Electronic device network data detection optimization control method based on edge computing

Abstract

The application relates to the technical field of electronic communication, and discloses an electronic device network data detection optimization control method based on edge computing. The application performs real-time detection on data flow flowing through an edge node, dynamically calculates data value parameters which are fused with time decay, event characteristics and context correlation, synchronously monitors network transmission cost and cache pressure parameters, inputs the above parameters into a preset decision function, and generates optimization control instructions such as instant forwarding, cache waiting or aggregation / discard for each data unit. The application combines the dynamic evaluation of data intrinsic value with the deep perception of external resource state, realizes the adaptive global optimization of transmission strategy, effectively guarantees the transmission efficiency of high-value data in a complex edge environment, and significantly improves the overall utilization efficiency of network bandwidth and cache resources.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electronic communication, in particular to an electronic device network data detection optimization control method based on edge computing. BACKGROUND

[0002] Under the background of rapid development of Internet of Things and Industrial Internet, network data generated by a large number of electronic devices needs to be preprocessed and forwarded by edge computing nodes; in the prior art, the processing of data flow by edge nodes depends on static strategies, for example, setting a fixed priority queue based on data source or protocol type, or performing traffic shaping and cache management according to a simple network bandwidth threshold; these methods usually regard data as homogeneous flow, lack of fine perception of the intrinsic value of data and its dynamic change, and the decision logic fails to fully consider the real-time characteristics of data content, the relevance of business context and the characteristics of data value decay over time.

[0003] However, the above prior art solution has obvious defects: first, the static and single value evaluation dimension cannot accurately reflect the real importance of data at a specific moment, resulting in delay of high timeliness and high event value data due to queuing or network congestion, affecting system real-time performance; second, the decision-making process often considers network state or cache condition in isolation, ignoring the dynamic coupling relationship between data value, transmission cost and cache pressure, making it difficult to achieve global optimization in complex and variable edge network environment, and easily causing bandwidth waste or cache overflow; finally, lacking of quantitative modeling of data value decay and associated value may lead to loss of key event context data or ineffective transmission of low value expired data, thereby reducing the overall resource utilization efficiency and service quality of edge computing nodes. SUMMARY

[0004] The technical problem to be solved by the present application is that the existing edge data processing method adopts a static and isolated decision strategy, lacks quantitative modeling and collaborative optimization of dynamic data value, multi-dimensional state coupling and value decay, resulting in low resource utilization efficiency and impaired service quality, therefore we propose an electronic device network data detection optimization control method based on edge computing.

[0005] In order to achieve the above purpose, the following technical solution is adopted: an electronic device network data detection optimization control method based on edge computing, applied to an edge computing node, the method comprising:

[0006] Dynamic data value evaluation: real-time detection and analysis of electronic device network data flow passing through the edge computing node, dynamic calculation of comprehensive value evaluation parameters of each data unit at the current moment based on data content, business type, real-time state and association relationship;

[0007] network state and cache pressure monitoring: real-time monitoring of the current state of the network transmission link between the edge computing node and the upstream node or the cloud, obtaining the network transmission cost parameter; at the same time, monitoring the occupation state of the cache space of the edge computing node itself, obtaining the cache pressure parameter;

[0008] Dynamic decision and optimization control: based on the comprehensive value evaluation parameter, the network transmission cost parameter and the cache pressure parameter, the decision function is calculated through the preset, and the real-time transmission control decision is generated for each data unit; the transmission control decision includes immediate forwarding, delayed caching waiting or executing aggregation / discard operation;

[0009] Control instruction execution: according to the transmission control decision, the corresponding data forwarding, cache scheduling or resource release operation is executed, the optimization control of the electronic device network data flow is realized, the high value data transmission efficiency is guaranteed, and the overall network bandwidth utilization and edge cache resource are optimized.

[0010] Preferably, the data value dynamic evaluation step specifically comprises:

[0011] S11: basic value assignment: according to the source electronic device type, service service level agreement and agreement type of the data unit, the initial basic value quantitative parameter is assigned;

[0012] S12: real-time feature detection and gain calculation: the data unit is subjected to lightweight real-time analysis, whether it contains predefined abnormal feature or key event feature is identified, and the corresponding event sensitive gain factor is calculated;

[0013] S13: context value correlation analysis: other data units logically associated with the current data unit are identified, and the context correlation gain factor is calculated according to the comprehensive value evaluation parameter of the associated other data units;

[0014] S14: time decay calculation: according to the business time effectiveness category of the data unit, the value decay coefficient is determined, and the time decay factor is calculated based on the coefficient and the data survival time;

[0015] S15: dynamic value synthesis: based on the basic value quantitative parameter, the time decay factor, the event sensitive gain factor and the context correlation gain factor, the comprehensive value evaluation parameter of the data unit at the current time is synthesized.

[0016] Preferably, in the network state and cache pressure monitoring step, the network transmission cost parameter is obtained specifically as follows:

[0017] Real-time acquisition of the current average transmission delay and available bandwidth of the network transmission link;

[0018] The current average transmission delay and the reciprocal of the available bandwidth are weighted and summed based on predefined weight coefficients, and the result is taken as a network transmission cost parameter representing the current network transmission difficulty level.

[0019] Preferably, in the network state and cache pressure monitoring step, the cache pressure parameter is specifically obtained by:

[0020] The total capacity and the current occupied amount of the cache space of the edge computing node are obtained in real time.

[0021] The ratio of the current occupied amount to the total capacity is calculated as the cache space occupancy rate, and the cache pressure parameter is determined based on the occupancy rate.

[0022] Preferably, in the dynamic decision and optimization control step, a real-time transmission control decision is generated for each data unit, which is specifically:

[0023] The comprehensive value evaluation parameter, the network transmission cost parameter and the cache pressure parameter of the data unit are input into the decision function, and a decision score is calculated.

[0024] The decision score is compared with the preset high threshold and low threshold:

[0025] If the decision score is higher than the high threshold, an "immediate forwarding" decision is generated;

[0026] If the decision score is between the low threshold and the high threshold, a "delayed cache waiting" decision is generated, and the optimal waiting time is estimated.

[0027] If the decision score is lower than or equal to the low threshold, an "aggregation or discard" decision is generated.

[0028] Preferably, the decision function considers the expected decay penalty of data value in its calculation process, which is determined based on the value decay coefficient of the data unit and the estimated delay waiting time, and is used to reduce the decision score of those data units whose value will significantly decay during the waiting period.

[0029] Preferably, the decision function further introduces a nonlinear suppression factor based on the cache pressure parameter in its calculation process. When the cache space occupancy rate exceeds the preset pressure activation threshold, the decision score is significantly reduced by the nonlinear suppression factor, so that only data units with extremely high comprehensive value evaluation parameters are allowed to obtain immediate forwarding qualification when the cache resource is tight.

[0030] Preferably, the estimated optimal waiting time of the "delayed cache waiting" decision is determined by predicting the future trend of the network transmission cost parameter and finding the future time point that can maximize the expected transmission utility.

[0031] Preferably, in the data value dynamic evaluation step, the integrated value evaluation parameter of the synthetic data unit at the current time is defined by the following formula:

[0032]

[0033] wherein: represents the basic value quantization parameter of the data unit ; represents the value decay coefficient of the data unit ; represents the survival time of the data unit from generation to the current time ; represents the event or abnormal feature score of the data unit obtained by real-time feature detection; represents the activation threshold of the event-sensitive gain; represents the weight coefficient of the event-sensitive gain; represents the set of data units logically associated with the data unit ; represents the integrated value evaluation parameter of the associated data unit at the previous time; represents the basic value quantization parameter of the associated data unit ; represents the association weight coefficient of the data unit and ; represents the weight coefficient of the contextual association gain.

[0034] Preferably, in the dynamic decision and optimization control step, the method for the decision function to calculate the decision score is defined by the following formula:

[0035]

[0036] wherein: represents the decision score of the data unit at time ; represents the integrated value evaluation parameter of the data unit at time ; represents the network transmission cost parameter at the current time; represents the network transmission cost adjustment coefficient; represents the current occupied amount of the edge computing node cache; represents the total capacity of the edge computing node cache; represents the activation threshold of the cache pressure nonlinear suppression. a steepness coefficient representing the cache pressure suppression; a value decay coefficient representing the data unit a value decay coefficient representing the data unit an estimated delay waiting time length; a weight coefficient representing the value decay expected penalty; a sensitivity adjustment coefficient representing the value decay penalty;

[0037] The transmission control decision is based on: setting a high threshold and a low threshold , when , the decision is "immediate forwarding", when , the decision is "delayed cache waiting", and when , the decision is "perform aggregation or discard".

[0038] Technical effects and advantages of the present application:

[0039] In the present application, the present application realizes fine value perception and self-adaptive optimization control of network data of electronic devices by constructing a data dynamic value evaluation model and an intelligent decision function integrating multi-dimensional state perception, thereby effectively improving the data processing and transmission performance of edge nodes. Specifically, the present application solves the problem that traditional methods cannot accurately identify high-value data by comprehensively evaluating the dynamic value of data timeliness, real-time event features and context association, and realizes global collaborative optimization of data forwarding strategy by coupling data value, network transmission cost and cache pressure in a unified decision framework. Finally, the method can guarantee the priority transmission and low-delay delivery of high-value and high-timeliness data under complex network conditions, while intelligently balancing network bandwidth and cache resources, reducing invalid transmission and data loss, and significantly improving the overall efficiency, resource utilization and operation reliability of edge computing services. BRIEF DESCRIPTION OF DRAWINGS

[0040] The disclosure of the present application will be described with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of the present application. In the drawings, the same reference numerals are used to refer to the same components:

[0041] Fig. 1 is a method step diagram of the present application;

[0042] Fig. 2 is a structure topology diagram of the present application. DETAILED DESCRIPTION

[0043] ​It is easy to understand that, according to the technical solutions of the present application, those skilled in the art can propose various structures and implementation modes that can be replaced with each other without changing the essential spirit of the present application. Therefore, the following specific embodiments and drawings are only exemplary descriptions of the technical solutions of the present application, and should not be considered as the whole or as a limitation or restriction on the technical solutions of the present application.

[0044] Referring to Figs. 1-2 As shown in the drawings, the present application provides a technical solution: an edge computing-based electronic device network data detection optimization control method applied to an edge computing node, aiming to solve the problems of existing edge computing nodes in processing massive electronic device network data, such as non-precise data value evaluation, insufficient network resource utilization, and unintelligent cache management. The method dynamically evaluates data value, real-time monitors network state and cache pressure, and makes intelligent decisions based on this, thereby optimizing data flow transmission and processing.

[0045] The method of the present application includes the following main steps:

[0046] Data value dynamic evaluation step: real-time detection and analysis of electronic device network data flow through the edge computing node; the core is to dynamically calculate the comprehensive value evaluation parameter of each data unit at the current time based on data content, business type, real-time state and association; this means that the data value is not fixed but dynamically adjusted with time, events and context.

[0047] Specifically, the data value dynamic evaluation step can include the following sub-steps:

[0048] S11: Basic value assignment: according to the source electronic device type, business service level agreement and protocol type of the data unit, assign it an initial basic value quantitative parameter; for example, data from a key industrial control device may have a higher basic value, while data from a normal environmental sensor may be lower; SLA defines the priority of data transmission and processing, and the protocol type may also imply the business importance of the data.

[0049] S12: Real-time feature detection and gain calculation: lightweight real-time analysis of data units to identify whether they contain predefined abnormal features or key event features, and calculate the corresponding event-sensitive gain factor; for example, when sensor data suddenly exceeds the normal threshold, or abnormal behavior is detected in a video stream, the value of the data unit will instantaneously increase; this real-time feature detection can be rule-based or based on a lightweight machine learning model.

[0050] S13: Contextual Value Correlation Analysis: Identify other data units that are logically related to the current data unit, and calculate a contextual correlation gain factor based on the integrated value evaluation parameters of the related data units; for example, a temperature sensor data itself has low value, but if it is related to a smoke sensor data that is alarming, its value will be significantly increased due to the correlation; such correlation can be based on time, space, device topology or business logic.

[0051] S14: Time Decay Calculation: Determine the value decay coefficient of the data unit according to its business time sensitivity category, and calculate its time decay factor based on the coefficient and the data survival time; for example, the time sensitivity of real-time control instructions is very high, and its value decay coefficient will be large, while the time sensitivity of historical log data is low, and its value decay coefficient will be small; the longer the data survival time, the more significant the value decay.

[0052] S15: Dynamic Value Synthesis: Synthesize the integrated value evaluation parameters of the data unit at the current time based on the basic value quantitative parameters, time decay factors, event sensitive gain factors and contextual correlation gain factors.

[0053] In a preferred embodiment, the synthesis method of the integrated value evaluation parameters of the data unit at the current time is defined by the following formula:

[0054]

[0055] Wherein:

[0056] represents the basic value quantitative parameters of the data unit ; represents the value decay coefficient of the data unit ; represents the survival time of the data unit from generation to the current time ; represents the event or abnormal feature score of the data unit obtained by real-time feature detection; represents the activation threshold of event sensitive gain; represents the weight coefficient of event sensitive gain; represents the set of data units that are logically related to the data unit ; represents the integrated value evaluation parameters of the related data unit at the previous time; represents the basic value quantitative parameters of the related data unit ; represents the data unit and The correlation weight coefficient; The weighting coefficients representing the contextual gain are used in this formula, which comprehensively considers the initial value of the data, its timeliness, the impact of real-time events, and its correlation with other data, thus enabling a dynamic and refined assessment of the data's value.

[0057] Network status and cache pressure monitoring steps: This step monitors the current status of the network transmission link between the edge computing node and the upstream node or the cloud in real time and obtains network transmission cost parameters; at the same time, it monitors the occupancy status of the edge computing node's own cache space and obtains cache pressure parameters.

[0058] In terms of obtaining network transmission cost parameters, the current average transmission latency and available bandwidth of the network transmission link can be collected in real time. Based on predefined weighting coefficients, the reciprocals of the current average transmission latency and available bandwidth are weighted and summed, and the result is used as a network transmission cost parameter to characterize the difficulty of current network transmission. For example, the higher the transmission latency and the lower the available bandwidth, the larger the network transmission cost parameter, indicating that the transmission is more difficult.

[0059] In terms of obtaining cache pressure parameters, the total capacity and current occupied amount of the edge computing node cache space can be obtained in real time; the ratio of the current occupied amount to the total capacity is calculated as the cache space occupancy rate, and the cache pressure parameters are determined based on this occupancy rate; the higher the cache occupancy rate, the greater the cache pressure.

[0060] Dynamic decision-making and optimization control steps: This step is based on comprehensive value assessment parameters, network transmission cost parameters, and buffer pressure parameters. It calculates the data using a preset decision function to generate real-time transmission control decisions for each data unit. Transmission control decisions include immediate forwarding, delayed buffering and waiting, or performing aggregation / drop operations.

[0061] In a preferred embodiment, the preset decision function generates real-time transmission control decisions for each data unit by: inputting the comprehensive value assessment parameters, network transmission cost parameters, and buffer pressure parameters of the data unit into the decision function to calculate a decision score; and comparing the decision score with preset high and low thresholds.

[0062] If the decision score is higher than the high threshold, an "immediately forward" decision is generated; this means that the data unit has high value, low transmission cost, and low cache pressure, and should be transmitted first.

[0063] If the decision score is between the low and high thresholds, a "delayed caching wait" decision is generated, and the optimal waiting time is estimated. This means that the data unit is of medium value or the current transmission conditions are not good, so it can be temporarily cached and wait for a better transmission opportunity.

[0064] If the decision score is lower than or equal to the low threshold, a "perform aggregation or discard" decision is generated; this means that the data unit is of low value, high transmission cost or large cache pressure, and should be considered for aggregation processing to reduce transmission volume, or directly discarded to release resources.

[0065] In a further embodiment, the decision function calculation process takes into account an expected decay penalty of data value; the expected decay penalty is determined based on the value decay coefficient of the data unit and the estimated delay latency, and is used to reduce the decision score tendency of those data units whose value will significantly decay during the waiting period; this makes the decision more intelligent, avoiding unnecessary delay for high-value data that will soon expire.

[0066] In yet another further embodiment, the decision function calculation process further introduces a nonlinear suppression factor based on the cache pressure parameter; when the cache space occupancy rate exceeds the preset pressure activation threshold, the nonlinear suppression factor significantly reduces the decision score, so that only data units with extremely high comprehensive value evaluation parameters are allowed to obtain immediate forwarding qualification when cache resources are tight; this ensures that the edge node can prioritize the most critical data in extreme resource shortage situations.

[0067] In terms of estimating the optimal waiting time of the "delay cache waiting" decision, a future time point that maximizes the expected transmission utility can be determined by predicting the trend of future network transmission cost parameters; for example, based on historical network data or prediction models, the network bandwidth and latency changes in the future can be estimated, so that the best time for transmission can be selected.

[0068] In a specific embodiment, the method for the decision function to calculate the decision score is defined by the following formula:

[0069]

[0070] wherein, represents the decision score of the data unit at time ; represents the comprehensive value evaluation parameter of the data unit at time ; represents the network transmission cost parameter at the current time; represents the network transmission cost adjustment coefficient; represents the current occupied amount of the edge computing node cache; represents the total capacity of the edge computing node cache; represents the activation threshold of the cache pressure nonlinear suppression; a steepness coefficient representing cache pressure suppression; a value decay coefficient representing data units ; a weight coefficient representing expected penalty of value decay; an estimated delay waiting duration of data units ; a sensitivity adjustment coefficient representing value decay penalty;

[0071] The transmission control decision is based on setting a high threshold and a low threshold , when , the decision is “immediate forwarding”, when , the decision is “delayed cache waiting”, and when , the decision is “perform aggregation or discard”, which comprehensively considers data value, transmission cost, cache pressure, and expected decay of data value, and realizes adaptive decision-making for complex edge environments through a nonlinear function.

[0072] The control instruction execution step includes:

[0073] According to the transmission control decision, the corresponding data forwarding, cache scheduling, or resource release operation is performed to realize the optimization control of the electronic device network data flow, which optimizes the overall network bandwidth utilization and edge cache resources while ensuring the transmission efficiency of high-value data.

[0074] For data units with “immediate forwarding” decision, the edge computing node will immediately send them to the upstream node or the cloud through the network transmission link.

[0075] For data units with “delayed cache waiting” decision, the edge computing node will store them in the local cache and schedule them according to the estimated optimal waiting duration; during the waiting period, the network state can be continuously monitored, and forwarding can be performed at the best opportunity.

[0076] For data units with “perform aggregation or discard” decision, the edge computing node will process them according to the preset strategy; for example, for low-value and aggregable data, they can be aggregated with other similar data to reduce transmission volume; for low-value and time-expired data, they can be directly discarded to release cache space.

[0077] Through the synergistic effect of the above steps, the method of the present application can dynamically and intelligently manage the data flow of the edge computing node, ensure that high-value data is given priority and processed efficiently, and effectively utilize limited network bandwidth and cache resources, thereby improving the overall performance and reliability of edge computing services.

[0078] The technical scope of the present application is not limited to the above description, and those skilled in the art can make various modifications and changes to the above embodiments without departing from the technical idea of the present application, and these modifications and changes should all belong to the protection scope of the present application.

Claims

1. A method for optimizing and controlling network data detection in electronic devices based on edge computing, characterized in that, Applied to edge computing nodes, the method includes: Dynamic data value assessment: Real-time detection and analysis of network data streams of electronic devices passing through the edge computing nodes; dynamic calculation of comprehensive value assessment parameters for each data unit at the current moment based on data content, service type, real-time status and correlation. Network status and cache pressure monitoring: Real-time monitoring of the current status of the network transmission link between the edge computing node and the upstream node or the cloud to obtain network transmission cost parameters; at the same time, monitoring the occupancy status of the edge computing node's own cache space to obtain cache pressure parameters; Dynamic decision-making and optimization control: Based on the comprehensive value assessment parameters, the network transmission cost parameters, and the cache pressure parameters, a real-time transmission control decision is generated for each data unit through a preset decision function; the transmission control decision includes immediate forwarding, delayed cache waiting, or performing aggregation / discarding operations; Control command execution: Based on the transmission control decision, execute corresponding data forwarding, cache scheduling or resource release operations to achieve optimized control of the network data flow of the electronic device, so as to optimize the overall network bandwidth utilization and edge cache resources while ensuring the efficiency of high-value data transmission.

2. The edge computing-based network data detection optimization control method for electronic devices according to claim 1, characterized in that, The specific steps for dynamic data value assessment include: S11: Basic Value Assignment: Assign initial basic value quantification parameters to the data unit based on the type of electronic device from which it originates, the service level agreement, and the agreement type. S12: Real-time feature detection and gain calculation: Perform lightweight real-time analysis on data units to identify whether they contain predefined abnormal features or key event features, and calculate the corresponding event-sensitive gain factor. S13: Contextual value correlation analysis: Identify other data units that are logically related to the current data unit, and calculate the contextual correlation gain factor based on the comprehensive value assessment parameters of the other related data units; S14: Timeliness decay calculation: Determine the value decay coefficient of the data unit according to its business timeliness category, and calculate its time decay factor based on the coefficient and the data's lifespan. S15: Dynamic value synthesis: Based on the basic value quantification parameters, the time decay factor, the event-sensitive gain factor, and the context-related gain factor, synthesize the comprehensive value assessment parameters of the data unit at the current moment.

3. The edge computing-based network data detection optimization control method for electronic devices according to claim 1, characterized in that, In the network status and cache pressure monitoring step, the network transmission cost parameter is obtained specifically as follows: Real-time acquisition of the current average transmission latency and available bandwidth of the network transmission links; Based on predefined weighting coefficients, the current average transmission delay and the reciprocal of the available bandwidth are weighted and summed, and the result is used as a network transmission cost parameter to characterize the current network transmission difficulty.

4. The edge computing-based network data detection optimization control method for electronic devices according to claim 1, characterized in that, In the network status and cache pressure monitoring steps, the cache pressure parameters are obtained specifically as follows: Real-time acquisition of the total capacity and current occupied amount of cache space of edge computing nodes; Calculate the ratio of the currently occupied amount to the total capacity as the cache space occupancy rate, and determine the cache pressure parameter based on this occupancy rate.

5. The edge computing-based network data detection optimization control method for electronic devices according to claim 1, characterized in that, In the dynamic decision-making and optimization control steps, generating real-time transmission control decisions for each data unit specifically involves: The comprehensive value assessment parameters of the data unit, the network transmission cost parameters, and the cache pressure parameters are input into the decision function to calculate a decision score; The decision score is compared with preset high and low thresholds: If the decision score is higher than the high threshold, then a "forward immediately" decision is generated; If the decision score is between the low threshold and the high threshold, a "delayed cache wait" decision is generated, and the optimal waiting time is estimated. If the decision score is lower than or equal to the low threshold, a "execute aggregation or discard" decision is generated.

6. The edge computing-based network data detection optimization control method for electronic devices according to claim 5, characterized in that, The decision function takes into account the expected decay penalty of data value in its calculation process. The expected decay penalty is determined based on the value decay coefficient of the data unit and the estimated delay waiting time, and is used to reduce the decision score tendency of data units whose value will significantly decay during the waiting period.

7. The edge computing-based network data detection optimization control method for electronic devices according to claim 5, characterized in that, The decision function further incorporates a nonlinear inhibition factor based on cache pressure parameters during its calculation process. When the cache space occupancy rate exceeds the preset pressure activation threshold, this nonlinear inhibition factor significantly reduces the decision score, thereby allowing only data units with extremely high comprehensive value evaluation parameters to obtain immediate forwarding qualification when cache resources are scarce.

8. The edge computing-based network data detection optimization control method for electronic devices according to any one of claims 5-7, characterized in that, The optimal waiting time for the "delayed caching wait" decision is determined by predicting the changing trend of future network transmission cost parameters and finding a future time point that maximizes the expected transmission utility.

9. The edge computing-based network data detection optimization control method for electronic devices according to claim 8, characterized in that, In the dynamic data value assessment step, the comprehensive value assessment parameters of the synthesized data unit at the current moment are... The method is defined by the following formula: ； in: Representing data unit Basic value quantification parameters; Representing data unit The value decay coefficient; Representing data unit From generation to the current moment Survival time; This represents a data unit obtained through real-time feature detection. The score of the event or abnormal feature; The activation threshold representing the event-sensitive gain; The weighting coefficients representing the event-sensitive gain; Representation and data unit A logically related set of data units; Represents associated data unit The comprehensive value assessment parameters at the previous moment; Represents associated data unit Basic value quantification parameters; Representing data unit and The correlation weight coefficient; The weighting coefficients represent the context-related gain.

10. The edge computing-based network data detection optimization control method for electronic devices according to claim 9, characterized in that, In the dynamic decision-making and optimization control step, the decision function calculates the decision score. The method is defined by the following formula: ； in, Representing data unit At any moment Decision score; Representing data unit At any moment Comprehensive value assessment parameters; This represents the network transmission cost parameter at the current moment; This represents the network transmission cost adjustment coefficient; This indicates the current amount of cache occupied by the edge computing node; This represents the total capacity of the edge computing node cache; The activation threshold represents the nonlinear suppression threshold for cache pressure. The steepness coefficient representing cache pressure suppression; Representing data unit The value decay coefficient; Represented as data unit Estimated delay time; The weighting coefficient representing the expected penalty for value decay; The sensitivity adjustment coefficient representing the value decay penalty; The transmission control decision is based on setting a high threshold. With low threshold ,when The decision was "forward immediately". The decision is "delayed buffering wait" when... The decision is "execute aggregation or discard".

Images