Resource allocation method between 5G network slices

Abstract

The invention provides a method for allocating resources among 5G network slices, which comprises the following steps of: firstly, calculating the bandwidth occupied by each service in each network slice based on the number of physical resource blocks occupied by the service and the bandwidth of a single resource block; calculating the theoretical capacity capability of each service according to the Shannon sampling theorem, and calculating the actual service capability at the current moment based on the service congestion rate; calculating the idle bandwidth capability of each slice based on the number of idle resource blocks of each slice and the mathematical expected value of the single resource block capability; calculating the controllable bandwidth capability of each slice based on the service busy ratio and the idle bandwidth capability of each slice; and finally, normalizing the controllable bandwidth capability of each slice to serve as a slice resource allocation priority index for resource allocation.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of 5G communication technology, and particularly relates to a 5G network slice inter-resource allocation method. BACKGROUND

[0002] As a key technology of 5G communication, network slicing can meet the service requirements of high speed, large capacity and ultra-low latency of 5G network, and can provide differentiated and customized services to customers. Among them, how to allocate resources between 5G network slices is involved. By allocating different resources to different types of data streams, higher network efficiency can be achieved while meeting the needs of different businesses.

[0003] The prior art patent with the publication number CN119421176A discloses a network slice access control method based on mixed service types, relating to the field of network slicing, aiming to solve the problem that the network slice access control method does not comprehensively consider the quality of the slice network, comprising: calculating a mixed service adjustment factor according to the average waiting time of all slices; evaluating the link residual capacity according to the residual bandwidth and signal-to-noise ratio of the slice; calculating the quality of service parameter of any slice according to the lower limit of the reference signal received power; calculating the business delay parameter and business throughput parameter of any slice, and determining the qualified set containing the slice; selecting the slice as the target slice for access control according to the comprehensive evaluation value of the slice in the qualified set.

[0004] The existing network slice inter-resource allocation method has many problems. First, the slice marking type is fixed and cannot account for mixed type services. Second, the signal quality level indicator is not refined according to the specific business type and level. Therefore, when allocating resources, the business type of each network slice needs to be evaluated, the business quality state needs to be analyzed, the business busy ratio and idle bandwidth capacity of each slice need to be analyzed, and based on the controllable bandwidth capacity of each slice, the inter-slice resources are dynamically allocated and controlled. SUMMARY

[0005] Each network slice has a certain amount of business. Under the premise of limited 5G network slice resources, differentiated strategies are needed to more intuitively allocate resources between slices. Therefore, the business type of the slice needs to be evaluated, the business quality state needs to be analyzed, the idle resource number needs to be calculated to calculate the Shannon sample, and finally the comprehensive allocation priority needs to be calculated to dynamically allocate and control the inter-slice resources. In this way, the resource needs of different network slices can be guaranteed to meet the index requirements of different businesses.

[0006] To solve the above technical problems, the technical solution provided by the present application is a 5G network slice inter-resource allocation method, comprising the following steps: S1, calculate the bandwidth occupied by each service in each network slice based on the number of physical resource blocks occupied by the service and the bandwidth of a single resource block; S2, calculate the theoretical capacity of each service according to the Shannon sampling theorem, and calculate the actual capacity of the service at the current moment based on the service congestion rate; S3, calculate the idle bandwidth capacity of each slice based on the number of idle resource blocks of each slice and the mathematical expectation value of the single resource block capacity; S4, calculate the controllable bandwidth capacity of each slice based on the service busy ratio of each slice and the idle bandwidth capacity; S5, normalize the controllable bandwidth capacity of each slice as a slice resource allocation priority indicator for resource allocation.

[0007] Specifically, in S1, the signal-to-noise ratio of all services in each slice is converted to a decimal value; then the bandwidth of a single resource block is calculated based on the set subcarrier bandwidth and the number of subcarriers occupied by a resource block, and the bandwidth occupied by each service is calculated by multiplying the bandwidth of a single resource block by the number of physical resource blocks occupied by each service.

[0008] Specifically, in S2, the actual capacity of each service at the current moment is calculated by multiplying the theoretical capacity of each service by 1 minus the difference between the service congestion rate.

[0009] Specifically, in S3, the number of idle resource blocks of each slice is obtained by subtracting the sum of the number of physical resource blocks occupied by each service from the total number of physical resource blocks allocated by the network slice.

[0010] Specifically, in S3, the single resource block capacity of each service is calculated based on the current actual capacity of each service and the number of resource blocks occupied by the service, and the average value of the single resource block capacity of all services on each slice is calculated as the mathematical expectation value of the single resource block capacity of the slice.

[0011] Specifically, in S3, the idle bandwidth capacity of each slice is obtained by multiplying the number of idle resource blocks at the current moment by the mathematical expectation value of the single resource block capacity of the slice.

[0012] Specifically, in S4, for any slice, the service busy ratio of the slice is 1 minus the ratio of the number of idle resource blocks of the slice to the total number of physical resource blocks allocated by the slice.

[0013] Specifically, the controllable bandwidth capacity of each slice is obtained by multiplying the service busy ratio of the slice by the idle bandwidth capacity of the slice calculated in S3.

[0014] Specifically, in S5, the controllable bandwidth capacity of all slices is normalized to the interval [0, 1] and used as a preliminary priority indicator for slice resource allocation.

[0015] Specifically, when resource allocation is performed on each slice, the priority indicators of the slices on which resource allocation is not performed in the last round are dynamically adjusted, the priority indicators of the slices are multiplied by a preset adjustment coefficient in the interval (1, 2], and the PF algorithm of proportional fairness is adopted to allocate the unallocated resources, and the total number of physical resource blocks of all slices is updated.

[0016] The application has the advantages that the bandwidth of a service can be calculated according to the number of physical resource blocks occupied by the service, the performance of all services can be estimated according to the Shannon sampling theorem, the single-resource-block capability in a differentiated environment can be calculated, and the idle bandwidth capability of a slice can be estimated, the controllable bandwidth of each slice can be calculated, resource allocation between slices can be performed according to the calculated dynamic priority, and the services can always be in a good slice environment, thereby providing real-time protection for improving customer experience. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 A flowchart of the method is provided.

[0018] Figure 2 A specific application flowchart of the second embodiment of the application is provided. DETAILED DESCRIPTION

[0019] The application will be described in detail below with reference to the drawings and embodiments.

[0020] Embodiment one: a 5G network slice resource allocation method, as shown in the figure, comprising the following steps: Figure 1 S1, based on the number of physical resource blocks occupied by the service and the single-resource-block bandwidth, the bandwidth occupied by each service in each network slice is calculated; S2, the theoretical capacity capability of each service is calculated according to the Shannon sampling theorem, and the actual capability of the service at the current moment is calculated based on the service congestion rate; S3, based on the number of idle resource blocks of each slice and the mathematical expectation value of the single-resource-block capability, the idle bandwidth capability of each slice is calculated; S4, based on the service busy ratio of each slice and the idle bandwidth capability, the controllable bandwidth capability of each slice is calculated; S5, the controllable bandwidth capability of each slice is normalized as a slice resource allocation priority indicator, and resource allocation is performed.

[0021] Specifically, in S1, the signal-to-noise ratio of all services of each slice is first converted to a decimal value; then the single-resource-block bandwidth is calculated based on the set subcarrier bandwidth and the number of subcarriers occupied by one resource block (RB), and the bandwidth occupied by each service is calculated by multiplying the single-resource-block bandwidth by the number of physical resource blocks occupied by each service.

[0022] ​Specifically, in S2, the actual capacity of each service at the current time is calculated by multiplying the theoretical capacity of each service by 1 minus the congestion rate of the service.

[0023] Specifically, in S3, the number of idle resource blocks of each slice is obtained by subtracting the sum of the number of physical resource blocks occupied by each service from the total number of physical resource blocks allocated by the network slice.

[0024] Specifically, in S3, the single-resource-block capacity of each service is first calculated based on the current actual capacity of each service and the number of resource blocks occupied by the service, and then the average of the single-resource-block capacities of all services on each slice is calculated as the mathematical expectation value of the single-resource-block capacity of the slice.

[0025] Specifically, in S3, the idle bandwidth capacity of each slice is obtained by multiplying the number of idle resource blocks of the slice at the current time by the mathematical expectation value of the single-resource-block capacity of the slice.

[0026] Specifically, in S4, for any slice, the service busy ratio of the slice is 1 minus the ratio of the number of idle resource blocks of the slice to the total number of physical resource blocks allocated by the slice.

[0027] Specifically, the controllable bandwidth capacity of each slice is obtained by multiplying the service busy ratio of the slice by the idle bandwidth capacity of the slice calculated in S3.

[0028] Specifically, in S5, the controllable bandwidth capacities of all slices are normalized to the interval [0, 1] and used as the preliminary priority indicators for slice resource allocation.

[0029] Specifically, each time resource allocation is performed for a slice, the priority indicators of the slices that have not implemented resource allocation in the previous round are dynamically adjusted by multiplying the priority indicators of these slices by a preset adjustment coefficient in the interval (1, 2], and the PF algorithm of proportional fairness is used to allocate the unallocated resources, and the total number of physical resource blocks of all slices is updated.

[0030] In the specific implementation process, for a plurality of network slices, use i to represent the network slice number and j to represent the number of services in the network slice. First, obtain the basic information of each network slice, including the number of allocated resource blocks (total RB number) of each network slice, the number of resource blocks occupied by each service on the network slice, the signal-to-noise ratio of each service, the congestion rate of each service, the subcarrier bandwidth of the network slice, and the number of subcarriers occupied by a resource block.

[0031] After obtaining the basic information required for subsequent calculation, such as Figure 2As shown, first traverse all network slices from i = 1, and traverse all services in each network slice from j = 1. For each service j, first convert the signal-to-noise ratio of the service j into a decimal, then calculate the single resource block bandwidth based on the set subcarrier bandwidth and the number of subcarriers occupied by a resource block, then calculate the occupied bandwidth of the service j by multiplying the single resource block bandwidth by the number of physical resource blocks occupied by each service, then calculate the theoretical capacity of the service j based on the Shannon sampling theorem, and calculate the actual capacity of the service j according to the congestion rate of the service j. Based on the above calculation results, the single RB capacity of the service j is calculated according to the actual capacity of the service j and the number of resource blocks occupied by the service j. Thus, the data required by the service j is calculated, and j is incremented. According to the above steps, the calculation of all service-related data in a slice is completed, and the expected value of the single RB capacity of the slice i and the number of idle RBs of the slice i are calculated after traversing all services in the slice i. Thus, the residual bandwidth capacity of the slice i is calculated, and the busy ratio and the actual residual bandwidth of the slice i are obtained. After completing the traversal calculation of all service data in the slice i and the calculation of the data of the slice i, i is incremented until all network slices are traversed, and the related data of all network slices is obtained.

[0032] After traversing all network slices, i is reset to 1, the actual residual bandwidth of the slice i is normalized, and the preliminary priority index of the slice i is set. Then, the slice i is judged. If the slice i is the slice that is starved in the last round of resource allocation, the priority index of the slice i is adjusted according to the preset adjustment coefficient. If the slice i is not the slice that is starved in the last round of resource allocation, the preliminary priority index of the slice i is maintained. After the judgment is completed, i is incremented until all network slices are traversed. After the traversal is completed, the PF algorithm is referenced according to the priority index of each slice to perform dynamic resource allocation.

[0033] Embodiment two: a 5G network slice resource allocation method, comprising the following steps: S1, based on the number of physical resource blocks occupied by the service and the single resource block bandwidth, calculating the bandwidth occupied by each service in each network slice; S2, calculating the theoretical capacity of each service according to the Shannon sampling theorem, and calculating the actual capacity of the service at the current time based on the congestion rate of the service; S3, calculating the idle bandwidth capacity of each slice based on the number of idle resource blocks of each slice and the mathematical expectation of the single resource block capacity; S4, calculating the controllable bandwidth capacity of each slice based on the service busy ratio and the idle bandwidth capacity of each slice; S5, normalizing the controllable bandwidth capacity of each slice as a slice resource allocation priority index, and performing resource allocation.

[0034] Specifically, in S1, the signal-to-noise ratio of each slice is converted into a decimal value; then the single resource block bandwidth is calculated based on the set subcarrier bandwidth and the number of subcarriers occupied by a single resource block; and the bandwidth occupied by each service is calculated by multiplying the single resource block bandwidth by the number of physical resource blocks occupied by each service.

[0035] Specifically, in S2, the actual capacity of each service at the current time is calculated by multiplying the theoretical capacity of each service by 1 minus the difference between the congestion rate of the service.

[0036] Specifically, in S3, the number of idle resource blocks of each slice is obtained by subtracting the sum of the number of physical resource blocks occupied by each service from the total number of physical resource blocks allocated by the network slice.

[0037] Specifically, in S3, the single resource block capacity of each service is calculated based on the actual capacity of each service at the current time and the number of resource blocks occupied by each service; and the average value of the single resource block capacity of all services on each slice is calculated as the mathematical expectation value of the single resource block capacity of the slice.

[0038] Specifically, in S3, the idle bandwidth capacity of each slice is obtained by multiplying the number of idle resource blocks at the current time by the mathematical expectation value of the single resource block capacity of the slice.

[0039] Specifically, in S4, for any slice, the service busy ratio of the slice is 1 minus the ratio of the number of idle resource blocks of the slice to the total number of physical resource blocks allocated by the slice.

[0040] Specifically, the controllable bandwidth capacity of each slice is obtained by multiplying the service busy ratio of the slice by the idle bandwidth capacity of the slice calculated in S3.

[0041] Specifically, in S5, the controllable bandwidth capacity of all slices is normalized to the interval [0, 1] and used as the preliminary priority indicator for slice resource allocation.

[0042] Specifically, each time resource allocation is performed for a slice, the priority indicators of the slices that have not implemented resource allocation in the previous round are dynamically adjusted by multiplying the priority indicators of these slices by a preset adjustment coefficient in the interval (1, 2]; the PF algorithm of proportional fairness is used to allocate the unallocated resources; and the total number of physical resource blocks of all slices is updated.

[0043] In the specific implementation process, for a plurality of network slices, use i to represent the network slice number and j to represent the number of services in the network slice. First, obtain the basic information of each network slice, including the number of allocated resource blocks (total RB number) of each network slice, the number of resource blocks occupied by each service on the network slice, the signal-to-noise ratio of each service, the congestion rate of each service, the subcarrier bandwidth of the network slice, and the number of subcarriers occupied by a single resource block.

[0044] After obtaining the basic information required for subsequent calculation, as shown in Figure 2 shown, starting from i = 1, all network slices are traversed, and starting from j = 1, a plurality of services in each network slice are traversed. For each service j, the signal-to-noise ratio of the service j is first converted to decimal, and then the single resource block bandwidth is calculated based on the set subcarrier bandwidth and the number of subcarriers occupied by a resource block. Then, the occupied bandwidth of the service j is calculated by multiplying the single resource block bandwidth by the number of physical resource blocks occupied by each service. Then, the theoretical capacity of the service j is calculated based on the Shannon sampling theorem, and the actual capacity of the service j is calculated according to the congestion rate of the service j. Based on the above calculation results, the single RB capacity of the service j is calculated according to the actual capacity of the service j and the number of resource blocks occupied by the service j. Thus, the data required for the service j is calculated, and j is incremented. According to the above steps, the calculation of all service-related data in a slice is completed, and the traversal of all services in the slice i is completed. The expected value of the single RB capacity of the slice i and the number of idle RBs of the slice i are calculated, and thus the residual bandwidth capacity of the slice i is calculated, and the busy ratio and the actual residual bandwidth of the slice i are obtained. After the traversal calculation of all service data in the slice i and the calculation of the data of the slice i are completed, i is incremented, and all network slices are traversed to obtain the related data of all network slices.

[0045] After traversing all network slices, i is reset to 1, the actual residual bandwidth of the slice i is normalized, and the preliminary priority index of the slice i is set. Then, the slice i is judged. If the slice i is the slice that is starved in the last round of resource allocation, the priority index of the slice i is adjusted according to the preset adjustment coefficient. If the slice i is not the slice that is starved in the last round of resource allocation, the preliminary priority index of the slice i is maintained. After the judgment is completed, i is incremented until all network slices are traversed. After the traversal is completed, the PF algorithm is used according to the priority index of each slice to perform dynamic resource allocation.

[0046] In this embodiment, the 5G network slice resource allocation method provided by the present application is simulated in a simulation environment. Three network slices are constructed to simulate a conventional network environment, and the resource allocation between the network slices is performed by the method provided by the present application.

[0047] The specific process is as follows: resource allocation is performed on three network slices, wherein the number of resource blocks allocated to network slice 1 is 40, and there are 4 services on network slice 1, i.e., services 1-1 to 1-4. Among them, service 1-1 occupies 8 resource blocks, the signal-to-noise ratio is -2 dB, and the service congestion rate is 0.5%; service 1-2 occupies 12 resource blocks, the signal-to-noise ratio is 0 dB, and the service congestion rate is 0.2%; service 1-3 occupies 4 resource blocks, the signal-to-noise ratio is 3 dB, and the service congestion rate is 0.8%; and service 1-4 occupies 6 resource blocks, the signal-to-noise ratio is 1 dB, and the service congestion rate is 0.3%.

[0048] The number of resource blocks allocated to network slice 2 is 34, and there are 2 services on network slice 2, i.e., services 2-1 and 2-2. Among them, service 2-1 occupies 7 resource blocks, the signal-to-noise ratio is -1 dB, and the service congestion rate is 0.5%; and service 2-2 occupies 12 resource blocks, the signal-to-noise ratio is 4 dB, and the service congestion rate is 1.1%.

[0049] The number of resource blocks allocated to network slice 3 is 26, and there are 3 services on network slice 3, i.e., services 3-1 to 3-3. Among them, service 3-1 occupies 4 resource blocks, the signal-to-noise ratio is 3 dB, and the service congestion rate is 0.5%; service 3-2 occupies 2 resource blocks, the signal-to-noise ratio is 4 dB, and the service congestion rate is 0.4%; and service 3-3 occupies 8 resource blocks, the signal-to-noise ratio is -2 dB, and the service congestion rate is 0.6%.

[0050] The specific working frequency of network slices 1 to 3 is 2.6 GHz, the working bandwidth is 100 MHz, the subcarrier bandwidth index is 1, the subcarrier bandwidth is 30 kHz, one resource block occupies 12 subcarriers, the preset adjustment coefficient of resource allocation is 1.1, and the last round of the starved slice before the implementation of resource allocation is network slice 1 and network slice 3.

[0051] In the above specific application scenario, dynamic resource allocation is performed on the three network slices. First, the signal-to-noise ratios of all services of the network slices are converted into decimal values, i.e., In network slice 1, the signal-to-noise ratios of services 1-1 to 1-4 are 0.63, 1, 2, and 1.26, respectively; In network slice 2, the signal-to-noise ratios of services 2-1 to 2-2 are 0.79 and 2.51, respectively; In network slice 3, the signal-to-noise ratios of services 3-1 to 3-3 are 2, 2.51, and 0.63, respectively.

[0052] According to the known subcarrier bandwidth and the number of subcarriers occupied by one resource block, the bandwidth of a single resource block is calculated to be 0.35 (MHz). On this basis, the bandwidth occupied by each service at the current time is calculated by multiplying the bandwidth of a single resource block by the number of resource blocks occupied by the service, i.e., In network slice 1, the current time bandwidth occupied by services 1-1 to 1-4 is respectively: 2.88 MHz, 4.32 MHz, 1.44 MHz, 2.16 MHz; In network slice 2, the current time bandwidth occupied by services 2-1 to 2-2 is respectively: 2.52 MHz, 4.32 MHz; In network slice 3, the current time bandwidth occupied by services 3-1 to 3-3 is respectively: 1.44 MHz, 0.72 MHz, 2.88 MHz.

[0053] According to the Shannon sampling theorem, the theoretical capacity of all services is calculated based on the bandwidth occupied by each service at the current time and the decimal service signal-to-noise ratio: In network slice 1, the theoretical capacity of services 1-1 to 1-4 is respectively: 1.98 Mbps, 4.22 Mbps, 2.23 Mbps, 2.48 Mbps; In network slice 2, the theoretical capacity of services 2-1 to 2-2 is respectively: 2.08 Mbps, 7.65 Mbps; In network slice 3, the theoretical capacity of services 3-1 to 3-3 is respectively: 2.23 Mbps, 1.27 Mbps, 1.98 Mbps.

[0054] According to the congestion state of the service, the actual capacity of all services at the current time is estimated, and then the actual capacity is divided by the number of resource blocks occupied by the service to calculate the single resource block capacity occupied by any service: In network slice 1, the single resource block capacity of services 1-1 to 1-4 is respectively: 0.25 Mbps, 0.35 Mbps, 0.55 Mbps, 0.41 Mbps; In network slice 2, the single resource block capacity of services 2-1 to 2-2 is respectively: 0.3 Mbps, 0.63 Mbps; In network slice 3, the single resource block capacity of services 3-1 to 3-3 is respectively: 0.55 Mbps, 0.63 Mbps, 0.25 Mbps.

[0055] The average value of the single resource block capacity of each slice is calculated to obtain the mathematical expectation value of the single resource block capacity of each slice: The mathematical expectation value of the single resource block capacity of network slice 1 is: 0.39 Mbps; The mathematical expectation value of the single resource block capacity of network slice 2 is: 0.46 Mbps; The mathematical expectation value of the single resource block capacity of network slice 3 is: 0.48 Mbps.

[0056] The current free resource block numbers of the three slices are 10, 15, and 12 respectively. Based on this, the free bandwidth capacity of each slice is calculated by multiplying the number of resource blocks by the mathematical expectation of the capacity of a single resource block: The free bandwidth capacity of network slice 1 is 3.9 Mbps; The free bandwidth capacity of network slice 2 is 6.94 Mbps; The free bandwidth capacity of network slice 3 is 5.74 Mbps.

[0057] For the three slices, the service busy ratio of each slice is calculated by subtracting 1 from the ratio of the number of free resource blocks to the total number of allocated physical resource blocks: The service busy ratio of network slice 1 is 0.75; The service busy ratio of network slice 2 is 0.56; The service busy ratio of network slice 3 is 0.54.

[0058] The controllable bandwidth capacity of each slice is calculated by multiplying the service busy ratio of each slice by the free bandwidth capacity of the slice: The controllable bandwidth capacity of network slice 1 is 2.93; The controllable bandwidth capacity of network slice 2 is 3.88; The controllable bandwidth capacity of network slice 3 is 3.09.

[0059] Dynamic resource allocation is performed based on the controllable bandwidth capacity: the adjustment coefficient for resource allocation is set to 1.1, and the controllable bandwidth capacities of all slices are normalized to the interval [0, 1] to obtain the slice resource allocation priority index: The slice resource allocation priority index of network slice 1 is 0.3; The slice resource allocation priority index of network slice 2 is 0.39; The slice resource allocation priority index of network slice 3 is 0.31.

[0060] Since network slice 1 and network slice 3 did not implement resource allocation in the last round, the slice resource allocation priority index is adjusted according to the adjustment coefficient, and the slice resource allocation priority index of network slice 1 and network slice 3 is multiplied by the adjustment coefficient to obtain: The adjusted slice resource allocation priority index of network slice 1 is 0.33; The adjusted slice resource allocation priority index of network slice 2 is 0.39; The adjusted slice resource allocation priority index of network slice 3 is 0.34.

[0061] Finally, the adjusted slice resource allocation priority index is taken as the final priority of the current resource allocation, and the PF algorithm of proportional fairness is adopted to allocate the unallocated resources, and the total number of physical resource blocks of all slices is updated.

[0062] Although the present application has been described in detail with reference to the foregoing embodiments, the technical solutions recorded in the foregoing embodiments can be modified or some of the technical features can be replaced by equivalents by those skilled in the art, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for allocating resources between 5G network slices, characterized in that: The following steps are involved: S1. Calculate the bandwidth occupied by each service in each network slice based on the number of physical resource blocks occupied by the service and the bandwidth of a single resource block; S2. Calculate the theoretical capacity of each service based on the Shannon sampling theorem, and calculate the actual capacity of the service at the current moment based on the service congestion rate. S3. Calculate the idle bandwidth capacity of each slice based on the number of idle resource blocks of each slice and the mathematical expectation value of the single resource block capacity; S4. Calculate the controllable bandwidth capacity of each slice based on the service busy ratio and idle bandwidth capacity of each slice; S5. Normalize the controllable bandwidth capabilities of each slice and use it as a slice resource allocation priority indicator for resource allocation.

2. The 5G network inter-slice resource allocation method according to claim 1, characterized in that: In S1, the signal-to-noise ratio of all services in each slice is first converted into a decimal value; then the single resource block bandwidth is calculated based on the set subcarrier bandwidth and the number of subcarriers occupied by a resource block, and the bandwidth occupied by each service is calculated by multiplying the single resource block bandwidth by the number of physical resource blocks occupied by each service.

3. The 5G network inter-slice resource allocation method according to claim 1, characterized in that: In S2, the actual capacity of the service at the current moment is calculated by multiplying the theoretical capacity of each service by 1 and subtracting the difference between the service congestion rate.

4. The method for allocating resources between 5G network slices according to claim 1, wherein: The number of idle resource blocks in each slice in S3 is obtained by subtracting the total number of physical resource blocks occupied by each service from the total number of physical resource blocks allocated to the network slice.

5. The 5G network inter-slice resource allocation method according to claim 1 or 4, characterized in that: In S3, the single resource block capacity of each service is first calculated based on the current actual capacity of each service and the number of resource blocks occupied by the service, and then the average single resource block capacity of all services on each slice is calculated as the mathematical expectation value of the single resource block capacity of the slice.

6. The 5G network inter-slice resource allocation method according to claim 1 or 4, characterized in that: The idle bandwidth capacity of each slice in S3 is calculated by multiplying the number of idle resource blocks of each slice at the current moment by the mathematical expectation value of the single resource block capacity of the slice.

7. The method for allocating resources between 5G network slices according to claim 1, wherein: In S4, for any slice, the service busy ratio of the slice is 1 minus the ratio of the number of idle resource blocks of the slice to the total number of physical resource blocks allocated to the slice.

8. The method for allocating resources between 5G network slices according to claim 1 or 7, wherein: The controllable bandwidth capacity of each slice is obtained by multiplying the service busyness ratio of the slice by the idle bandwidth capacity of the slice calculated in S3.

9. The method for allocating resources between 5G network slices according to claim 1, wherein: In S5, the controllable bandwidth capabilities of all slices are normalized to the interval [0,1] and used as the preliminary priority indicator for slice resource allocation.

10. The method for allocating resources between 5G network slices according to claim 9, wherein: Each time resources are allocated to slices, the priority indicators of slices that were not allocated resources in the previous round are dynamically adjusted. The priority indicators of these slices are multiplied by a preset adjustment coefficient in the interval (1, 2). The proportional fairness PF algorithm is used to allocate the unallocated resources and the total number of physical resource blocks of all slices is updated.

Images