Wireless network scheduling method driven by edge computing

Abstract

The invention relates to the technical field of network scheduling, in particular to a wireless network scheduling method driven by edge computing, which comprises the following steps of: acquiring task execution time delay of a current wireless network computing task, monitoring computing power occupation condition of the computing task at an edge computing node, and executing the task execution time delay; the execution interval time of the calculation task is analyzed, the calculation power occupation fluctuation trend is analyzed, the calculation power occupation ratio of the current wireless network is calculated, and calculation power utilization ratio data are output. According to the method, the execution time delay of the wireless network computing task is accurately monitored, analysis is performed in combination with the computing power occupancy fluctuation trend, different task computing requirements are met, the data rate ratio, the channel bandwidth occupancy rate and the data throughput ratio are comprehensively computed, network congestion is reduced, and the computing task execution node is dynamically adjusted, so that the computing efficiency is improved. And intelligent matching is performed on the channel adaptation capability and the task calculation requirement by combining the optimal allocation path of the calculation resources and combining the edge cache, so that the overall throughput capability and the task execution stability of the wireless network are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of network scheduling, and in particular, to a wireless network scheduling method driven by edge computing. Background Art

[0002] The technical field of network scheduling includes resource management and scheduling mechanisms in wireless networks, involving the reasonable allocation of communication resources such as wireless channels, time, and power. The core content is to ensure the stability and efficiency of wireless networks, mainly involving access control, interference management, power control, and data transmission optimization. The development of network scheduling technology has accompanied the evolution of wireless communication protocols, gradually developing from early static scheduling methods to intelligent scheduling strategies based on dynamic resource allocation. Currently, with the introduction of edge computing, artificial intelligence, and big data analysis, network scheduling technology has been continuously improved in high-density networks, ultra-low latency communication, and energy efficiency optimization, achieving fine-grained control and adaptive optimization of wireless network resources.

[0003] Among them, the wireless network scheduling method driven by edge computing refers to a technical solution that uses edge computing capabilities to optimize and adjust resource allocation and scheduling strategies in a wireless communication network. It mainly relies on the local computing capabilities of edge computing nodes to perform real-time analysis on aspects such as network environment perception, data processing, and task distribution, and combines historical data to establish scheduling strategies. Specifically, this method involves links such as wireless channel state monitoring, task load prediction, and computing resource allocation, and uses a prediction model based on machine learning to dynamically adjust scheduling parameters, thereby achieving optimization of user access, data transmission priority, and power allocation strategies in a wireless network. By offloading computing tasks to edge devices, the load pressure on the core network is reduced, and decisions are made locally, making network scheduling more intelligent and efficient.

[0004] In the existing wireless network scheduling process, there are problems such as low resource utilization rate, insufficient channel load balancing ability, unreasonable adjustment of task priorities, and lag in the response of computing resource scheduling. The allocation of computing resources is usually based on static policies, and the fluctuation trend of computing power occupancy cannot be effectively analyzed, resulting in a lack of precise matching between the computing requirements of tasks and resource allocation. Some computing tasks may experience execution delays due to unreasonable computing power allocation. Channel load monitoring mostly uses fixed bandwidth occupancy rate and packet loss rate analysis, and fails to perform dynamic optimization by combining data throughput ratios, making it difficult to improve the utilization efficiency of channel resources. It is vulnerable to burst traffic, resulting in overloading of some channels and affecting the overall network throughput capacity. The task scheduling strategy lacks precise matching of delay requirements, and the adjustment of task priorities fails to be optimized by combining channel load conditions, resulting in low-delay tasks being preempted by non-critical tasks, affecting the execution efficiency of high-timeliness tasks. The adjustment of the execution nodes of computing tasks lacks dynamic adaptability and fails to perform scheduling by combining the global distribution of computing resources, making some task migration plans unable to fully utilize computing resources and affecting the execution efficiency of computing tasks and the overall performance of network scheduling. Summary of the Invention

[0005] The object of the present invention is to solve the deficiencies existing in the prior art and propose a wireless network scheduling method driven by edge computing.

[0006] To achieve the above object, the present invention adopts the following technical solutions: A wireless network scheduling method driven by edge computing, comprising the following steps:

[0007] S1: Obtain the task execution delay of the current wireless network computing task, monitor the computing power occupancy of the computing task at the edge computing node, analyze the execution interval time of the computing task and analyze the fluctuation trend of computing power occupancy, calculate the computing power occupancy ratio of the current wireless network, and output the computing power utilization ratio data;

[0008] S2: Read the data transmission rate, channel bandwidth occupancy rate, and packet loss rate of the wireless network channel, calculate the ratio of the current data rate to the theoretical value of the channel bandwidth, analyze the change trend of the current packet loss rate, and calculate the ratio of the actual data transmission volume per unit time to the maximum available data transmission volume of the channel to obtain the channel load occupancy data;

[0009] S3: Combine the computing power utilization ratio data and the channel load occupancy data, screen the currently available channels, calculate the data throughput capacity ratio between the current channel and adjacent channels, select the channel with excellent throughput capacity as the target channel, and comprehensively output the target channel adaptation result;

[0010] S4: Based on the latency requirements of the computing tasks, combined with the channel load occupancy rate, allocate the basic task priorities, analyze the current channel occupancy rate, judge the changing trend of the wireless task queue load, adjust the task priority order, and obtain the task priority sorting list.

[0011] As a further solution of the present invention, the computing power utilization ratio data includes the proportion of task execution latency, the computing power occupancy fluctuation coefficient, and the record of the distribution of computing task execution intervals. The channel load occupancy data includes the proportion of data rate, the channel bandwidth occupancy rate, and the data throughput ratio per unit time. The target channel adaptation result includes the target channel throughput capacity, the amount of data loss during channel switching, and the edge cache enabling status. The task priority sorting list includes the basic task priority, the sorting adjustment result of low-latency tasks, and the analysis result of the changing trend of the wireless task queue load.

[0012] As a further solution of the present invention, the specific steps of S1 are as follows:

[0013] S111: Obtain the task execution latency of the current wireless network computing task, detect the computing power occupancy of this computing task on the edge computing node, extract its time series data, calculate the execution interval time of the computing task, and extract the computing power occupancy data for each time period to obtain the computing task execution interval time series;

[0014] S112: Based on the computing task execution interval time series, conduct a statistical analysis of the computing power occupancy at each time point, calculate the mean square error and the fluctuation amplitude, and use the formula:

[0015]

[0016] Calculate to obtain the computing power occupancy fluctuation trend coefficient T v , where C i represents the computing power occupancy ratio at the i-th moment, n represents the length of the time series, represents the average value of the computing power occupancy ratio, max(C) represents the maximum value of the computing power occupancy ratio in the time series, and min(C) represents the minimum value of the computing power occupancy ratio in the time series;

[0017] S113: According to the computing power occupancy fluctuation trend coefficient, calculate the computing power occupancy ratio of the current wireless network, normalize the computing power utilization rate, and extract the computing power utilization ratio data.

[0018] As a further solution of the present invention, the specific steps of S2 are as follows:

[0019] S211: Obtain the data transmission rate, channel bandwidth occupancy rate, and packet loss rate of the wireless network channel, calculate the ratio of the current data rate to the theoretical value of the channel bandwidth, analyze the channel utilization situation, and obtain the channel utilization analysis result;

[0020] S212: Analyze the changing trend of the current packet loss rate based on the analysis result of the channel utilization rate, calculate the ratio of the actual data transmission volume within a unit time to the maximum available data transmission volume of the channel, using the formula:

[0021]

[0022] Calculate and output the channel load occupancy ratio L, where D i represents the actual data transmission volume in the i-th time interval, Z i represents the maximum available data transmission volume of the channel in the i-th time interval, P i represents the packet loss rate in the i-th time interval, and H represents the total number of time intervals;

[0023] S213: Based on the channel load occupancy ratio, combined with the channel utilization rate ratio and the changing trend of the packet loss rate, judge the channel interference situation and generate channel load occupancy data.

[0024] As a further solution of the present invention, the specific steps of S3 are:

[0025] S311: Based on the computing power utilization ratio data and the channel load occupancy data, judge whether the load of the current wireless network access point exceeds the set load threshold, and screen the currently available channels to obtain a set of available channels;

[0026] S312: Based on the set of available channels, use the formula:

[0027]

[0028] Calculate the data throughput capacity ratio T between the current channel and the adjacent channels, and select the channel with excellent throughput capacity according to the ratio to obtain a set of target channels, where B i represents the bandwidth of the i-th channel, U i represents the utilization rate of the i-th channel, K i represents the packet loss rate of the i-th channel, D i represents the data traffic change volume of the i-th channel, and M represents the total number of channels;

[0029] S313: Based on the set of target channels, calculate the data loss volume generated during the channel switching process, and judge whether it is necessary to enable the edge cache unit to store the unfinished data packets, and comprehensively output the target channel adaptation result.

[0030] As a further solution of the present invention, the specific steps of S4 are:

[0031] S411: Statistically calculate the latency requirements of computing tasks and the channel load occupancy rate, calculate the latency requirement intensity of each computing task, calculate the channel load occupancy rate, calculate the change rate of the channel load occupancy rate, establish the ratio of the task latency requirement intensity to the change rate of the channel load occupancy rate, and obtain task latency pressure ratio data;

[0032] S412: Based on the task latency pressure ratio data and combined with the current channel occupancy rate, judge the load change trend of the wireless task queue, and use the formula:

[0033]

[0034] Calculate the task load adjustment coefficient P adj , and adjust and obtain the basic priority of the computing task in combination with the channel occupancy rate. Among them, T d represents the task latency requirement intensity, R ch represents the channel load occupancy rate, L var,i represents the i-th variable of the channel load change rate, m represents the total number of variables, represents the comprehensive influence amount of the channel load change;

[0035] S413: Based on the basic priority of the computing task, analyze the current channel occupancy rate, judge the load change trend of the wireless task queue, adjust the priority order of the low-latency tasks in the task queue, and obtain the task priority sorting list.

[0036] As a further solution of the present invention, the method further includes;

[0037] S5: Based on the target channel adaptation result and the task priority sorting list, analyze the matching degree between the current computing task and the available channels, judge whether the task computing requirements exceed the upper limit of the current channel adaptation ability. If it exceeds, adjust the execution node of the task, and use the computing resource allocation node as the task migration target to obtain the wireless network computing task allocation plan.

[0038] The wireless network computing task allocation plan includes the task execution node, the computing resource allocation target, and the task migration matching degree.

[0039] As a further solution of the present invention, the specific steps of S5 are:

[0040] S511: Based on the target channel adaptation result and the task priority sorting list, statistically calculate the total amount of computing resources required for the computing task, calculate the upper limit of the current channel adaptation ability, analyze the ratio of the total amount of resources required for the computing task to the upper limit of the channel adaptation ability, and obtain the task channel adaptation degree ratio data;

[0041] S512: Based on the task channel adaptability ratio data, determine whether the task computing requirement exceeds the upper limit of the current channel adaptability. If it exceeds, adjust the execution node of the task using the formula:

[0042]

[0043] Calculate the resource matching degree M after adjusting the execution node of the computing task adj , and in combination with the task priority ranking list, adjust the computing resource allocation node to obtain the task migration target, where R t represents the total amount of computing resources required for the task, R ch represents the upper limit of the current channel adaptability, C node represents the current resource capacity of the computing node, S var,j represents the j-th variable of the change in computing resources during the task migration, T d represents the task delay requirement intensity, L var represents the task channel load change rate, and m represents the total number of variables;

[0044] S513: According to the task migration target, in combination with the computing resource allocation node, dynamically adjust the allocation of the task computing resources to obtain the wireless network computing task allocation scheme.

[0045] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0046] In the present invention, through the precise monitoring of the execution delay of the wireless network computing task and the analysis in combination with the fluctuation trend of the computing power occupation, the computing resource scheduling can adapt to the computing requirements of different tasks, reduce resource waste, and improve the computing efficiency. In combination with the channel load occupation situation, based on the comprehensive calculation of the data rate ratio, the channel bandwidth occupancy rate, and the data throughput ratio per unit time, the utilization of the channel resources is made more reasonable, and the network congestion caused by the unbalanced channel load is reduced. Based on the delay requirements of the computing task, the priority order of the tasks is optimized, so that the low-delay tasks can obtain better execution opportunities and the processing efficiency of the high-priority tasks is improved. Through the dynamic adjustment of the computing task execution node and in combination with the optimal allocation path of the computing resources, the computing resource allocation of the wireless network task is made more efficient, and the impact of the uneven distribution of computing resources on the network performance is reduced. By comprehensively analyzing the channel throughput capacity and the data loss situation, the data integrity during the channel switching is ensured, and in combination with the edge caching technology, the continuity of the data transmission is improved, and the data loss caused by the channel switching is reduced. The channel adaptability and the task computing requirements are intelligently matched, so that the scheduling of the computing tasks is more targeted, and the overall throughput capacity and the task execution stability of the wireless network are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1This is the main process flow chart of the present invention;

[0048] Figure 2 This is the process flow chart of step S1 of the present invention;

[0049] Figure 3 This is the process flow chart of step S2 of the present invention;

[0050] Figure 4 This is the process flow chart of step S3 of the present invention;

[0051] Figure 5 This is the process flow chart of step S4 of the present invention;

[0052] Figure 6 This is the process flow chart of step S5 of the present invention. Detailed implementation manners

[0053] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0054] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation of the present invention. In addition, in the description of the present invention, "a plurality of" means two or more unless otherwise specifically defined.

[0055] Please refer to Figure 1 , a wireless network scheduling method driven by edge computing, including the following steps:

[0056] S1: Obtain the task execution delay of the current wireless network computing task, monitor the computing power occupancy of the computing task on the edge computing node, analyze the execution interval time of the computing task and the fluctuation trend of the computing power occupancy, calculate the current wireless network computing power occupancy ratio, and output the computing power utilization ratio data;

[0057] S2: Read the data transmission rate, channel bandwidth occupancy rate and packet loss rate of the wireless network channel, calculate the ratio of the current data rate to the theoretical value of the channel bandwidth, analyze the current packet loss rate change trend, and judge the channel interference situation, calculate the ratio of the actual transmitted data volume per unit time to the maximum available data transmission volume of the channel, and obtain the channel load occupancy data;

[0058] S3: Combine the computing power utilization ratio data and the channel load occupancy data, determine whether the load of the current wireless network access point exceeds the set load threshold, filter the currently available channels, calculate the data throughput capacity ratio between the current channel and adjacent channels, select the channel with better throughput capacity as the target channel, calculate the data loss amount generated during the channel switching process, and determine whether it is necessary to enable the edge cache unit to store the unfinished data packets, and comprehensively output the target channel adaptation result;

[0059] S4: Based on the latency requirements of the computing tasks, combine the channel load occupancy rate, assign a basic priority to the tasks, analyze the current channel occupancy rate, judge the load change trend of the wireless task queue, and adjust the priority order of the low-latency tasks in the task queue to obtain the task priority sorting list;

[0060] S5: According to the target channel adaptation result and the task priority sorting list, analyze the matching degree between the current computing tasks and the available channels, judge whether the task computing requirements exceed the upper limit of the current channel adaptation ability, if so, adjust the execution node of the task, and use the computing resource allocation node as the task migration target to obtain the wireless network computing task allocation plan.

[0061] The computing power utilization ratio data includes the task execution latency ratio, the computing power occupancy fluctuation coefficient, and the record of the computing task execution interval distribution. The channel load occupancy data includes the data rate ratio, the channel bandwidth occupancy rate, and the data throughput ratio per unit time. The target channel adaptation result includes the target channel throughput capacity, the channel switching data loss amount, and the edge cache enable status. The task priority sorting list includes the task basic priority, the sorting adjustment result of low-latency tasks, and the analysis result of the load change trend of the wireless task queue. The wireless network computing task allocation plan includes the task execution node, the computing resource allocation target, and the task migration matching degree.

[0062] Please refer to Figure 2 , the specific steps of S1 are as follows:

[0063] S111: Obtain the task execution latency of the current wireless network computing task, detect the computing power occupancy of the computing task on the edge computing node, extract its time series data, calculate the execution interval time of the task, and extract the computing power occupancy data for each time period to obtain the computing task execution interval time series;

[0064] To obtain the task execution latency of the current wireless network computing task, it needs to be completed through the actual monitoring of the wireless network environment. Specifically, the start time and end time of task execution can be recorded by setting monitoring points at fixed time intervals. For example, in a wireless local area network (WLAN) environment, the execution time of each task can be calculated through the timestamp of the probe packet, so as to obtain the task execution latency. Assume that the start time of a certain task is t s= 12.005 s, the end time is t e = 12.035 s, then the task execution delay is:

[0065] T d = t e - t s = 12.035 s - 12.005 s = 0.030 s;

[0066] To detect the computing power occupancy of the computing task on the edge computing node, usually use the resource monitoring tools of the computing node (such as CPU utilization, memory occupancy, etc.) for collection. For edge computing devices such as Raspberry Pi and embedded computing units, the CPU utilization can be obtained through system resource monitoring instructions. For example, the CPU occupancy rate of a certain node is C u = 65%, combined with the task execution time, the computing resources consumed by the task during execution can be obtained. Assuming that the total computing volume of the computing task is C t = 500 million instructions (MIPS), then the computing power requirement during the execution of the computing task is:

[0067]

[0068] To extract time series data, it is necessary to continuously monitor the execution information of multiple tasks to construct a time series, as shown in Table 1.1 below.

[0069] Table 1.1 Execution interval time series of computing tasks

[0070] Task ID Start Time (s) End Time (s) Task Execution Delay (s) CPU Occupancy (%) Computing Task Required Computing Power (MIPS) 1 10.000 10.030 0.030 60 300 2 10.050 10.085 0.035 70 350 3 10.100 10.132 0.032 65 325 4 10.150 10.181 0.031 68 340

[0071] As shown in Table 1.1, the execution delays and computing power occupancies of multiple computing tasks are obtained, and the execution intervals of the tasks are calculated based on the timestamp information of the task execution. For example, the execution interval between Task 1 and Task 2 is:

[0072] T gap = 10.050 s - 10.030 s = 0.020 s;

[0073] And so on, the execution interval data between all tasks can be obtained, and the execution interval time series of the computing tasks can be constructed.

[0074] S112: Based on the execution interval time series of the computing tasks, statistically analyze the computing power occupancy at each time point, calculate the mean square error and the fluctuation amplitude, using the formula:

[0075]

[0076] Calculate the computing power occupancy fluctuation trend coefficient T v , where, C irepresents the computing power utilization ratio at the i-th moment, n represents the length of the time series, Represents the average value of the computing power utilization ratio, max(C) represents the maximum value of the computing power utilization ratio in the time series, and min(C) represents the minimum value of the computing power utilization ratio in the time series;

[0077] Based on the time series of computing task execution intervals, the fluctuation trend of computing power usage is calculated. First, the mean square error (MSE) of the computing power usage ratio is calculated to measure the stability of computing power usage. The specific calculation method is as follows:

[0078]

[0079] Among them, C i is the computing power usage ratio at the i-th moment, is the average computing power usage ratio (calculated using the CPU usage ratio, because the CPU usage ratio is a common indicator for measuring computing power usage), n is the total number of computing tasks, taking the data in Table 1.1 as an example, first calculate the average computing power usage ratio:

[0080]

[0081] Then calculate the mean square error:

[0082]

[0083] In addition, the fluctuation range of computing power usage is calculated, that is, the ratio of the maximum value to the minimum value:

[0084]

[0085] Finally, calculate the fluctuation trend coefficient of computing power usage:

[0086] T v =MSE+ΔC=14.19+0.152=14.342;

[0087] The computing power occupies the fluctuation trend coefficient T v =14.342 represents the degree of change in the computing power usage of the current computing task at different time points. The larger the MSE, the more drastic the fluctuation in computing power usage, and the more unstable the computing resource usage of the task, while the fluctuation amplitude ΔC reflects the maximum change in the computing power usage of the task. From the calculation results, the MSE is much larger than the fluctuation amplitude value, indicating that the instability of computing power usage mainly comes from the deviation of computing power usage at different time points, rather than the extreme value change of overall computing power usage. That is, the computing power usage of the current computing task has a strong fluctuation trend, and there is an imbalance in task scheduling. It is necessary to further analyze the task allocation strategy to optimize the use of computing resources.

[0088] S113: Calculate the computing power occupancy ratio of the current wireless network according to the computing power occupancy fluctuation trend coefficient, perform normalization processing on the computing power utilization rate, and extract the computing power utilization rate ratio data;

[0089] Call the computing power occupancy fluctuation trend coefficient, calculate the computing power occupancy ratio of the current wireless network, perform normalization processing on the computing power utilization rate, first calculate the normalization value of the current task execution interval, using the normalization calculation formula:

[0090]

[0091] Set the task execution delay range in the wireless network environment to be between T min = 0.020s and T max = 0.040s. Then for task 3 (delay 0.032s):

[0092]

[0093] Then calculate the normalization value of the computing power occupancy ratio:

[0094]

[0095] Set the CPU utilization rate to be between C min = 50% and C max = 80%. Then for task 3 (CPU occupancy 65%):

[0096]

[0097] Finally, calculate the normalized computing power utilization rate ratio:

[0098]

[0099] This computing power utilization rate ratio U norm = 0.55 reflects the computing resource usage of the current task. The purpose of normalization calculation is to map the computing power occupancy and task execution delay to the same numerical range (0 to 1) so that they can be fairly compared. The calculation results show that the normalized delay value of the current task is 0.6, while the computing power occupancy value is 0.5, and their average value is 0.55, indicating that the task execution time and computing power occupancy level are above the medium level.

[0100] Specifically, the latency of this task is close to 60% of the maximum acceptable range, indicating that the task execution time is relatively long, while the computing power occupancy is at the 50% level, indicating that the utilization efficiency of computing resources is relatively moderate. Overall, the calculation of the computing power utilization ratio shows that this task has not fully optimized the utilization efficiency of computing resources. If we want to improve the computing performance of the wireless network, we can consider reducing the task execution latency or optimizing the computing power allocation strategy to further increase the ratio and make it closer to 1, thereby enhancing the utilization rate of computing resources.

[0101] Please refer to Figure 3 , and the specific steps of S2 are as follows:

[0102] S211: Obtain the data transmission rate, channel bandwidth occupancy rate, and packet loss rate of the wireless network channel, calculate the ratio of the current data rate to the theoretical value of the channel bandwidth, analyze the channel utilization situation, and obtain the channel utilization analysis result;

[0103] To obtain the data transmission rate, channel bandwidth occupancy rate, and packet loss rate of the wireless network channel, first, use a wireless network monitoring device to continuously sample the target channel and record the amount of transmitted data within a unit of time. For example, if the data transmission amount within a certain second is 50MB, then the transmission rate for this second is calculated as:

[0104]

[0105] At the same time, measure the channel bandwidth occupancy rate, that is, the ratio of the actual used bandwidth to the total channel bandwidth. Assume that the maximum bandwidth of the channel is 100MHz and the currently detected used bandwidth is 40MHz, then the bandwidth occupancy rate is calculated as follows:

[0106]

[0107] In addition, count the number of lost packets within a unit of time. If a total of 1000 data packets are sent within 1 second and 10 packets are lost, then the packet loss rate is calculated as follows:

[0108]

[0109] Next, calculate the ratio of the current data rate to the theoretical maximum rate of the channel. The theoretical maximum rate of the channel can be calculated by the Shannon formula:

[0110] C = B × log2(1 + S / N);

[0111] where B is the channel bandwidth and S / N is the signal-to-noise ratio. Assume that the channel bandwidth is 100MHz and the signal-to-noise ratio is 20dB (i.e., S / N = 100), then the theoretical maximum rate is calculated as follows:

[0112] C = 100MHz × log2(101) ≈ 665Mbps;

[0113] Then, calculate the ratio of the actual data rate to the maximum theoretical channel rate. Assuming the current data rate is 400 Mbps, the ratio is calculated as follows:

[0114]

[0115] Analyze the channel utilization based on this ratio. Generally speaking, the closer this ratio is to 1, the higher the channel utilization; otherwise, the utilization is lower. In this example, the calculated result is 0.6, indicating that the channel utilization is 60%. Finally, obtain the analysis result of the channel utilization.

[0116] S212: According to the analysis result of the channel utilization, analyze the changing trend of the current packet loss rate, and calculate the ratio of the actual data volume transmitted per unit time to the maximum available data transmission volume of the channel, using the formula:

[0117]

[0118] Calculate and output the channel load occupancy ratio L, where D i represents the actual data volume transmitted in the i-th time interval, and Z i represents the maximum available data transmission volume of the channel in the i-th time interval, P i represents the packet loss rate in the i-th time interval, and H represents the total number of time intervals;

[0119] Combined with the analysis result of the channel utilization, analyze the changing trend of the current packet loss rate. First, count the packet loss rate data in the recent 10 seconds. For example:

[0120] {1%, 1.2%, 1.5%, 1.8%, 2%, 2.2%, 2.5%, 2.8%, 3%, 3.2%};

[0121] Draw a trend graph of the packet loss rate changing with time, and it can be observed that it gradually increases. Next, calculate the ratio of the actual data volume transmitted per unit time to the maximum available data transmission volume of the channel. Assume the actual transmission rate data in the past 10 seconds is as follows (unit: Mbps):

[0122] {400, 420, 430, 440, 450, 460, 470, 480, 490, 500};

[0123] The maximum available data transmission volume of the channel remains 665 Mbps, then the calculation is as follows:

[0124] ∑D i = 4450 Mbps;

[0125] ∑Z i = 6650 Mbps;

[0126] The total packet loss rate is calculated as follows:

[0127] ∑P i = 20.2%;

[0128] H = 10;

[0129] Substitute into the formula for calculation:

[0130]

[0131] Finally, the channel load occupancy ratio is obtained.

[0132] S213: Based on the channel load occupancy ratio, combined with the channel utilization ratio and the trend of packet loss rate change, determine the channel interference situation and generate channel load occupancy data;

[0133] According to the channel load occupancy ratio, combined with the analysis result of channel utilization rate and the trend of packet loss rate change, determine the channel interference situation. First, compare the channel utilization ratio and the channel load occupancy ratio. According to the above calculation, the channel utilization ratio is 0.6 and the channel load occupancy ratio is 0.654. The two values are similar, indicating that the utilization and load conditions of the channel are consistent. However, considering the upward trend of the packet loss rate, if the packet loss rate continues to increase, it may indicate that there is interference in the channel. For example, due to wireless interference from other devices, packets are lost during transmission. Further calculate the change rate of the channel packet loss rate to identify the severity of the interference. Assume that the increment of the packet loss rate change trend in the past 10 seconds is as follows (unit: %):

[0134] {+0.2, +0.3, +0.3, +0.3, +0.2, +0.2, +0.3, +0.3, +0.2, +0.2};

[0135] Calculate the average packet loss rate growth rate:

[0136]

[0137] For the threshold of the packet loss rate growth rate of 0.2% / s, its setting basis comes from the stability requirements of the wireless channel. In a conventional communication environment, the normal fluctuation range of the packet loss rate is less affected by network load and short-term interference, usually between 0.1% / s and 0.2% / s. If the growth rate of the packet loss rate exceeds 0.2% / s, it indicates that the channel may be affected by strong external interference, such as high-power devices in adjacent channels, short-term burst interference sources, or physical occlusion. Therefore, 0.2% / s as the interference recognition threshold for the packet loss rate growth rate is obtained based on long-term network monitoring data statistics and verified in combination with the actual wireless network environment. This threshold may be slightly adjusted in different environments and usually varies with the fluctuations of parameters such as the channel signal-to-noise ratio S / N, bandwidth B, and average packet loss rate P. When S / N is low or P is high, the threshold may be reduced to 0.15% / s, while in a stable environment, this value can be increased to 0.25% / s. Generally, however, 0.2% / s is used as the general judgment standard. If the growth rate of the packet loss rate is higher than 0.2% / s, it can be determined that there may be interference in the channel. In this example, the calculated result is 0.25% / s, which is higher than the threshold. Therefore, it can be judged that the channel is interfered, and finally the channel load occupancy data is generated.

[0138] Please refer to Figure 4 , and the specific steps of S3 are as follows:

[0139] S311: According to the computing power utilization ratio data and the channel load occupancy data, judge whether the load of the current wireless network access point exceeds the set load threshold, and screen the currently available channels to obtain the available channel set;

[0140] In practical applications, the load monitoring of a wireless network access point is usually evaluated by calculating the ratio of computing power utilization to channel load occupancy. Suppose the computing power utilization ratio of an access point is 0.78 and the channel load occupancy ratio is 0.65, then calculate its total load ratio L t It can be calculated by linear weighting:

[0141] L t = w1 × U c + w2 × L s ;

[0142] Among them, w1 and w2 are the weights of computing power utilization and channel load respectively. If it is set that w1 = 0.6 and w2 = 0.4, then the calculation is as follows:

[0143] L t = 0.6 × 0.78 + 0.4 × 0.65 = 0.468 + 0.26 = 0.728;

[0144] The setting basis of the load threshold of 0.75 comes from the balance between the throughput capacity of the wireless network access point and the network stability requirements. The calculation of this threshold involves the estimation of the maximum allowable load occupancy ratio L max where L max is jointly determined by the channel bandwidth, the number of access users, the average throughput requirement, and the upper limit of the packet loss rate. Assume that the maximum bandwidth B max of this access point is 200 Mbps, the current maximum number of concurrent users U max is 50, the average throughput requirement D avg of each user is 3 Mbps, and the upper limit of the packet loss rate P limit is set to 5%. Then the calculation of the maximum allowable load ratio is as follows:

[0145]

[0146] To ensure that there is an appropriate margin for the load threshold, the set threshold L threshold is

[0147] L thresHold = L max + 0.05 = 0.7125 + 0.05 = 0.7625 ≈ 0.75;

[0148] This threshold will fluctuate with the changes in the number of access users, the average throughput requirement, or the packet loss rate. Therefore, it may need to be adjusted in a timely manner in a dynamic environment.

[0149] Assume that the set load threshold is 0.75. Then the current load occupancy ratio L t is less than the threshold, indicating that the load has not exceeded the standard, and the available channels can continue to be screened.

[0150] When screening channels, channels with too high channel interference degree need to be excluded, and the current occupancy rate of each channel needs to be calculated. Assume there are 6 channels, and their current occupancy rates are shown in the following table:

[0151] Table 3.1 Channel Occupancy Rate Table

[0152]

[0153]

[0154] As shown in Table 3.1, the current channels 2, 4, and 5 meet the screening conditions. Therefore, the available channel set {channel 2, channel 4, channel 5} is obtained.

[0155] S312: Based on the available channel set, use the formula:

[0156]

[0157] Calculate the data throughput capacity ratio T between the current channel and adjacent channels, select the channel with better throughput capacity according to the ratio, and obtain the target channel set, where B i represents the bandwidth of the i-th channel, U i represents the utilization rate of the i-th channel, K i represents the packet loss rate of the i-th channel, D i represents the data traffic change amount of the i-th channel, and M represents the total number of channels;

[0158] To calculate the throughput capacity ratio, it is necessary to obtain the bandwidth, utilization rate, packet loss rate, and data traffic change amount of each channel. Assume that the parameters corresponding to the available channel set {Channel 2, Channel 4, Channel 5} are as follows:

[0159] Table 3.2 Channel Parameter Table

[0160] Channel ID Bandwidth (MHz) Utilization Rate (%) Packet Loss Rate (%) Data Traffic Change (Mbps) 2 40 65 8 3.2 4 80 42 5 4.5 5 60 30 6 2.8

[0161] According to the throughput capacity calculation formula, substitute the parameters in Table 3.2:

[0162]

[0163] Calculate the numerator:

[0164] 40 × 0.65 = 26;

[0165] 80 × 0.42 = 33.6;

[0166] 60 × 0.30 = 18;

[0167] 26 + 33.6 + 18 = 77.6;

[0168] Calculate the denominator:

[0169] 8 + 5 + 6 = 19;

[0170]

[0171] Since the higher the throughput capacity ratio, the better the channel, Channel 4 is selected as the target channel set.

[0172] S313: According to the target channel set, calculate the data loss amount generated during the channel switching process, and determine whether it is necessary to enable the edge cache unit to store the unfinished data packets, and comprehensively output the target channel adaptation result;

[0173] Based on the target channel set, during the channel switching process, since the data packets may not be successfully transmitted instantaneously during the switching, resulting in partial data loss. Assume that the current data transmission rate R t is 120 Mbps, and the switching time T sIf it is 0.5 seconds, the data loss amount is calculated as follows:

[0174] D loss = R t × T s ;

[0175] D loss = 120 × 0.5 = 60 Mbps;

[0176] Assume the storage capacity C of the edge cache unit cacHe is 100 Mbps, which is greater than the data loss amount D loss , so the edge cache unit can be enabled to store incomplete data packets to ensure data integrity, and finally the target channel adaptation result is comprehensively output.

[0177] Please refer to Figure 5 , the specific steps of S4 are as follows:

[0178] S411: Statistically calculate the delay requirements of computing tasks and the channel load occupancy rate, calculate the delay requirement intensity of each computing task, calculate the channel load occupancy rate, calculate the change rate of the channel load occupancy rate, establish the ratio of the task delay requirement intensity to the change rate of the channel load occupancy rate, and obtain the task delay pressure ratio data;

[0179] Based on the delay requirements of computing tasks and the channel load occupancy rate, first, obtain the delay requirement data of each computing task. For example, the delay requirement of task a is 50 milliseconds, task b is 100 milliseconds, and task c is 150 milliseconds. At the same time, record the load occupancy rate of the current channel. For example, the occupancy rate of channel 1 is 60%, channel 2 is 75%, and channel 3 is 50%. Calculate the change rate of the channel load occupancy rate. Assume that in the previous time period, the occupancy rate of channel 1 was 55%, then the change rate is calculated as (60% - 55%) / 55% = 9.09%. Similarly, the change rates of channel 2 and channel 3 are (75% - 70%) / 70% = 7.14% and (50% - 45%) / 45% = 11.11% respectively. Subsequently, calculate the ratio of the task delay requirement intensity to the change rate of the channel load occupancy rate. Taking task a and channel 1 as an example, this ratio is 50 milliseconds / 9.09% = 5.5, the ratio of task b and channel 2 is 100 milliseconds / 7.14% = 14, and the ratio of task c and channel 3 is 150 milliseconds / 11.11% = 13.5. Finally, summarize all calculated values to obtain the task delay pressure ratio data.

[0180] S412: According to the task delay pressure ratio data, combined with the current channel occupancy rate, judge the load change trend of the wireless task queue, and use the formula:

[0181]

[0182] Calculate the computing task load adjustment coefficient P adj , and adjust and obtain the basic priority of the computing task in combination with the channel occupancy rate, where T d represents the task delay demand intensity, R cH represents the channel load occupancy rate, L var,i represents the i-th variable of the channel load change rate, m represents the total number of variables, represents the comprehensive influence amount of the channel load change;

[0183] Call the task delay pressure ratio data, and combine the current channel occupancy rate. First, analyze the task delay pressure ratio of each task. For example, the task delay pressure ratio of task a is 5.5, that of task b is 14, and that of task c is 13.5. At the same time, obtain the current channel occupancy rate. For example, the occupancy rate of channel 1 is 60%, that of channel 2 is 75%, and that of channel 3 is 50%. Subsequently, calculate the load increase and decrease rate of the wireless task queue. Taking channel 1 as an example, if the occupancy rate of channel 1 in the previous time period was 55%, its growth rate is calculated as (60% - 55%) / 55% = 9.09%. The growth rates of channel 2 and channel 3 are (75% - 70%) / 70% = 7.14% and (50% - 45%) / 45% = 11.11% respectively. Calculate the product of the task delay pressure ratio and the load increase and decrease rate. Taking task a and channel 1 as an example, the product is 5.5 × 9.09% = 0.5. The product of task b and channel 2 is 14 × 7.14% = 1. The product of task c and channel 3 is 13.5 × 11.11% = 1.5. Next, calculate the task load adjustment coefficient of each task. Substitute the actual data into the formula. Taking task a and channel 1 as an example, let T d = 50 milliseconds, R cH = 60% = 0.6. Assuming that the comprehensive influence amount of the channel load change is 0.1, then

[0184]

[0185] Similarly, calculate the task load adjustment coefficients of task b and task c. Assuming that the comprehensive influence amounts of the load changes of channel 2 and channel 3 are both 0.1, then

[0186]

[0187] Then, adjust and calculate the task priority in combination with the channel occupancy rate. The higher the task load adjustment coefficient, the lower the priority should be.

[0188] Therefore, the priority of task a is the highest, and those of task b and task c are the second. Obtain the basic priority of the computing task.

[0189] S413: Based on the basic priority of computing tasks, analyze the current channel occupancy rate, judge the load change trend of the wireless task queue, adjust the priority order of low-latency tasks in the task queue, and obtain the task priority sorting list;

[0190] According to the basic priority of computing tasks, analyze the current channel occupancy rate. First, obtain the basic priority of computing tasks. For example, task a has the highest priority, followed by task b and task c. At the same time, obtain the current channel occupancy rate. For example, channel 1 is 60%, channel 2 is 75%, and channel 3 is 50%. Subsequently, calculate the load change trend of the wireless task queue. Taking channel 1 as an example, if the occupancy rate of channel 1 in the previous time period was 55%, its load change is calculated as 60% - 55% = 5%. The load changes of channel 2 and channel 3 are 75% - 70% = 5% and 50% - 45% = 5% respectively. Next, adjust the priority order of low-latency tasks in the task queue. Assume that task a is a low-latency task and task b and task c are high-latency tasks. Then, keep the priority of task a the highest, followed by task b and task c. Finally, screen out the adjusted task priority order to obtain the task priority sorting list.

[0191] Please refer to Figure 6 , the specific steps of S5 are as follows:

[0192] S511: Based on the target channel adaptation result and the task priority sorting list, count the total amount of computing resources required for computing tasks, calculate the upper limit of the current channel adaptation ability, analyze the ratio of the total amount of resources required for computing tasks to the upper limit of channel adaptation ability, and obtain the task channel adaptation degree ratio data;

[0193] Based on the target channel adaptation result and the task priority sorting list, calculate the adaptation degree of each computing task to the available channels. First, it is necessary to extract the computing resource requirements of each task, including CPU occupancy rate, storage space requirements, and task execution time, and quantify the resource occupancy of different tasks. The task resource requirements can be calculated using the normalization method to make it convenient for comparison within the range of 0-1. Suppose a task's computing requirements are 5GHz CPU, 8GB RAM, and 50GB storage space, while the computing power of the current available channel only provides 4GHz CPU, 6GB RAM, and 40GB storage space. Then, it is necessary to calculate the matching degree. The matching degree can be defined as the reciprocal of the sum of the ratios of each resource requirement to the provided ability. For example, the matching degree of this task is calculated as follows:

[0194]

[0195] Substitute the data for calculation to obtain:

[0196]

[0197] Similarly, calculate the matching degree for all tasks to obtain the task channel adaptability ratio data.

[0198] S512: Based on the task channel adaptability ratio data, determine whether the task computing requirement exceeds the upper limit of the current channel adaptability. If it exceeds, adjust the execution node of the task using the formula:

[0199]

[0200] Calculate the resource matching degree M after adjusting the task execution node adj , combined with the task priority sorting list, adjust the computing resource allocation node to obtain the task migration target, where R t represents the total computing resource required for the task, R ch represents the upper limit of the current channel adaptability, C node represents the current resource capacity of the computing node, S var,j represents the jth variable of the change in computing resources during the task migration, T d represents the task delay requirement intensity, L var represents the task channel load change rate, and m represents the total number of variables;

[0201] Call the task channel adaptability ratio data to determine whether the task computing requirement exceeds the upper limit of the current channel adaptability. If the matching degree of a certain task is less than 0.5, it is determined that its computing requirement exceeds the adaptability, and the execution node of the task needs to be adjusted. Calculate the resource matching degree after adjusting the task execution node. Assume that the maximum available computing power of the current channel is 6GHz CPU, 10GB RAM, and 60GB storage space, while the computing requirement of a certain task is still 5GHz CPU, 8GB RAM, and 50GB storage space, and calculate the adjusted matching degree using the formula.

[0202] Among them, assume that the task delay requirement T d is 80ms, the load change rate L var is 20ms / s, and the computing resource adjustment amount S var is 2GHz CPU, 2GB RAM, and 10GB storage space, then:

[0203]

[0204] Perform operations to obtain the resource matching degree of the task execution node, call the task priority sorting list, adjust the computing resource allocation node, and obtain the task migration target.

[0205] Table 5.1 Matching Degree Table of Computing Tasks and Available Channel Resources

[0206] Task ID CPU Requirement (GHz) RAM Requirement (GB) Storage Requirement (GB) Computing Power Limit (GHz, GB, GB) Match Degree X1 5 8 50 4，6，40 0.261 X2 6 10 60 6，10，60 1.00 X3 7 12 70 6，10，60 0.86

[0207] As shown in Table 5.1, the matching degree of X1 is relatively low, and the execution node needs to be adjusted.

[0208] S513: According to the task migration target, combined with the computing resource allocation node, dynamically adjust the allocation of task computing resources to obtain a wireless network computing task allocation plan;

[0209] Invoke the task migration target, combined with the computing resource allocation node, adjust the task computing resource allocation plan, and screen out the adjusted computing resource scheduling plan. Assume that the system provides three computing nodes J1, J2, and J3, with computing capabilities of 6GHz, 10GB, 60GB, 8GHz, 12GB, 80GB, and 10GHz, 16GB, 100GB respectively. If the original matching degree of task X1 is lower than 0.5, select the computing node with the highest matching degree and adjust the computing resource scheduling plan of task X1 as follows:

[0210]

[0211] After adjustment, task X1 is migrated to node J2, and the matching degree is increased to 0.85 to obtain a wireless network computing task allocation plan.

[0212] The above is only a preferred embodiment of the present invention, and it does not limit the present invention in other forms. Any person skilled in the art may use the disclosed technical content to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution content of the present invention still belong to the protection scope of the technical solution of the present invention.

Claims

1. A wireless network scheduling method driven by edge computing, characterized in that: The following steps are involved: S1: Obtain the task execution delay of the current wireless network computing task, monitor the computing power usage of the computing task on the edge computing node, analyze the execution interval of the computing task and the fluctuation trend of the computing power usage, calculate the current wireless network computing power usage ratio, and output the computing power utilization ratio data; S2: Read the data transmission rate, channel bandwidth occupancy rate and packet loss rate of the wireless network channel, calculate the ratio of the current data rate to the theoretical value of the channel bandwidth, analyze the current packet loss rate change trend, calculate the ratio of the actual transmission data volume per unit time to the maximum available data transmission volume of the channel, and obtain the channel load occupancy data; S3: combining the computing power utilization ratio data with the channel load occupancy data, screening currently available channels, calculating the data throughput ratio between the current channel and the adjacent channels, selecting a channel with better throughput as the target channel, and comprehensively outputting the target channel adaptation result; S4: Based on the latency requirements of the computing tasks and combined with the channel load occupancy, the basic task priority is assigned, the current channel occupancy is analyzed, the load change trend of the wireless task queue is determined, the task priority sequence is adjusted, and a task priority sorting list is obtained.

2. The edge computing driven wireless network scheduling method according to claim 1, characterized in that: The computing power utilization ratio data includes the task execution delay ratio, the computing power occupancy fluctuation coefficient, and the calculation task execution interval distribution record; the channel load occupancy data includes the data rate ratio, the channel bandwidth occupancy rate, and the data throughput ratio per unit time; the target channel adaptation result includes the target channel throughput capacity, the channel switching data loss amount, and the edge cache enablement status; the task priority sorting list includes the task basic priority, the low-latency task sorting adjustment result, and the wireless task queue load change trend analysis result.

3. The edge computing driven wireless network scheduling method according to claim 1, characterized in that: The specific steps of S1 are: S111: Obtain the task execution delay of the current wireless network computing task, detect the computing power occupancy of the computing task on the edge computing node, extract its time series data, calculate the execution interval time of the task, and extract the computing power occupancy data of each time period to obtain the computing task execution interval time series; S112: Based on the time series of the computing task execution intervals, statistical analysis is performed on the computing power usage at each time point, and the mean square error and fluctuation range are calculated using the formula: Calculate the computing power usage fluctuation trend coefficient T v , where C i represents the computing power utilization ratio at the i-th moment, n represents the length of the time series, Represents the average value of the computing power utilization ratio, max(C) represents the maximum value of the computing power utilization ratio in the time series, and min(C) represents the minimum value of the computing power utilization ratio in the time series; S113: Calculate the computing power occupancy ratio of the current wireless network according to the computing power occupancy fluctuation trend coefficient, normalize the computing power utilization rate, and extract computing power utilization rate ratio data.

4. The edge computing driven wireless network scheduling method according to claim 1, characterized in that: The specific steps of S2 are: S211: Acquire the data transmission rate, channel bandwidth occupancy rate and packet loss rate of the wireless network channel, calculate the ratio of the current data rate to the theoretical value of the channel bandwidth, analyze the channel utilization rate, and obtain the channel utilization rate analysis result; S212: Analyze the change trend of the current packet loss rate according to the channel utilization analysis result, and calculate the ratio of the actual transmission data volume per unit time to the maximum available data transmission volume of the channel, using the formula: Calculate and output the channel load occupancy ratio L, where D i represents the actual amount of data transmitted in the ith time interval, Z i represents the maximum available data transmission capacity of the channel in the i-th time interval, P i represents the packet loss rate of the ith time interval, and H represents the total number of time intervals; S213: Based on the channel load occupancy ratio, combined with the channel utilization ratio and the packet loss rate change trend, determine the channel interference situation and generate channel load occupancy data.

5. The edge computing driven wireless network scheduling method according to claim 1, characterized in that: The specific steps of S3 are: S311: judging whether the load of the current wireless network access point exceeds a set load threshold according to the computing power utilization ratio data and the channel load occupancy data, screening currently available channels, and obtaining an available channel set; S312: Based on the available channel set, using the formula: Calculate the data throughput ratio T between the current channel and the adjacent channel, select the channel with the best throughput according to the ratio, and get the target channel set, where B i represents the bandwidth of the ith channel, U i represents the utilization rate of the i-th channel, K i represents the packet loss rate of the ith channel, D i represents the data flow change of the ith channel, and M represents the total number of channels; S313: Calculate the amount of data loss generated during the channel switching process according to the target channel set, determine whether it is necessary to enable the edge cache unit to store unfinished data packets, and comprehensively output the target channel adaptation result.

6. The edge computing driven wireless network scheduling method according to claim 1, characterized in that: The specific steps of S4 are: S411: Count the delay requirements and channel load occupancy of computing tasks, calculate the delay requirement intensity of each computing task, calculate the channel load occupancy, calculate the rate of change of the channel load occupancy, establish a ratio of the task delay requirement intensity to the channel load occupancy rate, and obtain task delay pressure ratio data; S412: According to the task delay pressure ratio data and the current channel occupancy rate, the load change trend of the wireless task queue is determined by using the formula: Calculate the task load adjustment coefficient P adj , combined with the channel occupancy rate to adjust and obtain the basic priority of the computing task, where T d represents the task delay requirement intensity, R ch Represents the channel load occupancy, L var,i represents the i-th variable of the channel load change rate, m represents the total number of variables, Represents the comprehensive impact of channel load changes; S413: Based on the calculation task basic priority, analyze the current channel occupancy rate, determine the load change trend of the wireless task queue, adjust the priority order of the low-latency tasks in the task queue, and obtain a task priority sorting list.

7. The edge computing driven wireless network scheduling method according to claim 1, characterized in that: The method further comprises: S5: According to the target channel adaptation result and the task priority sorting list, analyze the matching degree between the current computing task and the available channel, determine whether the task computing demand exceeds the upper limit of the current channel adaptation capacity, and if so, adjust the execution node of the task, and use the computing resource allocation node as the task migration target to obtain the wireless network computing task allocation plan; The wireless network computing task allocation scheme includes task execution nodes, computing resource allocation targets, and task migration matching degrees.

8. The edge computing driven wireless network scheduling method according to claim 7, characterized in that: The specific steps of S5 are: S511: Based on the target channel adaptation result and the task priority sorting list, the total amount of computing resources required for the computing task is counted, and the upper limit of the current channel adaptation capability is calculated, and the ratio of the total amount of resources required for the computing task to the upper limit of the channel adaptation capability is analyzed to obtain task channel adaptation ratio data; S512: Based on the task channel adaptability ratio data, determine whether the task computing requirement exceeds the current channel adaptation capability upper limit. If it exceeds, adjust the execution node of the task, using the formula: Calculate the resource matching degree M of the task execution node after adjustment adj , combined with the task priority sorting list, adjust the computing resource allocation node and obtain the task migration target, where R t Represents the total amount of computing resources required for the task, R ch Represents the upper limit of the current channel adaptation capability, C node Represents the current resource capacity of the computing node, S var,j represents the jth variable of computing resource changes during task migration, T d Represents the task delay requirement intensity, L var represents the rate of change of task channel load, and m represents the total number of variables; S513: According to the task migration target, in combination with the computing resource allocation node, dynamically adjust the allocation of task computing resources to obtain a wireless network computing task allocation plan.

Images