side is 0.

[0137] Scenario 2: The user's resource demand far exceeds the resource provider's resource supply, i.e. The effective utilization rate of network resources on the user side is 0.

[0138] Scenario 3: The user's maximum resource requirement is met, but the minimum resource requirement is not met. and The effective utilization rate of network resources on the user side is ;

[0139] Scenario 4: The user's resource requirements are fully met, that is... and The effective utilization rate of network resources on the user side is ;

[0140] Scenario 5: The user's minimum resource requirement is met, but the maximum resource requirement is not met. and The effective utilization rate of network resources on the user side is ;

[0141] Scenario 6: The user's partial intermediate resource requirements are met, i.e. and The effective utilization rate of network resources on the user side is .

[0142] The remaining network and storage performance metrics for the user side are calculated in the same way and will not be elaborated further.

[0143] Step 23: Constructing the Initial Candidate List for Users. Following steps S21 and S22, the overall coverage and effective utilization rate of user network resources are calculated. The overall coverage and effective utilization rate of user network resources, along with the corresponding user identification numbers, form the initial candidate list for users, completing the initial screening and preventing resource waste by large-scale resource providers and scenarios where user needs cannot be met.

[0144] Furthermore, the overall coverage rate and effective utilization rate of the computing network resources of the user in the initial candidate list of the user are both non-zero values, so as to avoid infeasible resource combinations from participating in the subsequent matching process.

[0145] Step 3: Based on the performance indicators of the resource providers, determine the overall coverage rate and effective utilization rate of the computing network resources of the resource providers, and form an initial candidate list of resource providers with the corresponding resource provider IDs;

[0146] Specifically, based on the computing power performance indicators, network performance indicators, and storage performance indicators of the resource provider, the overall coverage rate and effective utilization rate of the computing network resources of the resource provider are determined. If the input resource is the resource provider, the computing network resource matching similarity of the resource provider is calculated by defining the overall resource coverage rate of the resource provider. Among them, the overall computing network resource coverage rate is used to measure the attribute coverage ratio of the demand being met, which aims to reflect the accuracy of resource matching and thus characterize the computing network resource matching similarity.

[0147] Based on the calculation of the overall coverage rate of the computing network resources provided by the resource provider, the effectiveness of the use of computing network resources by the resource provider is defined and calculated through the effective utilization rate of computing network resources by the resource provider. Among them, the effective utilization rate of computing network resources is used to measure the true efficiency of resource use, aiming to reflect the sufficiency of resource utilization and characterize the effectiveness of computing network resource use.

[0148] Given that there are m types of network resource indicators, including CPU computing power performance indicators ( GPU computing power performance indicators ( ), network bandwidth ( ), shaking ( ), delay ( Packet loss rate Total disk storage ( ), disk usage ( ), disk read speed ( Disk write speed )wait.

[0149] like Figure 4 As shown, the specific steps include:

[0150] Step 31: Determine the overall coverage of the network resources provided by the resource provider:

[0151] If the product of the resource provider's network resource metric and the performance fluctuation rate of the resource provider's network resource metric is greater than or equal to the upper limit of the user's network resource metric expectation, or if the resource provider's network resource metric is less than or equal to the lower limit of the user's network resource metric expectation, then the overall coverage rate of the resource provider's network resources is 0. Otherwise, the overall coverage rate of the resource provider's network resources is calculated based on the upper limit of the user's network resource metric expectation, the lower limit of the user's network resource metric expectation, the resource provider's network resource metric, and the performance fluctuation rate of the resource provider's network resource metric. The formula for calculating the overall coverage rate of the resource provider's network resources is as follows:

[0152] (5)

[0153] in, This represents the upper limit of the expected network resource indicators for the m-th user. This represents the lower bound of the expected network resource indicators for the m-th user. This represents the network resource index of the m-th resource provider. This represents the rate of change in the performance of the network resource indicators of the m-th resource provider. These indicators have been defined previously.

[0154] Taking CPU computing power performance metrics as an example, the known range of CPU computing power performance metrics from resource providers is... and The user's CPU computing power performance index range is [value missing]. The calculation of the overall coverage of network resources by resource providers includes the following six scenarios, as shown below.

[0155] Scenario 1: The resource provider offers far more resources than the user requires, i.e. The overall coverage rate of the network resources provided by the resource provider is 0.

[0156] Scenario 2: The resource provider offers far fewer resources than the user requires, i.e. The overall coverage rate of the network resources provided by the resource provider is 0.

[0157] Scenario 3: The resource provider provides resources to meet the user's minimum resource requirements, but not the maximum resource requirements, i.e. and The overall coverage rate of the resource provider's computing network resources is ;

[0158] Scenario 4: The resources provided by the resource provider are fully utilized by the user, i.e. and The overall coverage rate of the resource provider's computing network resources is ;

[0159] Scenario 5: The resource provider offers resources that meet the user's maximum resource requirement, but fail to meet the minimum resource requirement, i.e. and The overall coverage rate of the resource provider's computing network resources is ;

[0160] For example: the resource provider needs [10,19], and the user can provide [12,23]. The resource provider's coverage is: [(19-12)+(19-10)] / [(23-10)+(19-10)]=16 / 22.

[0161] Scenario 6: The resource provider only satisfies the resource needs of a portion of the intermediary users, i.e. and The overall coverage rate of the resource provider's computing network resources is .

[0162] Taking GPU computing power performance metrics as an example, the known range of GPU computing power performance metrics from resource providers is... and The user's GPU computing power performance index range is [value missing]. The calculation of the overall coverage of network resources by resource providers includes the following six scenarios, as shown below.

[0163] Scenario 1: The resource provider offers far more resources than the user requires, i.e. The overall coverage rate of the network resources provided by the resource provider is 0.

[0164] Scenario 2: The resource provider offers far fewer resources than the user requires, i.e. The overall coverage rate of the network resources provided by the resource provider is 0.

[0165] Scenario 3: The resource provider provides resources to meet the user's minimum resource requirements, but not the maximum resource requirements, i.e. and The overall coverage rate of the resource provider's computing network resources is ;

[0166] Scenario 4: The resources provided by the resource provider are fully utilized by the user, i.e. and The overall coverage rate of the resource provider's computing network resources is ;

[0167] Scenario 5: The resource provider offers resources that meet the user's maximum resource requirement, but fail to meet the minimum resource requirement, i.e. and The overall coverage rate of the resource provider's computing network resources is ;

[0168] Scenario 6: The resource provider only satisfies the resource needs of a portion of the intermediary users, i.e. and The overall coverage rate of the resource provider's computing network resources is .

[0169] The overall coverage of network and storage performance metrics for the remaining resource providers is calculated in the same way, and will not be elaborated further.

[0170] Step 32: Determine the effective utilization rate of network resources provided by the resource provider.

[0171] If the product of the resource provider's network resource index and the performance fluctuation rate of the resource provider's network resource index is greater than or equal to the upper limit of the user's network resource index expectation, or if the resource provider's network resource index is less than or equal to the lower limit of the user's network resource index expectation, then the effective utilization rate of the resource provider's network resources is 0; otherwise, the effective utilization rate of the resource provider's network resources is calculated based on the upper limit of the user's network resource index expectation, the lower limit of the user's network resource index expectation, the resource provider's network resource index, and the performance fluctuation rate of the resource provider's network resource index.

[0172] The actual utilization efficiency of the resource provider in the matching of computing network resources is calculated. The formula for calculating the effective utilization rate of computing network resources for the resource provider is as follows:

[0173] (6)

[0174] in, This represents the upper limit of the expected network resource indicators for the m-th user. This represents the lower bound of the expected network resource indicators for the m-th user. This represents the network resource index of the m-th resource provider. This represents the rate of change in the performance of the network resource indicators of the m-th resource provider. These indicators have been defined previously.

[0175] Taking CPU computing power performance metrics as an example, the known range of CPU computing power performance metrics from resource providers is... and The user's CPU computing power performance index range is [value missing]. The calculation of the effective utilization rate of network resources by resource providers includes the following six scenarios, as shown below.

[0176] Scenario 1: The resource provider offers far more resources than the user requires, i.e. The effective utilization rate of network resources provided by the resource provider is 0.

[0177] Scenario 2: The resource provider offers far fewer resources than the user requires, i.e. The effective utilization rate of network resources provided by the resource provider is 0.

[0178] Scenario 3: The resource provider provides resources to meet the user's minimum resource requirements, but not the maximum resource requirements, i.e. and The effective utilization rate of network resources provided by the resource provider is ;

[0179] Scenario 4: The resources provided by the resource provider are fully utilized by the user, i.e. and The effective utilization rate of network resources provided by the resource provider is ;

[0180] Scenario 5: The resource provider offers resources that meet the user's maximum resource requirement, but fail to meet the minimum resource requirement, i.e. and The effective utilization rate of network resources provided by the resource provider is ;

[0181] Scenario 6: The resource provider only satisfies the resource needs of a portion of the intermediary users, i.e. and The effective utilization rate of network resources provided by the resource provider is .

[0182] Taking GPU computing power performance metrics as an example, the known range of GPU computing power performance metrics from resource providers is... and The user's GPU computing power performance index range is [value missing]. The calculation of the effective utilization rate of network resources by resource providers includes the following six scenarios, as shown below.

[0183] Scenario 1: The resource provider offers far more resources than the user requires, i.e. The effective utilization rate of network resources provided by the resource provider is 0.

[0184] Scenario 2: The resource provider offers far fewer resources than the user requires, i.e. The effective utilization rate of network resources provided by the resource provider is 0.

[0185] Scenario 3: The resource provider provides resources to meet the user's minimum resource requirements, but not the maximum resource requirements, i.e. and The effective utilization rate of network resources provided by the resource provider is ;

[0186] Scenario 4: The resources provided by the resource provider are fully utilized by the user, i.e. and The effective utilization rate of network resources provided by the resource provider is ;

[0187] Scenario 5: The resource provider offers resources that meet the user's maximum resource requirement, but fail to meet the minimum resource requirement, i.e. and The effective utilization rate of network resources provided by the resource provider is ;

[0188] Scenario 6: The resource provider only satisfies the resource needs of a portion of the intermediary users, i.e. and The effective utilization rate of network resources provided by the resource provider is .

[0189] The remaining performance metrics for resource providers, such as computing power, network, and storage, are calculated in the same way and will not be elaborated upon.

[0190] Step 33: Constructing the Initial Candidate List of Resource Providers. Through steps S31 and S32, the overall coverage and effective utilization rate of the computing network resources of each resource provider are calculated. The overall coverage and effective utilization rates of the computing network resources of each resource provider, along with their corresponding resource provider IDs, constitute the initial candidate list of resource providers; this completes the initial screening, preventing large-scale resource waste by resource providers and scenarios where user needs cannot be met.

[0191] Furthermore, the overall coverage rate and effective utilization rate of computing network resources of the resource providers in the initial candidate list of resource providers are both non-zero values, in order to avoid infeasible resource combinations from participating in the subsequent matching process.

[0192] The proposed method for calculating the similarity and effectiveness of computing network resource matching establishes a two-way calculation mechanism from the perspectives of resource matching accuracy and resource utilization sufficiency. On the user side, it calculates the overall coverage rate and effective utilization rate of the user's computing network resources; on the resource side, it calculates the overall coverage rate and effective utilization rate of the resource provider's computing network resources. Through the joint calculation and verification of the above two-way indicators, a comprehensive evaluation of the quality of resource matching is achieved. Values ​​with zero overall coverage rate or effective utilization rate of computing network resources are deleted to avoid infeasible resource combinations from participating in the matching, thereby reducing the risk of resource waste while ensuring matching accuracy.

[0193] After the initial screening of users and resource providers, based on the similarity of computing network resource matching and the effectiveness of computing network resource usage, user satisfaction scores and resource provider satisfaction scores were calculated respectively. The final candidate lists for users and resource providers were then constructed by combining the resource provider ID and user ID corresponding to the satisfaction scores, with both parties marked as "not matched". Details are as follows:

[0194] Step 4: Determine the user satisfaction score based on the initial candidate list of users, and generate the final candidate list of users based on the user satisfaction score and the corresponding resource provider ID.

[0195] Specifically, based on the overall coverage and effective utilization rate of computing network resources in the initial candidate list of users, a user satisfaction score is determined, and a final candidate list of users is generated based on the user satisfaction score and the corresponding resource provider ID. The user satisfaction score represents the user's satisfaction with the resource provider, that is, it characterizes the degree to which the user's needs are met.

[0196] The method for constructing the candidate list of users is as follows: the product of the overall coverage and effective utilization rate of the user's computing network resources for each indicator is weighted and summed to calculate the user satisfaction score, with the weights being the preferences for each indicator; the candidate list of users is obtained based on the user satisfaction score and the corresponding resource provider ID.

[0197] like Figure 5 As shown, the specific steps include:

[0198] Step 41: Determine user satisfaction rating;

[0199] In calculating user satisfaction scores, the traditional stable matching algorithm (Gale-Shapely, GS) only benefits the initiator, harming the interests of the other party. In this invention, the resource provider is the initiator, and the user is the receiver. The overall coverage rate of the user's computing network resources is combined with the effective utilization rate of those resources to ensure user satisfaction. The user satisfaction score is calculated by weighted summation of the products of the overall coverage rate and effective utilization rate of each indicator, with the weights representing the preference for each indicator. The formula for calculating the user satisfaction score is as follows.

[0200] (7)

[0201] in, This indicates user satisfaction rating; This represents the overall coverage rate of user-side computing network resources for the l-th indicator; This represents the effective utilization rate of user-side computing network resources for the first type of indicator; This represents the user's preference for the l-th indicator, and its value is determined by the user. .

[0202] Step 42: Based on the user satisfaction score calculation results and the corresponding resource provider ID, construct the final candidate list of users. In this embodiment, the user satisfaction scores are arranged in descending or ascending order.

[0203] Step 5: Determine the resource provider satisfaction score based on the initial candidate list of resource providers, and generate the final candidate list of resource providers based on the resource provider satisfaction score and the corresponding user ID identifier;

[0204] Specifically, based on the overall network resource coverage and effective utilization rate of the network resources in the initial candidate list of resource providers, a resource provider satisfaction score is determined. Then, a final candidate list of resource providers is generated based on the resource provider satisfaction score and the corresponding user identification number. The resource provider satisfaction score represents the resource provider's satisfaction with the user, i.e., it characterizes resource utilization efficiency.

[0205] like Figure 6 As shown, the specific steps include:

[0206] Step 51: Determine the resource provider satisfaction score;

[0207] As the initiator, the resource provider considers that if the provided resources are not fully utilized by users, a large amount of idle computing power will be generated. Therefore, the calculation method of the resource provider satisfaction score is different from that of the user satisfaction score. The effective utilization rate of the resource provider's computing network resources for each indicator is weighted and summed to calculate the resource provider satisfaction score. The weights are the preferences of each indicator. The specific calculation formula is as follows.

[0208] (8)

[0209] in, This indicates the satisfaction rating of the resource provider; This represents the effective utilization rate of network resources provided by the resource provider for the l-th indicator; This represents the resource provider's preference for the l-th indicator, the value of which is determined by the resource provider. .

[0210] Step 52: Based on the resource provider satisfaction score calculation results and the corresponding user ID identifier, construct the final candidate list of resource providers. In this embodiment, the resource provider satisfaction scores are arranged in descending or ascending order.

[0211] Step 6: Based on the final candidate list of users and the final candidate list of resource providers, a stable matching mechanism driven by two-sided satisfaction is used to complete the allocation of computing network resources. This stable matching mechanism means that the resource provider proposes a match, the user accepts or rejects the match, and the rejected resource provider continues to propose a match in the next round until both parties have completed their matching, at which point the resource allocation ends. Figure 7 As shown, it specifically includes:

[0212] Step 61: Resource provider proposes matching. After the final candidate lists for users and resource providers are constructed, each unmatched resource provider initiates a matching request to the user with the highest current satisfaction rating in its own final candidate list.

[0213] Step 62, User Response. After receiving all matching requests, each user selects the resource provider with the highest satisfaction rating from their final candidate list and temporarily accepts the match. The remaining unselected resource providers are rejected, and these unselected resource providers are removed from their final candidate lists, resulting in an updated final resource provider candidate list.

[0214] Step 63: Update Status. The accepting user and resource provider enter a temporary matching state. The rejected resource provider continues to send matching requests to the user with the highest current satisfaction rating in the updated final resource provider candidate list in the next round.

[0215] Step 64: Allocation terminates. Once one of the participating parties has completed matching, the resource allocation is considered complete, and the final matching status of the user and resource provider is obtained, along with a list of matching candidates.

[0216] By introducing a stable matching mechanism driven by two-sided satisfaction, the unilateral optimality defect of traditional stable matching algorithms is effectively overcome while ensuring matching stability.

[0217] The dual-satisfaction-driven stable matching mechanism calculates user satisfaction scores and resource provider satisfaction scores based on the overall coverage and effective utilization of computing network resources during the stable matching process, and constructs a final candidate list for users and a final candidate list for resource providers accordingly. During the matching process, the final candidate lists for users and resource providers are allowed to be dynamically updated as resource status changes. Through this dual-satisfaction-driven stable matching iterative process, the interests of resource providers and users are balanced while ensuring matching stability, achieving precise allocation of computing network resources.

[0218] By modeling computing power, network, and storage resources using interval values, the uncertainty of resource capabilities is explicitly introduced into the resource allocation process. Initial candidate lists for users and resource providers are determined before the matching phase using overall network resource coverage and effective utilization rates, avoiding infeasible resource combinations from participating in the matching process. Furthermore, a two-sided satisfaction index is constructed based on user-side demand satisfaction and resource-side utilization efficiency. A matching candidate list is dynamically generated based on these two-sided satisfaction indices, and finally, a stable matching mechanism driven by two-sided satisfaction completes the network resource allocation. Each step is executed sequentially according to a predetermined order, collaboratively completing the network resource allocation process. The aim is to achieve accurate evaluation of network resources through interval values ​​and ensure the fairness and accuracy of the allocation results through an improved stable matching method.

[0219] In summary, the research on a stable matching mechanism based on interval constraint-driven and two-sided satisfaction-driven methods for computing network resource allocation is of great significance for improving user satisfaction, enhancing allocation fairness, and expanding the application of stable matching theory in computing network resource allocation. This strategy improves the accuracy and reliability of resource allocation by introducing an interval value evaluation mechanism. Simultaneously, by improving traditional stable matching methods, it comprehensively considers the overall coverage and effective utilization rate of computing network resources, further optimizing the allocation results while ensuring matching fairness. This research provides a novel approach to solving multi-dimensional and uncertain resource allocation problems and opens up new directions for the application of stable matching theory in computing network resource allocation scenarios.

[0220] The above optional implementation methods can achieve at least the following beneficial effects:

[0221] 1. This invention aims to address the problem that existing computing network resource allocation methods struggle to simultaneously ensure accuracy, fairness, and satisfaction in multi-dimensional and uncertain resource allocation. It designs a computing network resource allocation method based on interval constraint-driven and a two-sided stable matching mechanism. First, by modeling the capabilities of computing power, network, and storage resources in the form of interval values, the uncertainty of resource capabilities is explicitly introduced into the resource allocation process. Based on the overall coverage and effective utilization rate of computing network resources, an initial candidate list for users and an initial candidate list for resource providers are determined before the matching stage to avoid infeasible resource combinations from participating in the matching. Furthermore, a two-sided satisfaction index is constructed from two dimensions: user-side demand satisfaction and resource-side utilization efficiency. A matching candidate list is dynamically generated based on this two-sided satisfaction index, and finally, the computing network resource allocation is completed through a two-sided stable matching mechanism.

[0222] 2. This invention designs a method for classifying and modeling network resource indicators based on interval values. Combining the characteristics of user network resource requirements and the network resource capabilities of resource providers, network resource indicators are classified. These indicators include at least computing power performance indicators, network performance indicators, and storage performance indicators. Corresponding interval value models are then constructed for each type of resource indicator, resulting in interval values ​​for computing power performance indicators, network performance indicators, storage performance indicators, and a comprehensive resource indicator. These interval values ​​characterize the available range of resource capabilities, thereby improving the reliability and stability of resource assessment.

[0223] 3. This invention designs a method for calculating the similarity and effectiveness of network resource matching. A two-way calculation mechanism is established from the perspectives of resource matching accuracy and resource utilization sufficiency. On the user side, the overall resource coverage rate and effective utilization rate of the user's resources are calculated; on the resource side, the overall resource coverage rate and effective utilization rate of the resource provider's resources are calculated. Through joint calculation and cross-validation of the above two-way indicators, a comprehensive evaluation of resource matching quality is achieved. Furthermore, by combining the initial candidate lists of the user and the initial candidate lists of resource providers, infeasible resource combinations are avoided from participating in the matching process, thereby reducing the risk of resource waste while ensuring matching accuracy.

[0224] 4. This invention designs a stable matching method driven by two-sided network resource satisfaction. During the stable matching process, user satisfaction scores and resource satisfaction scores are calculated based on overall resource coverage and effective resource utilization, respectively, and preference lists for both parties are constructed accordingly. During the matching process, these preference lists are allowed to be dynamically updated as resource status changes. Through this two-sided satisfaction-driven stable matching iterative process, the interests of resource providers and users are balanced while ensuring matching stability, achieving precise allocation of network resources.

[0225] Example 2:

[0226] Embodiment 2 of the present invention also provides a computing network resource allocation system, comprising:

[0227] The user-side initial candidate list construction unit is used to determine the overall coverage and effective utilization rate of the user-side computing network resources based on the user-side performance indicators, and to form the user-side initial candidate list with the corresponding user-side ID identifier.

[0228] The initial candidate list construction unit for resource providers is used to determine the overall coverage and effective utilization rate of computing network resources of resource providers based on the performance indicators of resource providers, and to form the initial candidate list of resource providers with the corresponding resource provider ID identifier;

[0229] The user-side final candidate list construction unit is used to determine the user-side satisfaction score based on the user-side initial candidate list, and generate the user-side final candidate list based on the user-side satisfaction score and the corresponding resource provider ID identifier.

[0230] The resource provider final candidate list construction unit is used to determine the resource provider satisfaction score based on the initial resource provider candidate list, and generate the resource provider final candidate list based on the resource provider satisfaction score and the corresponding user ID identifier;

[0231] The resource allocation unit is used to complete the allocation of computing network resources based on the final candidate list of users and the final candidate list of resource providers, using a stable matching mechanism driven by two-sided satisfaction. The stable matching mechanism driven by two-sided satisfaction means that the resource provider proposes a match, the user accepts or rejects the match, and the rejected resource provider continues to propose a match in the next round until one of the two parties has completed the match, at which point the resource allocation ends.

[0232] Example 3:

[0233] Embodiment 3 of the present invention also provides an electronic device, including: a memory and a processor, wherein the memory is used to store a computer program; the processor is used to execute the computer program to implement the above-described computer network resource allocation method.

[0234] Example 4:

[0235] Embodiment 4 of the present invention also provides a non-volatile storage medium for storing a computer program, wherein the computer program implements the aforementioned computer network resource allocation method when executed by a processor.

[0236] Example 5:

[0237] Embodiment 5 of the present invention also provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the aforementioned computer network resource allocation method.

[0238] The aforementioned network resource allocation system, electronic equipment, non-volatile storage medium, and computer program products have the same beneficial effects as the network resource allocation method described above, and will not be repeated here.

Claims

1. A method for allocating computing network resources, characterized in that, include: Based on the user's performance indicators, determine the overall coverage and effective utilization of the user's computing network resources, and form an initial candidate list of users with the corresponding user ID identifier. Based on the performance indicators of the resource providers, the overall coverage rate and effective utilization rate of the computing network resources of the resource providers are determined, and an initial candidate list of resource providers is formed by combining the corresponding resource provider IDs. The user satisfaction score is determined based on the initial candidate list of users, and the final candidate list of users is generated based on the user satisfaction score and the corresponding resource provider ID. Based on the initial candidate list of resource providers, a satisfaction score for the resource provider is determined, and a final candidate list of resource providers is generated based on the satisfaction score and the corresponding user ID. Based on the final candidate list of users and the final candidate list of resource providers, a stable matching mechanism driven by two-sided satisfaction is adopted to complete the allocation of computing network resources. The stable matching mechanism driven by two-sided satisfaction means that the resource provider proposes a match, the user accepts or rejects the match, and the rejected resource provider continues to propose a match in the next round until one of the two parties has completed the match, and the resource allocation ends.

2. The method according to claim 1, characterized in that, Based on the user's performance metrics, determine the overall coverage and effective utilization rate of the user's computing network resources, including: If the product of the resource provider's network resource indicator and the performance fluctuation rate of the resource provider's network resource indicator is greater than or equal to the upper limit of the user's network resource indicator expectation, or if the resource provider's network resource indicator is less than or equal to the lower limit of the user's network resource indicator expectation, then the overall coverage rate of the user's network resources is 0; otherwise, the overall coverage rate of the user's network resources is determined based on the upper limit of the user's network resource indicator expectation, the lower limit of the user's network resource indicator expectation, the resource provider's network resource indicator, and the performance fluctuation rate of the resource provider's network resource indicator. If the product of the resource provider's network resource index and the performance fluctuation rate of the resource provider's network resource index is greater than or equal to the upper limit of the user's network resource index expectation, or if the resource provider's network resource index is less than or equal to the lower limit of the user's network resource index expectation, then the effective utilization rate of the user's network resources is 0; otherwise, the effective utilization rate of the user's network resources is determined based on the upper limit of the user's network resource index expectation, the lower limit of the user's network resource index expectation, the resource provider's network resource index, and the performance fluctuation rate of the resource provider's network resource index.

3. The method according to claim 1, characterized in that, Based on the resource provider's performance indicators, determine the overall coverage and effective utilization rate of the network resources provided by the resource provider, including: If the product of the resource provider's network resource index and the performance fluctuation rate of the resource provider's network resource index is greater than or equal to the upper limit of the user's network resource index expectation, or if the resource provider's network resource index is less than or equal to the lower limit of the user's network resource index expectation, then the overall coverage rate of the resource provider's network resources is 0; otherwise, the overall coverage rate of the resource provider's network resources is determined based on the upper limit of the user's network resource index expectation, the lower limit of the user's network resource index expectation, the resource provider's network resource index, and the performance fluctuation rate of the resource provider's network resource index. If the product of the resource provider's network resource index and the performance fluctuation rate of the resource provider's network resource index is greater than or equal to the upper limit of the user's network resource index expectation, or if the resource provider's network resource index is less than or equal to the lower limit of the user's network resource index expectation, then the effective utilization rate of the resource provider's network resources is 0; otherwise, the effective utilization rate of the resource provider's network resources is determined based on the upper limit of the user's network resource index expectation, the lower limit of the user's network resource index expectation, the resource provider's network resource index, and the performance fluctuation rate of the resource provider's network resource index.

4. The method according to claim 1, characterized in that, The overall coverage rate and effective utilization rate of computing network resources of the users in the initial candidate list of users are both non-zero values; the overall coverage rate and effective utilization rate of computing network resources of the resource providers in the initial candidate list of resource providers are both non-zero values.

5. The method according to claim 1, characterized in that, Determine user satisfaction scores based on the initial candidate list of users, including: The user satisfaction score is calculated by weighting and summing the products of the overall coverage and effective utilization of network resources for each indicator, with the weights representing the preferences for each indicator.

6. The method according to claim 1, characterized in that, Based on the initial candidate list of resource providers, a resource provider satisfaction score is determined, including: The resource provider's effective utilization rate of computing network resources for each indicator is weighted and summed to calculate the resource provider's satisfaction score, with the weights representing the preference for each indicator.

7. The method according to claim 1, characterized in that, Based on the final candidate lists of users and resource providers, a stable matching mechanism driven by two-sided satisfaction is used to complete the allocation of computing network resources, specifically including: After the final candidate lists for users and resource providers are completed, each unmatched resource provider initiates a matching request to the user with the highest current satisfaction rating in its own final candidate list. After each user receives all matching requests, it selects the resource provider with the highest satisfaction rating from its final candidate list and temporarily accepts the match; the remaining unselected resource providers will be rejected, and the unselected resource providers will remove the rejected users from their final candidate list and obtain an updated final candidate list of resource providers. The user and the resource provider enter a temporary matching state. The rejected resource provider continues to send matching requests to the user with the highest current satisfaction rating in the updated final candidate list of resource providers in the next round. Once one of the participating parties has completed the matching process, the resource allocation ends, and the final matching status between the user and the resource provider is obtained.

8. A computing network resource allocation system, characterized in that, include: The user-side initial candidate list construction unit is used to determine the overall coverage and effective utilization rate of the user-side computing network resources based on the user-side performance indicators, and to form the user-side initial candidate list with the corresponding user-side ID identifier. The initial candidate list construction unit for resource providers is used to determine the overall coverage and effective utilization rate of computing network resources of resource providers based on the performance indicators of resource providers, and to form the initial candidate list of resource providers with the corresponding resource provider ID identifier; The user-side final candidate list construction unit is used to determine the user-side satisfaction score based on the user-side initial candidate list, and generate the user-side final candidate list based on the user-side satisfaction score and the corresponding resource provider ID identifier. The resource provider final candidate list construction unit is used to determine the resource provider satisfaction score based on the initial resource provider candidate list, and generate the resource provider final candidate list based on the resource provider satisfaction score and the corresponding user ID identifier; The resource allocation unit is used to complete the allocation of computing network resources based on the final candidate list of users and the final candidate list of resource providers, using a stable matching mechanism driven by two-sided satisfaction. The stable matching mechanism driven by two-sided satisfaction means that the resource provider proposes a match, the user accepts or rejects the match, and the rejected resource provider continues to propose a match in the next round until one of the two parties has completed the match, at which point the resource allocation ends.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the network resource allocation method as described in any one of claims 1-7.

10. A non-volatile storage medium, characterized in that, Used to store a computer program, wherein the computer program, when executed by a processor, implements the computer network resource allocation method as described in any one of claims 1-7.

Images