A method for 5g communication system energy efficiency optimization and resource allocation

Abstract

This invention discloses a method for energy efficiency optimization and resource allocation in a 5G communication system. The method includes: acquiring operational status data of multiple cells and multiple service slices in a target 5G communication system during the current scheduling period. The operational status data includes cell resource occupancy data, cell channel quality data, cell actual energy consumption data, slice resource occupancy data, slice service quality data, and slice actual energy consumption data; and generating a cell-level corrected energy consumption benchmark value based on the cell resource occupancy data, cell channel quality data, and cell service quality constraints. This invention relates to the field of mobile communication network energy efficiency management technology. This method, by constructing a corrected energy efficiency deviation fingerprint, correlates the energy efficiency deviation value, deviation duration, resource occupancy status, and service quality response relationship. This allows the identification of resource inefficiency objects to no longer rely solely on the actual energy consumption magnitude, but rather to determine whether the increase in energy consumption brings a corresponding service quality benefit.

Description

Technical Field

[0001] This invention relates to the field of energy efficiency management in mobile communication networks, and more specifically, to a method for energy efficiency optimization and resource allocation in 5G communication systems. Background Technology

[0002] With the development of 5G network slicing, Massive MIMO, and multi-service concurrent transport, while communication systems improve speed, reduce latency, and enhance reliability, they also bring about problems such as base station energy consumption and redundant occupation of slice resources. In existing technologies, CN112235817B discloses a resource allocation method for 5G communication systems, which optimizes system and energy efficiency through user channel association and power allocation; CN110536321B discloses an optimization method for balancing energy efficiency and spectrum efficiency in 5G IoT communication, which optimizes energy and frequency efficiency through MIMO antenna allocation and power allocation; CN117998594B discloses a 5G power multi-service slice resource allocation method, which predicts network resource requirements based on historical data and current service needs and updates the resource mapping table. While these solutions can optimize resource allocation to some extent, they focus primarily on power, antenna, or resource demand prediction, making it difficult to identify inefficient resource objects by combining reasonable energy consumption benchmarks for cells and slices, and lacking an energy debt scheduling mechanism tailored to the rigidity of different services. Summary of the Invention

[0003] The purpose of this invention is to provide a method for energy efficiency optimization and resource allocation in 5G communication systems. This method addresses the energy efficiency issues of existing 5G communication systems. While it can optimize resource allocation to some extent, it focuses primarily on power, antenna, or resource demand prediction. It is difficult to identify inefficient resource objects by combining reasonable energy consumption benchmarks for cells and slices, and it also lacks an energy consumption debt scheduling mechanism for different service rigidity levels.

[0004] This invention achieves the above objective through the following technical solution: a method for energy efficiency optimization and resource allocation in a 5G communication system, the method comprising: The system acquires operational status data of multiple cells and multiple service slices in the target 5G communication system during the current scheduling period. The operational status data includes cell resource occupancy data, cell channel quality data, cell actual energy consumption data, slice resource occupancy data, slice service quality data, and slice actual energy consumption data. A cell-level corrected energy consumption benchmark value is generated based on the cell resource occupancy data, cell channel quality data, and cell service quality constraints. A slice-level corrected energy consumption benchmark value is generated based on the slice resource occupancy data, slice service type, and slice service quality constraints. The deviation between the actual energy consumption data of the community and the community-level corrected energy consumption benchmark value is calculated, the deviation between the actual energy consumption data of the slice and the slice-level corrected energy consumption benchmark value is calculated, and the corrected energy efficiency deviation fingerprint is generated by combining the direction of resource occupation change and the direction of service quality change. Based on the aforementioned energy efficiency deviation correction fingerprint identification, inefficient cells and inefficient service slices are identified. Calculate the energy consumption debt value of inefficient business segments; The priority of resource allocation adjustment is determined based on the energy efficiency deviation fingerprint, sliced ​​energy consumption debt value, business priority, service quality safety margin, and historical debt processing records. According to the priority of resource allocation adjustment, debt repayment-style resource allocation is performed on business slices that meet the debt repayment conditions, and service quality assurance resources are maintained for business slices that meet the rigid business protection conditions. Collect energy consumption change data and service quality change data after debt repayment-based resource allocation, and update the corrected energy efficiency deviation fingerprint, slice energy consumption debt value and resource allocation strategy for the next scheduling cycle.

[0005] Furthermore, the operational status data is jointly acquired by the base station management unit, the slice management unit, and the energy consumption metering unit; The cell resource occupancy data includes physical resource block occupancy rate, carrier bandwidth occupancy, transmit power occupancy, and multiple input multiple output layer occupancy. The cell channel quality data includes channel quality indication, reference signal received power, reference signal received quality, and uplink interference intensity; The slice service quality data includes slice average latency, slice packet loss rate, slice reliability, slice throughput, and slice user connections; The system then performs sampling period alignment, abnormal sampling point removal, and missing sampling point completion on the operational status data to obtain standardized operational status data.

[0006] Furthermore, the generation of the cell-level corrected energy consumption benchmark value includes: The cell state vector is constructed based on the cell service load, cell channel quality, cell coverage constraints, and cell service quality constraints. Filter historical cell samples that match the current cell status vector from historical operational status data; Remove samples from historical community samples that do not meet community service quality constraints; The remaining historical cell samples are sorted from low to high according to their energy consumption values, and the energy consumption values ​​that meet the preset percentile positions are determined as the cell-level correction energy consumption benchmark values.

[0007] Furthermore, the generation of the slice-level corrected energy consumption reference value includes: Construct a slice state vector based on slice service type, slice physical resource block usage, slice transmit power usage, slice multi-input multi-output layer usage, and slice quality of service constraints; Filter historical slice samples from historical slice scheduling data that match the current slice state vector; Remove samples from historical slice samples that do not meet the corresponding slice service quality constraints; The lowest energy consumption quantile in the remaining historical slice samples is determined as the slice-level corrected energy consumption benchmark.

[0008] Furthermore, the corrected energy efficiency deviation fingerprint includes a deviation intensity field, a deviation duration field, a deviation source field, a resource usage field, and a service quality response field; The deviation intensity field is used to record the normalized deviation of the actual energy consumption from the corresponding corrected energy consumption benchmark value. The deviation persistence field is used to record the number of scheduling cycles in which the energy efficiency deviation continuously exceeds the dynamic threshold. The deviation source field is used to record cell-side energy consumption deviation, slice-side energy consumption deviation, or cell-slice coupling energy consumption deviation. The resource occupancy field is used to record changes in the occupancy of physical resource blocks, transmit power, and multiple input multiple output layers; The service quality response field is used to record whether service quality has improved, remained stable, or declined.

[0009] Furthermore, the identification of resource-inefficient cells and resource-inefficient service slices includes: When the energy efficiency deviation of the target cell exceeds the dynamic threshold of the cell in multiple consecutive scheduling cycles, and the throughput, coverage quality and edge user experience of the target cell do not reach the preset improvement range, the target cell will be identified as a resource-inefficient cell. When the slice energy efficiency deviation value of the target service slice exceeds the slice dynamic threshold in multiple consecutive scheduling cycles, and the average latency, packet loss rate, reliability and throughput of the target service slice do not reach the preset improvement range, the target service slice will be identified as a resource-inefficient service slice. The cell dynamic threshold and slice dynamic threshold are determined based on the median and absolute deviation of the corresponding energy efficiency deviation values ​​within the sliding time window.

[0010] Furthermore, the slice energy consumption debt value is determined based on the slice energy consumption excess, excess resource utilization, service quality revenue value, business priority correction coefficient, and business rigidity correction coefficient; Wherein, the excess energy consumption of the slice is the difference between the actual energy consumption data of the slice and the corrected energy consumption benchmark value at the slice level; The excess resource usage includes the physical resource block usage, transmit power usage, and multiple input multiple output layer usage that exceed the slice service quality constraints. The service quality benefit value is determined based on the reduction in latency, the reduction in packet loss rate, the improvement in reliability, and the improvement in throughput. The business priority correction coefficient and business rigidity correction coefficient are used to reduce the debt repayment intensity of high-priority business slices and rigid business slices in the current scheduling cycle.

[0011] Furthermore, the priority of resource allocation adjustment is determined jointly by energy efficiency anomaly level, debt status level, and business protection level; The energy efficiency anomaly level is determined by the deviation intensity field and the deviation duration field in the corrected energy efficiency deviation fingerprint. The debt status level is determined by the slice energy consumption debt value, the number of historical debt repayments, and the number of historical debt delays. The business protection level is determined by business priority, service quality security margin, and risk of user experience degradation. The business slices are sorted in order of energy efficiency anomaly level from high to low, debt status level from high to low, and business protection level from low to high to obtain the priority of resource allocation adjustment.

[0012] Furthermore, the debt-repayment-based resource allocation includes: When the energy consumption debt value of an inefficient business slice exceeds the debt repayment trigger threshold, and the inefficient business slice does not meet the rigid business protection conditions, resource compression processing is initiated. The resource compression process includes one or more of the following: reducing the allocation ratio of non-critical physical resource blocks, reducing redundant transmit power, reducing unnecessary multiple input multiple output layers, reducing the scheduling frequency of deferred services, and migrating deferred services. During the resource compression process, the average latency, packet loss rate, reliability, and throughput of the corresponding business slice are continuously monitored. When any service quality indicator reaches the corresponding service quality constraint boundary, the resource compression process for that business slice will be stopped, and the energy consumption reduction generated by the resource compression process will be recorded as the debt repayment amount.

[0013] Furthermore, the updates to the rigid business protection conditions and resource allocation strategies include: When a service slice has low latency constraints, high reliability constraints, or continuous service guarantee constraints, and the service quality security margin of the service slice is less than the security margin threshold, the service slice is determined to meet the rigid service protection conditions. For business slices that meet the rigid business protection conditions, maintain the minimum quality of service guarantee resources and record their slice energy consumption debt value as the delayed debt value; After the current scheduling cycle ends, collect data on actual energy consumption changes, resource usage changes, and service quality changes. When resource compression reduces actual energy consumption and service quality indicators meet the corresponding service quality constraints, the corresponding resource adjustment parameters are recorded as an effective debt repayment strategy. When resource compression processing causes any service quality indicator to fail to meet the corresponding service quality constraint, the corresponding resource adjustment parameter will be recorded as an invalid debt repayment strategy. Based on the effective debt repayment strategy, ineffective debt repayment strategy, debt repayment amount, and delayed debt value, update the resource allocation strategy for the next scheduling cycle.

[0014] The beneficial effects of this invention are as follows: 1. By using cell-level and slice-level energy consumption benchmark values ​​to characterize the reasonable energy consumption level under different operating conditions, the energy consumption can be evaluated and corrected in combination with service load, channel quality, coverage constraints, and service quality constraints, thereby reducing misjudgments of resource inefficiency caused by changes in service volume.

[0015] 2. By constructing a correction energy efficiency deviation fingerprint, the energy efficiency deviation value, deviation duration, resource occupancy status, and service quality response relationship are correlated, so that the identification of resource inefficiency objects no longer depends solely on the actual energy consumption, but rather on whether the increase in energy consumption brings a corresponding service quality benefit.

[0016] 3. This invention introduces a slice energy consumption debt value, which incorporates the actual excess energy consumption of a slice, the degree of resource redundancy, service quality benefits, business priority, and business rigidity into a unified evaluation, so that resource adjustments for different business slices have a comparable quantitative basis.

[0017] 4. This invention adopts a debt repayment-based resource allocation method, which gradually reduces the allocation ratio of non-critical physical resource blocks, redundant transmit power, or unnecessary multi-input multi-output layers for business slices that meet the debt repayment conditions, and stops compression when the service quality approaches the constraint boundary, thereby balancing energy consumption reduction and service quality maintenance.

[0018] 5. This invention sets up a rigid business protection mechanism to maintain minimum service quality guarantee resources for low-latency, high-reliability, or continuous business guarantee slices, and delays the corresponding slice energy consumption debt value to subsequent scheduling cycles for processing, which can avoid the energy efficiency optimization process from affecting the continuity of critical businesses.

[0019] 6. Based on the actual energy consumption changes and service quality changes after debt repayment-based resource allocation, this invention updates the corrected energy efficiency deviation fingerprint, sliced ​​energy consumption debt value, resource allocation adjustment priority, and debt repayment trigger threshold, forming a closed-loop optimization process and improving the adaptability of resource allocation strategies in subsequent scheduling cycles. Attached Figure Description

[0020] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart illustrating the overall process of the 5G communication system energy efficiency optimization and resource allocation method of the present invention. Figure 2 This is a schematic diagram of the module structure of the 5G communication system energy efficiency optimization and resource allocation system of the present invention; Figure 3 This is a flowchart of the data acquisition and preprocessing process for the operating status of this invention; Figure 4 This is a flowchart illustrating the process of generating the energy consumption benchmark value for this invention. Figure 5 This is a flowchart of the energy efficiency deviation correction fingerprint construction and resource inefficient object identification process of the present invention; Figure 6 This is a flowchart illustrating the closed-loop update process of energy consumption debt repayment and resource allocation in this invention. Detailed Implementation

[0021] The present application will now be described in further detail with reference to the accompanying drawings. It should be noted that the following specific embodiments are only used to further illustrate the present application and should not be construed as limiting the scope of protection of the present application. Those skilled in the art can make some non-essential improvements and adjustments to the present application based on the above application content.

[0022] Example 1: Please refer to Figure 1-6 This invention provides a method for energy efficiency optimization and resource allocation in a 5G communication system, applicable to 5G communication systems comprising multiple cells and multiple service slices. Service slices include enhanced mobile broadband service slices, low-latency service slices, high-reliability service slices, and IoT connectivity service slices. The method is executed collaboratively by a base station management unit, a slice management unit, an energy consumption metering unit, and edge computing nodes. Specifically, the base station management unit acquires cell resource occupancy data and channel quality data; the slice management unit acquires service slice resource occupancy data and service quality data; the energy consumption metering unit acquires actual energy consumption data for cells and slices; and the edge computing nodes perform energy efficiency deviation identification, slice energy consumption debt calculation, and resource allocation strategy updates.

[0023] This embodiment uses a scheduling cycle as the data processing unit, and the scheduling cycle is denoted as . In one implementation, Take 15 minutes; when the network management platform already has a fixed energy consumption statistics granularity, The energy consumption statistics period of this platform is used. Both the actual energy consumption values ​​of the cell and the actual energy consumption values ​​of the slice are uniformly converted to a scheduling period. The energy value within, in units of Transmit power data is converted into energy consumption data based on the scheduling cycle; physical resource block occupancy, MIMO layer occupancy, average latency, packet loss rate, reliability, and throughput are normalized according to the corresponding resource requirements or service quality targets, so that the data used for subsequent comparison are converted into dimensionless values ​​before calculation, avoiding the direct addition of data with different dimensions.

[0024] The system acquires operational status data for multiple cells and service slices in the target 5G communication system within the current scheduling period. This operational status data includes cell resource occupancy data, cell channel quality data, actual cell energy consumption data, slice resource occupancy data, slice service quality data, and slice actual energy consumption data. Cell resource occupancy data includes physical resource block occupancy rate, carrier bandwidth occupancy, transmit power occupancy, and MIMO tier occupancy. Cell channel quality data includes channel quality indication, reference signal received power, reference signal received quality, and uplink interference intensity. Slice service quality data includes slice average latency, slice packet loss rate, slice reliability, slice throughput, and slice user connections. After collection, the system performs sampling period alignment, abnormal sampling point removal, and missing sampling point completion on the above operational status data to obtain standardized operational status data.

[0025] When normalizing the collected operational status data, the following expression is used: , in, Indicates the first Class operation status indicators in the first The original value within each scheduling cycle; This represents the normalized index value, which is dimensionless. and They represent the first The lower and upper limits of the category indicator; To prevent extremely small positive numbers with a denominator of zero, take When the number of historical samples is no less than 1000 scheduling cycles, Take the 1st percentile value from the historical sample. The 99th percentile value from the historical samples is used to reduce the impact of outliers on the normalization results; when the number of historical samples is less than 1000 scheduling cycles, and Take the minimum and maximum allowable values ​​recorded in the device operation specifications or network management platform configuration file.

[0026] Cell-level and slice-level corrected energy consumption baseline values ​​are generated based on standardized operational status data. For the target cell, a cell state vector is constructed based on cell service load, cell channel quality, cell coverage constraints, and cell quality of service constraints. For the target service slice, a slice state vector is constructed based on slice service type, slice physical resource block occupancy, slice transmit power occupancy, slice MIMO layer occupancy, and slice quality of service constraints. For uniform description, the cell or service slice is denoted as the target object, and its corrected energy consumption baseline value is determined according to the following expression:

[0027] in, Represents the target object The corrected energy consumption benchmark value, in units of When the target object When it was a community, This serves as the baseline value for energy consumption correction at the community level; when the target object When slicing business data, This serves as a baseline value for energy consumption correction at the slice level. Representing historical samples target object Energy consumption value, in units of ; This represents the state vector of the current scheduling cycle; Representing historical samples The state vector; This represents the normalized distance between two state vectors; Indicates the sample matching threshold; This indicates the service quality satisfaction status in the historical samples; This represents the service quality constraints corresponding to the target object. Represents the quantile function; Indicates quantile.

[0028] The sample matching threshold is automatically determined based on historical samples. Specifically, the 30th percentile of the state vector distance is calculated among historical samples from the same hourly period within the last 7 days, and this value is used as the sample matching threshold. When the number of matching samples is less than 30, the sample range is expanded to the same hourly period within the last 14 days, and the 30th percentile is recalculated. The quantile is set to 0.25, meaning that the lowest energy consumption quantile among historical samples that meet the service quality constraints is selected as the corrected energy consumption benchmark. This rule avoids directly using the lowest energy consumption sample, which would result in an excessively low benchmark, and it better reflects the energy-saving scheduling target compared to the median energy consumption sample.

[0029] The deviation between the actual energy consumption data of the residential area and the corrected energy consumption benchmark value at the residential area level is calculated, and the deviation between the actual energy consumption data of the slice and the corrected energy consumption benchmark value at the slice level is also calculated to obtain the corresponding energy efficiency deviation value. For any target object, its energy efficiency deviation value is determined according to the following expression:

[0030] in, Represents the target object The energy efficiency deviation value is a dimensionless value. Represents the target object The actual energy consumption value within the current scheduling cycle, in units of ; Represents the target object The corrected energy consumption benchmark value, in units of ; For a very small positive number, take .when When, it indicates that the actual energy consumption of the target object is higher than the reasonable energy consumption level under similar operating conditions; when When this time, it indicates that the current energy consumption of the target object has not exceeded the corrected energy consumption baseline value. Since both the numerator and denominator are... ,therefore It is a dimensionless value and can be compared with a dynamic threshold.

[0031] After obtaining the energy efficiency deviation value, a corrected energy efficiency deviation fingerprint is generated by combining the direction of resource usage change and the direction of service quality change. The corrected energy efficiency deviation fingerprint includes a deviation intensity field, a deviation duration field, a deviation source field, a resource usage field, and a service quality response field. The deviation intensity field records the normalized deviation of actual energy consumption from the corresponding corrected energy consumption benchmark value; the deviation duration field records the number of scheduling cycles in which the energy efficiency deviation continuously exceeds the dynamic threshold; the deviation source field records the energy consumption deviation on the cell side, the energy consumption deviation on the slice side, or the energy consumption deviation coupled between the cell and the slice; the resource usage field records the changes in the usage of physical resource blocks, transmit power, and multiple input multiple output layers; the service quality response field records service quality improvement, service quality stability, or service quality decline. Service quality improvement refers to at least one of the following: reduction in average latency, reduction in packet loss rate, improvement in reliability, or improvement in throughput reaching more than 5% of the corresponding service quality target; if this level is not reached, the service quality is considered stable; when any service quality indicator is lower than the service quality constraint, the service quality is considered to have declined.

[0032] To avoid the problem that a fixed threshold cannot adapt to business fluctuations at different times, this embodiment determines a dynamic threshold based on the energy efficiency deviation distribution within a sliding time window, as shown in the following expression:

[0033] in, Represents the target object The dynamic threshold is a dimensionless value. Represents the target object In the sliding time window The set of energy efficiency deviation values ​​within; Represents the median function; Indicates the deviation amplification factor. Sliding time window. Take the most recent 12 scheduling cycles; when the scheduling cycle... When it is 15 minutes, This corresponds to the operating status of the last 3 hours. The initial value is 3.0; if the false trigger rate of resource-inefficient objects exceeds 5% in the past 24 hours, then... Increase by 0.5, with a maximum of 4.0; if the system checks and finds two consecutive missed detections of confirmed inefficient resource objects, then... The threshold is reduced by 0.5, but not lower than 2.5. Therefore, the dynamic threshold can be adjusted according to historical fluctuations and has a defined update rule.

[0034] Resource-inefficient cells and service slices are identified based on energy efficiency deviation fingerprinting. A target cell is identified as a resource-inefficient cell when its energy efficiency deviation exceeds the cell dynamic threshold for at least two consecutive scheduling cycles, and its throughput, coverage quality, and edge user experience do not meet the aforementioned service quality improvement rules. Similarly, a target service slice is identified as a resource-inefficient service slice when its slice energy efficiency deviation exceeds the slice dynamic threshold for at least two consecutive scheduling cycles, and its average latency, packet loss rate, reliability, or throughput do not meet the aforementioned service quality improvement rules.

[0035] For identified resource-inefficient business slices, the slice energy debt value is further calculated. This embodiment does not use a direct summation of multiple weights, but rather bases the calculation on the degree of energy excess, adjusting for resource redundancy, service quality benefits, business priority, and business rigidity. The slice energy debt value for resource-inefficient business slices is determined according to the following expression:

[0036] in, Indicates business slice The energy consumption debt value of the slice is a dimensionless value; Indicates business slice The slice energy efficiency deviation value; Indicates business slice Resource redundancy coefficient; Indicates business slice The service quality revenue coefficient; This represents the business protection correction factor.

[0037] The resource redundancy coefficient is determined based on the maximum value among physical resource block redundancy, transmit power redundancy, and MIMO layer redundancy. Specifically, it takes the maximum normalized excess ratio among the three redundancy indicators, without subjective weighting. Physical resource block redundancy is determined by the proportion by which the actual allocated quantity exceeds the quantity required to meet quality of service constraints; transmit power redundancy is determined by the proportion by which the actual transmit power exceeds the transmit power required to meet quality of service constraints; and MIMO layer redundancy is determined by the proportion by which the actual activated layers exceed the layers required to meet quality of service constraints.

[0038] The service quality benefit coefficient ranges from 0 to 1. When increased energy consumption in a service slice simultaneously leads to reduced average latency, reduced packet loss rate, improved reliability, or increased throughput, the service quality benefit coefficient is determined based on the service quality indicator with the greatest improvement, and its upper limit is set to 1. When the service quality indicator does not improve or the improvement does not reach 5% of the corresponding service quality target, the service quality benefit coefficient is set to 0. This rule distinguishes between two scenarios: increased energy consumption leading to effective business benefits and increased energy consumption without significant service quality benefits.

[0039] The service protection correction factor is obtained by multiplying the service priority factor and the service rigidity factor. For ordinary enhanced mobile broadband service slices, the service priority factor is 1.0; for key protection service slices, it is 1.25; and for emergency communication or private network protection service slices, it is 1.5. For non-rigid service slices, the service rigidity factor is 1.0; for service slices with continuous service guarantee requirements, it is 1.25; and for service slices with low latency or high reliability hard constraints, it is 1.5. These factors are determined based on the service level agreement of the service slice, the operator's slice configuration file, or the service level registered on the network management platform, and are not temporarily adjusted during the scheduling process. Because the service protection correction factor is in the denominator, the debt repayment pressure on high-priority and highly rigid service slices is correspondingly reduced.

[0040] Resource allocation adjustment priorities are determined based on the energy efficiency deviation fingerprint, slice energy consumption debt value, business priority, service quality safety margin, and historical debt processing records. Specifically, the energy efficiency anomaly level is determined by the deviation intensity and deviation duration fields in the energy efficiency deviation fingerprint; the debt status level is determined by the slice energy consumption debt value, historical debt repayment count, and historical debt delay count; and the business protection level is determined by business priority, service quality safety margin, and user experience degradation risk. Then, each business slice is sorted in descending order of energy efficiency anomaly level, descending order of debt status level, and ascending order of business protection level to obtain the resource allocation adjustment priorities. The higher the business protection level, the lower the resource compression priority of the corresponding business slice in the current scheduling cycle.

[0041] After determining the priority of resource allocation adjustments, debt repayment-based resource allocation is performed on business slices that meet the debt repayment conditions but not the rigid business protection conditions. The resource compression ratio is determined according to the following expression:

[0042] in, Indicates business slice The resource compression ratio within the current scheduling cycle; Indicates the maximum resource compression ratio allowed in a single scheduling cycle; Indicates the debt conversion factor; Indicates business slice The energy consumption debt value of the slice; This indicates the safe compression ratio allowed without exceeding the service quality constraints.

[0043] In one implementation, the maximum allowed resource compression ratio for a single scheduling cycle is set to 15% to avoid excessive resource fluctuations within a single scheduling cycle; the debt conversion coefficient is set to 0.2, meaning that as the energy consumption debt value of a slice increases, the resource compression ratio increases by a predetermined percentage. The security compression ratio is determined based on the service quality safety margin. For upper limit constraints such as average latency and packet loss rate, the service quality safety margin is determined by the ratio of the difference between the upper limit of the constraint and the current indicator value to the upper limit of the constraint; for lower limit constraints such as reliability and throughput, the service quality safety margin is determined by the ratio of the difference between the current indicator value and the lower limit of the constraint to the lower limit of the constraint. If the safety margin of any service quality indicator is less than 10%, the security compression ratio does not exceed 5%; if the safety margin of all service quality indicators is not less than 10% and any indicator is less than 20%, the security compression ratio does not exceed 10%; if the safety margin of all service quality indicators is not less than 20%, the security compression ratio does not exceed 15%. For low-latency service slices, high-reliability service slices, or continuous service assurance service slices, the safety margin threshold is increased to 20%.

[0044] Debt-paying resource allocation includes one or more of the following: reducing the allocation ratio of non-critical physical resource blocks, reducing redundant transmit power, reducing unnecessary MIMO layers, reducing the scheduling frequency of deferred services, and migrating deferred services. During resource compression, the average latency, packet loss rate, reliability, and throughput of the corresponding service slice are continuously monitored. When any service quality indicator reaches the corresponding service quality constraint boundary, resource compression processing for that service slice is stopped, and the energy consumption reduction resulting from resource compression is recorded as debt repayment.

[0045] For service slices that meet the rigid service protection conditions, service quality assurance processing is performed. When a service slice has low latency constraints, high reliability constraints, or continuous service assurance constraints, and its service quality safety margin is less than the corresponding safety margin threshold, the service slice is determined to meet the rigid service protection conditions. For service slices that meet the rigid service protection conditions, minimum service quality assurance resources are maintained. Resource compression processing that reduces critical physical resource blocks, critical transmit power, or critical MIMO layers is not performed, and its slice energy consumption debt value is recorded as latency debt value. The minimum service quality assurance resources are the lower limit of resources required to meet the corresponding average latency, packet loss rate, reliability, and throughput constraints.

[0046] The debt repayment trigger threshold is determined based on the median and absolute deviation of the energy consumption debt values ​​of all service slices over the most recent 12 scheduling cycles, and its update rule is consistent with that of the dynamic threshold. To prevent frequent resource compression triggers, debt repayment-based resource allocation is only initiated when the energy consumption debt value of a slice exceeds the debt repayment trigger threshold for at least two consecutive scheduling cycles.

[0047] After the current scheduling cycle ends, data on actual energy consumption changes, resource usage changes, and service quality changes are collected. When resource compression reduces actual energy consumption and service quality indicators still meet the corresponding service quality constraints, the corresponding resource adjustment parameters are recorded as an effective debt repayment strategy. When resource compression causes any service quality indicator to fail to meet the corresponding service quality constraints, the corresponding resource adjustment parameters are recorded as an invalid debt repayment strategy, and the service protection level of that service slice is increased in subsequent scheduling cycles. Based on the effective debt repayment strategy, invalid debt repayment strategy, debt repayment amount, and delayed debt value, the corrected energy efficiency deviation fingerprint, slice energy consumption debt value, resource allocation adjustment priority, and debt repayment trigger threshold for the next scheduling cycle are updated, thus forming a closed-loop scheduling process for energy efficiency optimization and resource allocation of 5G communication systems.

[0048] Example 2: In a specific application scenario, the target 5G communication system includes three cells and four types of service slices: enhanced mobile broadband service slice, low-latency service slice, high-reliability service slice, and IoT connectivity service slice. During each scheduling cycle, the base station management unit reports physical resource block occupancy, transmit power occupancy, MIMO layer occupancy, channel quality indicator, and actual cell energy consumption data to the edge computing nodes; the slice management unit reports the average latency, packet loss rate, reliability, throughput, number of user connections, and actual slice energy consumption data for each service slice to the edge computing nodes.

[0049] When an enhanced mobile broadband service slice has high physical resource block occupancy and high transmit power occupancy within two consecutive scheduling cycles, but its throughput improvement does not reach 5% of the service quality target, the deviation intensity and deviation persistence fields in the energy efficiency deviation fingerprint increase. Edge computing nodes calculate the slice energy efficiency deviation value based on the slice's actual energy consumption data and the slice-level corrected energy consumption benchmark value. If this deviation value exceeds the slice's dynamic threshold, the enhanced mobile broadband service slice is identified as a resource-inefficient service slice.

[0050] Edge computing nodes calculate the slice energy consumption debt value based on the slice energy efficiency deviation, resource redundancy coefficient, service quality benefit coefficient, and service protection correction coefficient of the enhanced mobile broadband service slice. Since this enhanced mobile broadband service slice does not belong to the low-latency, high-reliability, or continuous service assurance service slice categories, and its service quality safety margin is not less than 20%, the edge computing node determines that it does not meet the rigid service protection conditions. When the slice energy consumption debt value exceeds the debt repayment trigger threshold for two consecutive scheduling cycles, the edge computing node initiates debt repayment-based resource allocation.

[0051] When performing debt-repayment-based resource allocation, edge computing nodes determine the resource compression ratio for the current scheduling cycle based on the slice's energy consumption debt value and the security compression ratio. Priority is given to reducing the allocation ratio of non-critical physical resource blocks, followed by reducing redundant transmit power, and then reducing the scheduling frequency of delayable services. During execution, edge computing nodes continuously monitor the throughput and average latency of the enhanced mobile broadband service slice. Resource compression continues as long as throughput still meets service quality constraints and average latency has not reached the service quality constraint boundary. When any service quality indicator reaches the corresponding service quality constraint boundary, resource compression stops, and the reduced energy consumption is recorded as debt repayment.

[0052] For low-latency service slices, if their current slice energy consumption debt is high, but the average latency is close to the corresponding service quality constraint boundary, and the service quality safety margin is less than 20%, then the edge computing node determines that it meets the rigid service protection conditions. In this case, resource compression processing such as reducing critical physical resource blocks, reducing critical transmit power, or reducing critical MIMO layers is not performed on the low-latency service slice. Instead, its minimum service quality guarantee resources are maintained, and its slice energy consumption debt is recorded as the latency debt value. The debt repayment timing is reassessed after the service load decreases or the service quality safety margin recovers to no less than 20% in subsequent scheduling cycles.

[0053] Through the above implementation methods, the present invention can perform energy consumption debt repayment for compressible service slices based on the identification of resource-inefficient cells and resource-inefficient service slices, perform rigid service protection for low-latency, high-reliability or continuous service guarantee slices, and update the resource allocation strategy for subsequent scheduling cycles according to actual energy consumption changes and service quality changes, thereby reducing the energy consumption of 5G communication systems while maintaining stable service quality.

[0054] Example 3: In one specific implementation, the target 5G communication system is applied to a large-scale campus communication scenario. The large campus is equipped with two macro base stations, six micro base stations, and one edge computing node. Service slices include enhanced mobile broadband service slices, low-latency control service slices, and IoT data acquisition service slices. Scheduling cycle. Take 15 minutes of data, and convert the energy consumption data of each cell and each service slice into the data for the scheduling cycle. The values, physical resource block usage, transmit power usage, MIMO layer usage, and quality of service metrics are all processed according to the aforementioned normalization rules.

[0055] Within a certain scheduling period, the actual energy consumption value of the enhanced mobile broadband service slice obtained by the edge computing node is... The energy consumption baseline value obtained based on slice-level correction from historical samples of similar operating conditions is Based on the aforementioned energy efficiency deviation calculation rules, the slice energy efficiency deviation value for this enhanced mobile broadband service slice is obtained as follows: The edge computing node further obtains the energy efficiency deviation distribution of the service slice over the most recent 12 scheduling cycles, and determines the slice dynamic threshold based on the median and absolute deviation. Because the slice energy efficiency deviation of this enhanced mobile broadband service slice is greater than [value missing] in two consecutive scheduling cycles. Moreover, its throughput increase was only Service quality targets not met Based on the improved judgment criteria, this enhanced mobile broadband service slice was therefore identified as a resource-inefficient service slice.

[0056] Edge computing nodes calculate the resource redundancy coefficient of this enhanced mobile broadband service slice. After comparison, its physical resource block redundancy ratio is... The transmit power redundancy ratio is The redundancy ratio of multi-input multi-output layers is: The resource redundancy coefficient is determined according to the maximum normalized excess ratio. for Because the throughput increase of this service slice did not meet the service quality target. Service quality revenue coefficient Pick This service slice is a standard enhanced mobile broadband service slice, not a low-latency, high-reliability, or continuous service guarantee service slice. The service protection correction factor... Pick Therefore, the energy consumption debt value of this business segment is obtained. If the current system determines the debt repayment trigger threshold based on the debt value distribution of all business slices over the most recent 12 scheduling cycles, then... If so, the enhanced mobile broadband service slice meets the debt repayment conditions.

[0057] During the resource compression ratio determination phase, edge computing nodes determine the security compression ratio based on the service quality security margin of the enhanced mobile broadband service slice. It is determined that the average latency, packet loss rate, reliability, and throughput of this service slice all have security margins relative to their respective service quality constraint boundaries that are not less than [a certain value]. Therefore, the safe compression ratio Pick The maximum resource compression ratio allowed in a single scheduling cycle. Pick Debt conversion coefficient Pick The resource compression ratio within the current scheduling period is then taken as... Based on this resource compression ratio, edge computing nodes prioritize reducing the allocation ratio of non-critical physical resource blocks, then reduce redundant transmit power, while keeping the resources for carrying critical services unchanged.

[0058] After resource compression, the actual energy consumption of this enhanced mobile broadband service slice in the next scheduling cycle is reduced by [previous value]. Descending to The average latency is from Change to It is still below the service quality constraints of this business slice. Upper limit; packet loss rate from Change to Still lower than The upper limit of the constraint; the throughput meets the minimum throughput requirement of this service slice. Therefore, the edge computing node records the resource adjustment parameters of reducing the allocation ratio of non-critical physical resource blocks and reducing redundant transmit power as an effective debt repayment strategy, and records the energy consumption reduction amount. Record the debt repayment amount for this business segment.

[0059] For low-latency control service slices within the same scheduling cycle, the edge computing node obtains its slice energy consumption debt value. Above the debt repayment trigger threshold However, the average latency of this low-latency control service slice is... Its service quality constraint upper limit is The service quality safety margin is Smaller than the applicable low-latency service slice Safety margin threshold. Therefore, the edge computing node determines that the low-latency control service slice meets the rigid service protection conditions, and does not perform resource compression processing such as reducing critical physical resource blocks, reducing critical transmit power, or reducing critical MIMO layers. Instead, it maintains its minimum quality of service guarantee resources and records its slice energy consumption debt value as latency debt value.

[0060] In this embodiment, the enhanced mobile broadband service slice is subject to debt-repayment resource allocation due to its persistent energy efficiency deviation and lack of matching service quality benefits. While the low-latency control service slice has a high energy consumption debt, it is included in rigid service protection due to insufficient service quality safety margin. Therefore, this invention can implement differentiated processing for different service slices within the same scheduling cycle, avoiding inappropriate compression of low-latency, high-reliability services, while simultaneously enabling controllable energy-saving adjustments for service slices with redundant resource usage.

[0061] The above embodiments provide a detailed description of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for energy efficiency optimization and resource allocation in a 5G communication system, characterized in that, The method includes: The system acquires operational status data of multiple cells and multiple service slices in the target 5G communication system during the current scheduling period. The operational status data includes cell resource occupancy data, cell channel quality data, cell actual energy consumption data, slice resource occupancy data, slice service quality data, and slice actual energy consumption data. A cell-level corrected energy consumption benchmark value is generated based on the cell resource occupancy data, cell channel quality data, and cell service quality constraints. A slice-level corrected energy consumption benchmark value is generated based on the slice resource occupancy data, slice service type, and slice service quality constraints. The deviation between the actual energy consumption data of the community and the community-level corrected energy consumption benchmark value is calculated, the deviation between the actual energy consumption data of the slice and the slice-level corrected energy consumption benchmark value is calculated, and the corrected energy efficiency deviation fingerprint is generated by combining the direction of resource occupation change and the direction of service quality change. Based on the aforementioned energy efficiency deviation correction fingerprint identification, inefficient cells and inefficient service slices are identified. Calculate the energy consumption debt value of inefficient business segments; The priority of resource allocation adjustment is determined based on the energy efficiency deviation fingerprint, sliced ​​energy consumption debt value, business priority, service quality safety margin, and historical debt processing records. According to the priority of resource allocation adjustment, debt repayment-style resource allocation is performed on business slices that meet the debt repayment conditions, and service quality assurance resources are maintained for business slices that meet the rigid business protection conditions. Collect energy consumption change data and service quality change data after debt repayment-based resource allocation, and update the corrected energy efficiency deviation fingerprint, slice energy consumption debt value and resource allocation strategy for the next scheduling cycle.

2. The method for energy efficiency optimization and resource allocation of a 5G communication system according to claim 1, characterized in that: The operational status data is jointly acquired by the base station management unit, the slice management unit, and the energy consumption metering unit. The cell resource occupancy data includes physical resource block occupancy rate, carrier bandwidth occupancy, transmit power occupancy, and multiple input multiple output layer occupancy. The cell channel quality data includes channel quality indication, reference signal received power, reference signal received quality, and uplink interference intensity; The slice service quality data includes slice average latency, slice packet loss rate, slice reliability, slice throughput, and slice user connections; The system then performs sampling period alignment, abnormal sampling point removal, and missing sampling point completion on the operational status data to obtain standardized operational status data.

3. The method for energy efficiency optimization and resource allocation of a 5G communication system according to claim 2, characterized in that, The generation of the cell-level corrected energy consumption baseline value includes: The cell state vector is constructed based on the cell service load, cell channel quality, cell coverage constraints, and cell service quality constraints. Filter historical cell samples that match the current cell status vector from historical operational status data; Remove samples from historical community samples that do not meet community service quality constraints; The remaining historical cell samples are sorted from low to high according to their energy consumption values, and the energy consumption values ​​that meet the preset percentile positions are determined as the cell-level correction energy consumption benchmark values.

4. The method for energy efficiency optimization and resource allocation of a 5G communication system according to claim 1 or 2, characterized in that, The generation of the slice-level corrected energy consumption baseline includes: Construct a slice state vector based on slice service type, slice physical resource block usage, slice transmit power usage, slice multi-input multi-output layer usage, and slice quality of service constraints; Filter historical slice samples from historical slice scheduling data that match the current slice state vector; Remove samples from historical slice samples that do not meet the corresponding slice service quality constraints; The lowest energy consumption quantile in the remaining historical slice samples is determined as the slice-level corrected energy consumption benchmark.

5. The method for energy efficiency optimization and resource allocation of a 5G communication system according to claim 4, characterized in that: The energy efficiency deviation fingerprint includes a deviation intensity field, a deviation duration field, a deviation source field, a resource usage field, and a service quality response field. The deviation intensity field is used to record the normalized deviation of the actual energy consumption from the corresponding corrected energy consumption benchmark value. The deviation persistence field is used to record the number of scheduling cycles in which the energy efficiency deviation continuously exceeds the dynamic threshold. The deviation source field is used to record cell-side energy consumption deviation, slice-side energy consumption deviation, or cell-slice coupling energy consumption deviation. The resource occupancy field is used to record changes in the occupancy of physical resource blocks, transmit power, and multiple input multiple output layers; The service quality response field is used to record whether service quality has improved, remained stable, or declined.

6. The method for energy efficiency optimization and resource allocation of a 5G communication system according to claim 5, characterized in that, The identification of resource-inefficient cells and resource-inefficient service slices includes: When the energy efficiency deviation of the target cell exceeds the dynamic threshold of the cell in multiple consecutive scheduling cycles, and the throughput, coverage quality and edge user experience of the target cell do not reach the preset improvement range, the target cell will be identified as a resource-inefficient cell. When the slice energy efficiency deviation value of the target service slice exceeds the slice dynamic threshold in multiple consecutive scheduling cycles, and the average latency, packet loss rate, reliability and throughput of the target service slice do not reach the preset improvement range, the target service slice will be identified as a resource-inefficient service slice. The cell dynamic threshold and slice dynamic threshold are determined based on the median and absolute deviation of the corresponding energy efficiency deviation values ​​within the sliding time window.

7. The method for energy efficiency optimization and resource allocation of a 5G communication system according to claim 6, characterized in that: The slice energy consumption debt value is determined based on the slice energy consumption excess, excess resource usage, service quality revenue value, business priority correction coefficient, and business rigidity correction coefficient. Wherein, the excess energy consumption of the slice is the difference between the actual energy consumption data of the slice and the corrected energy consumption benchmark value at the slice level; The excess resource usage includes the physical resource block usage, transmit power usage, and multiple input multiple output layer usage that exceed the slice service quality constraints. The service quality benefit value is determined based on the reduction in latency, the reduction in packet loss rate, the improvement in reliability, and the improvement in throughput. The business priority correction coefficient and business rigidity correction coefficient are used to reduce the debt repayment intensity of high-priority business slices and rigid business slices in the current scheduling cycle.

8. The method for energy efficiency optimization and resource allocation of a 5G communication system according to claim 7, characterized in that: The priority of resource allocation adjustment is determined jointly by energy efficiency anomaly level, debt status level, and business protection level; The energy efficiency anomaly level is determined by the deviation intensity field and the deviation duration field in the corrected energy efficiency deviation fingerprint. The debt status level is determined by the slice energy consumption debt value, the number of historical debt repayments, and the number of historical debt delays. The business protection level is determined by business priority, service quality security margin, and risk of user experience degradation. The business slices are sorted in order of energy efficiency anomaly level from high to low, debt status level from high to low, and business protection level from low to high to obtain the priority of resource allocation adjustment.

9. The method for energy efficiency optimization and resource allocation of a 5G communication system according to claim 8, characterized in that, The debt-repayment-based resource allocation includes: When the energy consumption debt value of an inefficient business slice exceeds the debt repayment trigger threshold, and the inefficient business slice does not meet the rigid business protection conditions, resource compression processing is initiated. The resource compression process includes one or more of the following: reducing the allocation ratio of non-critical physical resource blocks, reducing redundant transmit power, reducing unnecessary multiple input multiple output layers, reducing the scheduling frequency of deferred services, and migrating deferred services. During the resource compression process, the average latency, packet loss rate, reliability, and throughput of the corresponding business slice are continuously monitored. When any service quality indicator reaches the corresponding service quality constraint boundary, the resource compression process for that business slice will be stopped, and the energy consumption reduction generated by the resource compression process will be recorded as the debt repayment amount.

10. The method for energy efficiency optimization and resource allocation of a 5G communication system according to claim 9, characterized in that, The updates to the rigid business protection conditions and resource allocation strategies include: When a service slice has low latency constraints, high reliability constraints, or continuous service guarantee constraints, and the service quality security margin of the service slice is less than the security margin threshold, the service slice is determined to meet the rigid service protection conditions. For business slices that meet the rigid business protection conditions, maintain the minimum quality of service guarantee resources and record their slice energy consumption debt value as the delayed debt value; After the current scheduling cycle ends, collect data on actual energy consumption changes, resource usage changes, and service quality changes. When resource compression reduces actual energy consumption and service quality indicators meet the corresponding service quality constraints, the corresponding resource adjustment parameters are recorded as an effective debt repayment strategy. When resource compression processing causes any service quality indicator to fail to meet the corresponding service quality constraint, the corresponding resource adjustment parameter will be recorded as an invalid debt repayment strategy. Based on the effective debt repayment strategy, ineffective debt repayment strategy, debt repayment amount, and delayed debt value, update the resource allocation strategy for the next scheduling cycle.

Images