Device multi-parameter statistical calculation method and system based on general task scheduling platform

Abstract

The invention discloses an equipment multi-parameter statistical calculation method and system based on a general task scheduling platform. The method comprises the steps of preliminary scheduling strategy generation, dynamic scheduling, sampling frequency adjustment, virtual data integration and task priority adjustment. The system comprises an equipment task mode extraction module, a dynamic scheduling module, a sampling frequency adjustment module, a virtual data integration module and a task priority adjustment module. Through intelligent task scheduling and efficient resource management, the efficiency and stability of cooperative work of multiple devices are remarkably improved. Through accurate analysis and identification of an equipment task mode and a cooperative network, resource allocation is optimized, task conflicts and equipment overload conditions are reduced, and dynamic scheduling and priority adjustment of tasks are realized. And virtual data prediction is carried out on the parameters of the discontinuous position, so that the monitoring of the running state of the equipment is still not influenced under the condition of data discontinuity. The continuity and accuracy of data acquisition are enhanced, and monitoring blind spots caused by data missing are avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of task scheduling, and more specifically, to a device multi-parameter statistical calculation method and system based on a general task scheduling platform. Background Art

[0002] Task scheduling primarily involves effectively allocating and managing task resources in a multi-device environment to optimize system performance and device utilization. The core task in this area is to ensure that individual devices operate at peak efficiency when working together through intelligent scheduling algorithms and parameter statistical analysis. Technical implementation typically includes analyzing device task patterns, real-time monitoring of device status, dynamic adjustment of task priorities, and device load balancing. The goal of this area is to maximize resource utilization, reduce task conflicts, and avoid device overload. Real-time collection and analysis of device parameters are also crucial to improving data integrity and accuracy. This provides robust data support for more precise device status monitoring and management decisions, thereby enhancing overall system efficiency and reliability.

[0003] In the collaborative work of multiple devices, low efficiency is often encountered. The main reason is the uneven allocation of resources, which often leads to task conflicts and overload of individual devices. This unbalanced resource allocation not only limits the overall operating efficiency of the system, but also easily causes some devices to suffer performance degradation or failure due to excessive load. In addition, the sampling frequency of traditional systems is usually fixed and cannot be adjusted according to the actual situation of parameter changes. This not only leads to low data collection efficiency, but also adds unnecessary data processing overhead, increasing the burden on the system. What is more serious is that frequent data interruptions lead to blind spots in equipment status monitoring, resulting in incomplete equipment status data. This data incompleteness weakens the accurate assessment of the equipment operating status, thereby affecting the scientificity and accuracy of equipment management and decision-making. When faced with complex environments, existing technologies lack effective response measures and are unable to handle data anomalies and equipment status fluctuations in a timely manner, which in turn affects the stability of equipment operation. Therefore, this is a technical problem that urgently needs to be solved in this field.

[0004] A Chinese patent document (Application No. 202411767248.0, Application Date: December 4, 2024) discloses a task allocation method, system, electronic device, storage medium, and product, including: for any distributed task in the distributed system, determining the task requirement of the distributed task and obtaining the resource status of each distributed device; calculating a weighted score for each distributed device to execute the distributed task based on the task requirement, each resource status, and a dynamic weighting algorithm; and allocating the distributed task to the optimal device with the highest weighted score for execution. However, it does not address the blind spot issue in monitoring the operating status of devices in the event of data interruptions. Summary of the Invention

[0005] In view of this, the present invention provides a device multi-parameter statistical calculation method and calculation system based on a general task scheduling platform to solve technical problems such as low overall operating efficiency and data interruptions, and blind spots in monitoring the operating status of equipment.

[0006] In a first aspect, the present application provides a device multi-parameter statistical calculation method based on a general task scheduling platform, comprising:

[0007] Generate preliminary scheduling strategy: Analyze historical task data of each device, extract device task patterns, build a multi-device task collaboration network, identify collaborative device groups and task conflict nodes, generate task pressure indicators, adjust device task priorities, and generate preliminary scheduling strategy;

[0008] Dynamic scheduling: Based on the status of each device after the implementation of the preliminary scheduling strategy as a reference baseline, the current load level of each device is determined, a load threshold is set, the operating status of nodes exceeding the load threshold is read, and nodes with load levels below the average load level of all nodes are identified, and tasks are dynamically scheduled;

[0009] Adjust the sampling frequency: Based on the real-time device data, calculate and mark the change rate of each parameter, set the change fluctuation range, and adjust the sampling frequency according to the fluctuation range of the parameter change rate of different devices;

[0010] Virtual data integration: Continuously collect parameters of the device based on the adjusted sampling frequency. For the collected parameters, analyze whether the data collected at each frequency interval is complete, extract information about devices with parameter discontinuities, and combine the current status of the device with the historical parameter change rate and change pattern to determine the parameter stability of the device. Predict virtual parameter data at the discontinuity location and integrate it with the sampled parameters.

[0011] Adjust task priority: For devices whose parameters are not collected and whose collection time exceeds the set time, a high-speed mode rule is generated to determine whether the device status corresponding to the parameters meets the high-speed mode rule. If the device status does not meet the high-speed mode rule requirements, dynamic scheduling is repeated to assign tasks. If the high-speed mode rule requirements are met, the task priority is evaluated according to the high-speed mode rule, and tasks that need to be prioritized and can be suspended are identified, and the task execution order is adjusted.

[0012] The extracting device task mode includes: analyzing historical task data of each device, identifying task execution frequency and average execution time, and generating a task mode, wherein the task mode includes a high-frequency task mode and a periodic task mode;

[0013] The construction of a multi-device task collaboration network includes: analyzing the similarity of parameter types of each device, identifying the task collaboration relationship between devices, constructing a multi-device task collaboration network diagram, and identifying the collaboration frequency between each device based on the task collaboration network diagram, and evaluating the impact on the overall task completion efficiency; wherein,

[0014] The collaboration frequency of the devices is:

[0015]

[0016] Among them, CD(v) is the degree centrality, which indicates the collaboration frequency of task node v, D(v) is the number of edges of task node v, that is, the number of devices collaborating with task node v, and n is the total number of devices;

[0017] The overall task completion efficiency is:

[0018]

[0019] Among them, E is the overall task completion efficiency, w i is the edge weight, i.e. the frequency of collaboration between devices. The larger the weight, the more frequent the collaboration. i The efficiency of the equipment in performing its tasks;

[0020] Generating the task pressure index includes: establishing a task pressure prediction model based on the identified task mode and collaborative relationship, predicting future task pressure, and calculating the pressure index of the device in different task scenarios; wherein,

[0021] The pressure indicators of the equipment in different mission scenarios are:

[0022] P i =αF i +βU i +γC i ;

[0023] C i =C D (i)×(n-1);

[0024] Among them, P i is the pressure index of device i in a certain task scenario, F i is the task execution frequency, U i is the resource utilization rate of device i, C i is the collaboration frequency, which indicates the frequency of device i collaborating with other devices in the task collaboration network, and α, β, and γ are weight coefficients.

[0025] Optionally, the rate of change of each parameter is:

[0026]

[0027] Among them, R is the rate of change, which represents the approximate value of the derivative involved in the change, V t+1 and V t-1 Respectively represent the parameter values ​​collected at the time point and , which is the time interval between two adjacent time points;

[0028] The frequency adjustment includes: dividing the parameters into three stages of change levels according to the change rate of the parameters, and setting sampling strategies for different levels. The change levels include high level, stable level and medium level, among which, the definition of increasing the sampling frequency strategy is assigned to the high level; the definition of maintaining the sampling frequency strategy is assigned to the stable level; and the definition of reducing the sampling frequency strategy is assigned to the medium level.

[0029] Optionally, determining the parameter stability of the device includes: extracting device information with data discontinuities, recording the location and frequency of the discontinuities, analyzing the impact of the discontinuities on the monitoring of the device's operating status, and generating a discontinuity severity level; evaluating whether the current parameters are within a corresponding normal fluctuation range, and detecting abnormal patterns in the parameter fluctuations; wherein,

[0030] The interruption impact index in the analysis of the impact of interruptions on equipment operation status monitoring is:

[0031]

[0032] Among them, I is the interruption impact index, which represents the comprehensive impact of interruptions on equipment operation status monitoring, ΔKPI k is the rate of change of the kth performance indicator, ω k is the weight of the kth performance indicator, δ is the weight factor of the interruption duration, D is the interruption duration, n is the total number of all performance indicators, KPI k，after is the kth value measured after the interruption occurs, KPI k，before is the kth value measured before the discontinuity occurs;

[0033] The method of predicting the virtual parameter data of the discontinuous position includes: establishing a time series prediction model, predicting the virtual parameter data of the discontinuous position for the parameters of the discontinuous position, and marking the parameters; wherein the method of predicting the virtual parameter data of the discontinuous position includes:

[0034] S t =ε·X t +(1-ε)·(S t-1 +b t-1 );

[0035] b t =∈·(S t -S t-1 )+(1+∈)·b t-1 ;

[0036]

[0037] in, is the virtual parameter data at the discontinuous position t+m, S t Indicates the basic level state of the parameter at time t, b t is the changing trend of the parameter, X t is the actual observation value at time t, ε is the smoothing coefficient, which is used to control the smoothness of the horizontal component, and 0<ε<1, ∈ is the trend smoothing coefficient, which is used to control the smoothness of the trend component, and 0<∈<1, and m is the prediction duration, that is, the number of time steps predicted from the current time t.

[0038] Optionally, the high-speed mode rule is:

[0039] Increase the sampling frequency to twice the current frequency;

[0040] Increase the CPU resource allocation ratio by 2 times;

[0041] Increase cache space by 1.5 times.

[0042] In a second aspect, the present application provides a device multi-parameter statistical calculation system based on a general task scheduling platform, which is used to execute the above-mentioned device multi-parameter statistical calculation method based on the general task scheduling platform, including:

[0043] The device task pattern extraction module is used to analyze the historical task data of each device, extract the device task pattern, build a multi-device task collaboration network, identify collaborative device groups and task conflict nodes, generate task pressure indicators, adjust device task priorities, and generate preliminary scheduling strategies;

[0044] A dynamic scheduling module, coupled to the device task mode extraction module, is used to determine the current load level of each device based on the status of each device after the implementation of the preliminary scheduling strategy as a reference baseline, set a load threshold, read the operating status of nodes that exceed the load threshold, and identify nodes whose load level is below the average load level of all nodes, and dynamically schedule tasks;

[0045] A sampling frequency adjustment module, coupled to the dynamic scheduling module, is used to calculate and mark the change rate of each parameter based on the real-time sampled device data, set the change fluctuation range, and adjust the sampling frequency according to the fluctuation range of the parameter change rate of different devices;

[0046] a virtual data integration module, coupled to the sampling frequency adjustment module, configured to continuously collect parameters of the device according to the adjusted sampling frequency, analyze the completeness of the data collected at each frequency interval for the collected parameters, extract information about devices with parameter discontinuities, and determine the parameter stability of the device based on the current state of the device and the historical parameter change rate and change pattern, predict virtual parameter data at the discontinuity location, and integrate the virtual parameter data with the sampled parameters;

[0047] a task priority adjustment module, coupled to the virtual data integration module, for generating a high-speed mode rule for a device for which no parameters have been collected and the collection time exceeds a set time, and determining whether the device status corresponding to the parameters satisfies the high-speed mode rule; if not, separating and pausing the task based on the task priority;

[0048] The device task mode extraction module includes:

[0049] The task pattern recognition unit is used to identify the task pattern of the equipment, analyze the historical task data of each equipment, identify the task execution frequency and average execution time, and generate the task pattern;

[0050] A collaboration frequency analysis unit, coupled to the task pattern recognition unit, is used to evaluate collaboration frequency, analyze the similarity of parameter types of each device, identify task collaboration relationships between devices, construct a multi-device task collaboration network diagram, and identify the collaboration frequency between each device based on the task collaboration network diagram to evaluate the impact on the overall task completion efficiency;

[0051] a task conflict point marking unit, coupled to the collaboration frequency analysis unit, for analyzing task conflict points between devices, identifying resource contention during task execution, and marking potential task conflict nodes based on historical data and the collaboration network;

[0052] a pressure index calculation unit, coupled to the task conflict point marking unit, for establishing a task pressure prediction model based on the identified task patterns and collaborative relationships, predicting future task pressures, and calculating the pressure index of the device under different task scenarios;

[0053] a task priority division unit, coupled to the pressure index calculation unit, for setting a decision factor affecting the task priority, and prioritizing the tasks executed by the device based on a comparison relationship between the device pressure index and the decision factor;

[0054] a preliminary scheduling strategy generating unit, coupled to the task priority division unit, for generating a preliminary scheduling strategy based on the task pressure index and the task priority;

[0055] The collaboration frequency of the devices is:

[0056]

[0057] Among them, C D (v) is the degree centrality, which indicates the collaboration frequency of task node v, D(v) is the number of edges of task node v, that is, the number of devices collaborating with task node v, and n is the total number of devices;

[0058] The overall task completion efficiency is:

[0059]

[0060] Among them, E is the overall task completion efficiency, w i is the edge weight, i.e. the frequency of collaboration between devices. The larger the weight, the more frequent the collaboration. i The efficiency of the equipment in performing its tasks;

[0061] The pressure indicators of the equipment in different mission scenarios are:

[0062] P i =αF i +βU i +γC i ;

[0063] C i =C D (i)×(n-1);

[0064] Among them, P i is the pressure index of device i in a certain task scenario, F i is the task execution frequency, U i is the resource utilization rate of device i, C i is the collaboration frequency, which indicates the frequency of device i collaborating with other devices in the task collaboration network, and α, β, and γ are weight coefficients.

[0065] Optionally, the dynamic scheduling module includes:

[0066] The equipment operation status collection unit is used to collect the current operation status data of each device after the initial scheduling strategy is implemented, and use the collected status data as a reference baseline;

[0067] a real-time load level calculation unit, coupled to the device operation status acquisition unit, for comparing the collected current operation status data with a reference baseline in real time to calculate the load level of each device;

[0068] a load threshold setting unit, coupled to the real-time load level calculation unit, for setting a load threshold by analyzing historical performance data, and marking the device as being in an overloaded state when the load level of a device exceeds the set threshold;

[0069] an average load calculation unit, coupled to the load threshold setting unit, for calculating the average load of all devices, and extracting all devices with load levels lower than the average load and marking them as task receiving devices;

[0070] The task allocating unit is coupled to the average load calculating unit and is used to allocate the device tasks in the overload state to the task receiving devices and maintain the load of the task receiving devices within a set threshold.

[0071] Optionally, the sampling frequency adjustment module includes:

[0072] A parameter change rate calculation unit is used to compare the collected data and calculate the change rate of each parameter;

[0073] a normal fluctuation interval setting unit, coupled to the parameter change rate calculation unit, for analyzing historical data of the device and setting a normal fluctuation interval for each parameter based on the statistical characteristics of the change rate;

[0074] a parameter change level classification unit, coupled to the normal fluctuation range setting unit, for classifying the parameter change level into three stages according to the parameter change rate, including a high level, a stable level, and a medium level;

[0075] A sampling strategy setting unit is coupled to the parameter change level division unit and is used to set sampling strategies for different levels, wherein: the strategy of increasing the sampling frequency is defined as the strategy of assigning to a high level; the strategy of maintaining the sampling frequency is defined as the strategy of assigning to a stable level; and the strategy of reducing the sampling frequency is defined as the strategy of assigning to a medium level.

[0076] Optionally, the virtual data integration module includes:

[0077] The device status monitoring unit is used to continuously collect device parameters for the adjusted sampling frequency, check whether the data collected in each sampling frequency interval is complete, and compare whether the data coverage of different frequency intervals is complete;

[0078] a data interruption analysis unit, coupled to the device status monitoring unit, for extracting information about devices with data interruptions, recording the locations and frequencies of the interruptions, analyzing the impact of the interruptions on device operating status monitoring, and generating an interruption severity level;

[0079] a parameter fluctuation interval evaluation unit, coupled to the data discontinuity analysis unit, for evaluating whether the current parameter is within the corresponding normal fluctuation interval and performing abnormal pattern detection on the parameter fluctuation;

[0080] A virtual data generating unit is coupled to the parameter fluctuation interval evaluating unit and is used to establish a time series prediction model, predict virtual parameter data of the discontinuous position for the parameters, and mark them;

[0081] The state monitoring update unit is coupled to the virtual data generation unit and is used to align and integrate the collected parameter data with the predicted virtual data, and update the integrated complete data stream to the device state monitoring unit.

[0082] Optionally, the task priority adjustment module includes:

[0083] The unit for calculating the duration of non-data collection is used to set a collection duration quota. For devices whose parameters have not been collected, the unit calculates the duration of the non-data collection, compares whether the collection duration quota has been exceeded, and reads the current operating status and pressure index of the device that has exceeded the quota.

[0084] The high-speed mode intervention demand analysis unit is coupled to the non-data collection time calculation unit and is used to analyze whether high-speed mode intervention is needed based on the current operating status and pressure indicators, and to determine whether the equipment meets the high-speed mode rule requirements.

[0085] Compared with the prior art, the present invention provides at least the following beneficial effects:

[0086] First, through in-depth analysis of device task patterns and comprehensive assessment of collaborative relationships, scheduling strategies can factor in multiple factors and improve overall task execution efficiency. By identifying task patterns, we can predict device task loads and collaboration requirements in advance, reducing the likelihood of task conflicts. Furthermore, analysis of collaborative networks enables more refined task prioritization, ensuring that critical tasks are executed under optimal conditions.

[0087] Second, the system significantly improves the efficiency and stability of collaborative work among multiple devices through intelligent task scheduling and efficient resource management. By accurately analyzing and identifying device task patterns and collaborative networks, it optimizes resource allocation, reduces task conflicts and device overload, and enables dynamic task scheduling and priority adjustment. This optimization mechanism ensures excellent performance in device load balancing, improves overall operational efficiency, and reduces resource waste.

[0088] Third, through real-time analysis of the parameter change rate and timely adjustment of the sampling frequency, the system not only ensures the timeliness and accuracy of key data, but also optimizes the overall sampling efficiency and effectively reduces unnecessary data processing overhead; at the same time, using the time series prediction model, the system performs virtual data prediction on the parameters at discontinuous positions to ensure that even in the case of data interruptions, the monitoring of the equipment operating status is still unaffected, enhancing the continuity and accuracy of data collection and avoiding monitoring blind spots caused by data missing; through the completion and integration of virtual data, the system can provide more comprehensive and reliable equipment status data support, improving the scientificity and accuracy of equipment management and decision-making; overall, this mechanism not only improves the refinement of equipment status monitoring, but also enhances the system's ability to respond to data anomalies and equipment status fluctuations, ensuring efficient and stable operation of equipment in complex environments.

[0089] Of course, any product implementing the present invention does not necessarily need to achieve all of the technical effects described above at the same time.

[0090] Further features and advantages of the present invention will become apparent from the following detailed description of exemplary embodiments of the present invention with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention.

[0092] Figure 1 This is a structural block diagram of a device multi-parameter statistical calculation system based on a general task scheduling platform provided by an embodiment of the present invention;

[0093] Figure 2 This is a structural block diagram of a device task mode extraction module provided by an embodiment of the present invention;

[0094] Figure 3 is a structural block diagram of a dynamic scheduling module provided by an embodiment of the present invention;

[0095] Figure 4 is a structural block diagram of a sampling frequency adjustment module provided by an embodiment of the present invention;

[0096] Figure 5 is a structural block diagram of a virtual data integration module provided by an embodiment of the present invention;

[0097] Figure 6 is a structural block diagram of a task priority adjustment module provided by an embodiment of the present invention;

[0098] Figure 7 It is a flow chart of a device multi-parameter statistical calculation method based on a general task scheduling platform provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0099] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that unless otherwise specifically stated, the relative arrangement of components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present invention.

[0100] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.

[0101] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.

[0102] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0103] It should be noted that like reference numerals and letters refer to like items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0104] Example 1

[0105] Reference Figures 1-6 As shown, Figure 1 This is a structural block diagram of a device multi-parameter statistical calculation system based on a general task scheduling platform provided by an embodiment of the present invention; Figure 2 This is a structural block diagram of a device task mode extraction module provided by an embodiment of the present invention; Figure 3 is a structural block diagram of a dynamic scheduling module provided by an embodiment of the present invention; Figure 4 is a structural block diagram of a sampling frequency adjustment module provided by an embodiment of the present invention; Figure 5 is a structural block diagram of a virtual data integration module provided by an embodiment of the present invention; Figure 6 This is a structural block diagram of a task priority adjustment module provided by an embodiment of the present invention.

[0106] like Figure 1 As shown, the device multi-parameter statistical calculation system based on the general task scheduling platform includes: a device task pattern extraction module 100, a dynamic scheduling module 200 coupled to the device task pattern extraction module 100, a sampling frequency adjustment module 300 coupled to the dynamic scheduling module 200, a virtual data integration module 400 coupled to the sampling frequency adjustment module 300, and a task priority adjustment module 500 coupled to the virtual data integration module 400.

[0107] The device task pattern extraction module 100 is used to analyze the historical task data of each device, extract the device task pattern, build a multi-device task collaboration network, identify collaborative device groups and task conflict nodes, generate task pressure indicators, adjust device task priorities, and generate a preliminary scheduling strategy.

[0108] like Figure 2 As shown, the device task mode extraction module 100 specifically includes:

[0109] The task pattern recognition unit 110 is used for device task pattern recognition and analysis of historical task data of each device, identifying task execution frequency and average execution time, and generating task patterns, including high-frequency task patterns and periodic task patterns;

[0110] The collaboration frequency analysis unit 120 is coupled to the task pattern recognition unit 110 and is used to evaluate the collaboration frequency, analyze the similarity of parameter types of each device, identify the task collaboration relationship between devices, construct a multi-device task collaboration network diagram, and identify the collaboration frequency between each device based on the task collaboration network diagram to evaluate the impact on the overall task completion efficiency;

[0111] The task conflict point marking unit 130 is coupled to the collaboration frequency analysis unit 120 and is used to analyze task conflict points between devices, identify resource contention during task execution, and mark potential task conflict nodes based on historical data and the collaboration network;

[0112] The pressure index calculation unit 140 is coupled to the task conflict point marking unit 130 and is used to establish a task pressure prediction model based on the identified task patterns and collaborative relationships, predict future task pressures, and calculate the pressure index of the device under different task scenarios;

[0113] The task priority division unit 150 is coupled to the pressure index calculation unit 140 and is used to set a decision factor that affects the task priority and prioritize the tasks executed by the device based on the comparison relationship between the device pressure index and the decision factor. The specific rule for prioritizing tasks based on the comparison relationship between the pressure index and the decision factor is to perform the priority division by setting a threshold range of the pressure index, as follows:

[0114] If the equipment pressure index P i If the value is higher than the maximum value of the threshold range and the task urgency is "high", the task priority is raised to the highest level and is scheduled first.

[0115] If P i If the resource demand is low and the resource is within the threshold range, the default priority is maintained.

[0116] If P iIf the task is below the minimum value of the threshold range and is non-critical, its priority will be lowered or its processing will be delayed.

[0117] The preliminary scheduling strategy generating unit 160 is coupled to the task priority dividing unit 150 and is used to generate a preliminary scheduling strategy based on the task pressure index and the task priority. The generation rule of the preliminary scheduling strategy is:

[0118] 1. Task priority: High-priority tasks are assigned to low-load devices first to ensure timely execution of critical tasks.

[0119] 2. Load balancing: Avoid allocating high-frequency, high-resource consumption tasks to the same device, and use the pressure index P i Balance the device load.

[0120] 3. Collaboration optimization: Frequent collaborative tasks should be assigned to the same device group or adjacent devices as much as possible to reduce cross-device communication overhead.

[0121] Based on the task collaboration network diagram, the frequency of collaboration between devices is identified and the impact on the overall task completion efficiency is evaluated. Specifically:

[0122] Calculate the collaboration frequency of each device as:

[0123]

[0124] Among them, C D (v) is the degree centrality, which indicates the frequency of collaboration of task node v. D(v) is the number of edges of task node v, that is, the number of devices collaborating with task node v. n is the total number of devices. Among them, node v represents the device. C D The higher the (v), the more active the device is in task collaboration and may become a scheduling bottleneck, requiring priority consideration for load balancing.

[0125] Evaluate the impact on overall task completion efficiency:

[0126]

[0127] Among them, E is the overall task completion efficiency, w i is the edge weight, i.e. the frequency of collaboration between devices. The larger the weight, the more frequent the collaboration. i The efficiency of the equipment in performing its tasks;

[0128] The overall task completion efficiency E can be used to quantify the impact of inter-device collaboration on the overall task execution speed.

[0129] For example, a group of devices that collaborate frequently and have high execution efficiency will increase E, and vice versa. If E is lower than a reasonable value, the system can adjust the collaboration strategy, such as reducing the collaborative tasks of inefficient devices or reallocating tasks to improve overall efficiency.

[0130] Generating task pressure indicators includes: establishing a task pressure prediction model based on the identified task patterns and collaborative relationships, predicting future task pressures, and calculating the pressure indicators of the equipment in different task scenarios;

[0131] The stress indicators of computing equipment in different task scenarios are as follows:

[0132] P i =αF i +βU i +γC i ;

[0133] C i =C D (i)×(n-1);

[0134] Among them, P i is the pressure index of device i in a certain task scenario, F i is the task execution frequency, U i is the resource utilization rate of device i, C i is the collaboration frequency, which indicates the frequency of device i collaborating with other devices in the task collaboration network, and α, β, and γ are weight coefficients.

[0135] The dynamic scheduling module 200 is used to determine the current load level of each device based on the status of each device after the implementation of the preliminary scheduling strategy as a reference baseline, set a load threshold, read the operating status of nodes that exceed the load threshold, and identify nodes whose load level is below the average load level of all nodes, and dynamically schedule tasks; wherein, the status of each device is evaluated through real-time monitoring data such as CPU usage, memory occupancy, task queue length, network bandwidth utilization, etc., and the load level is evaluated according to the following formula: Load level = (current resource usage / resource capacity) × 100%.

[0136] like Figure 3 As shown, the dynamic scheduling module 200 specifically includes:

[0137] The device operation status collection unit 210 is used to collect the current operation status data of each device after implementing the preliminary scheduling strategy, and use the collected status data as a reference baseline;

[0138] The real-time load level calculation unit 220 is coupled to the device operation status collection unit 210 and is used to compare the collected current operation status data with the reference baseline in real time to calculate the load level of each device;

[0139] The load threshold setting unit 230 is coupled to the real-time load level calculation unit 220 and is configured to analyze historical performance data to set a load threshold. When the load level of a device exceeds the set threshold, the device is marked as being in an overloaded state.

[0140] The average load calculation unit 240 is coupled to the load threshold setting unit 230 and is used to calculate the average load of all devices and extract all devices with load levels lower than the average load and mark them as task receiving devices;

[0141] The task allocating unit 250 is coupled to the average load calculating unit 240 and is used to allocate the tasks of the devices in the overload state to the task receiving devices and maintain the load of the task receiving devices within a set threshold.

[0142] The sampling frequency adjustment module 300 is used to calculate and mark the change rate of each parameter based on the real-time sampled device data, set the change fluctuation range, and adjust the sampling frequency according to the fluctuation range of the parameter change rate of different devices.

[0143] like Figure 4 As shown, the sampling frequency adjustment module 300 specifically includes:

[0144] The parameter change rate calculation unit 310 is used to compare the collected data and calculate the change rate of each parameter. The collected data mainly includes real-time operating parameters of the equipment, such as temperature, pressure, flow values ​​collected by sensors, or internal state parameters of the equipment, which are used to analyze the equipment status at the parameter level.

[0145] The normal fluctuation interval setting unit 320 is coupled to the parameter change rate calculation unit 310 and is used to analyze the historical data of the device and set the normal fluctuation interval of each parameter based on the statistical characteristics of the change rate;

[0146] The parameter change level classification unit 330 is coupled to the normal fluctuation range setting unit 320 and is used to classify the parameter change level into three stages, including a high level, a stable level, and a medium level, according to the parameter change rate;

[0147] The sampling strategy setting unit 340 is coupled to the parameter change level division unit 330 and is used to set sampling strategies for different levels, wherein: the strategy of increasing the sampling frequency is defined as the strategy of assigning to a high level; the strategy of maintaining the sampling frequency is defined as the strategy of assigning to a stable level; and the strategy of reducing the sampling frequency is defined as the strategy of assigning to a medium level.

[0148] The virtual data integration module 400 is used to continuously collect parameters of the device based on the adjusted sampling frequency. For the collected parameters, it analyzes whether the data collected at each frequency interval is complete, extracts the device information with parameter discontinuities, and combines the current status of the device and the historical parameter change rate and change pattern to determine the parameter stability of the device, predict the virtual parameter data at the discontinuity position, and integrate it with the sampling parameters.

[0149] like Figure 5 As shown, the virtual data integration module 400 specifically includes:

[0150] The device status monitoring unit 410 is configured to continuously collect device parameters at the adjusted sampling frequency, check whether the data collected within each sampling frequency interval is complete, and compare whether the data coverage of different frequency intervals is complete;

[0151] The data interruption analysis unit 420 is coupled to the device status monitoring unit 410 and is used to extract information about devices with data interruptions, record the location and frequency of the interruptions, analyze the impact of the interruptions on device operating status monitoring, and generate an interruption severity level;

[0152] The parameter fluctuation interval evaluation unit 430 is coupled to the data discontinuity analysis unit 420 and is used to evaluate whether the current parameter is within the corresponding normal fluctuation interval and perform abnormal pattern detection on the parameter fluctuation;

[0153] The virtual data generating unit 440 is coupled to the parameter fluctuation interval evaluating unit 430 and is used to establish a time series prediction model, predict virtual parameter data at the discontinuous position, and mark the parameters;

[0154] The state monitoring update unit 450 is coupled to the virtual data generation unit 440 and is configured to align and integrate the collected parameter data with the predicted virtual data, and update the integrated complete data stream to the device state monitoring unit.

[0155] The task priority adjustment module 500 is used to generate high-speed mode rules for devices that have not collected parameters and whose collection time exceeds the set time, and to determine whether the device status corresponding to the parameters can meet the high-speed mode rules. If not, the task is separated and suspended based on the task priority.

[0156] like Figure 6 As shown, the task priority adjustment module 500 specifically includes:

[0157] The data collection time calculation unit 510 is used to set a collection time quota. For devices that have not collected parameters, the unit calculates the time during which the device has not collected parameters, compares whether the collection time quota has been exceeded, and reads the current operating status and pressure index of the device that has exceeded the collection time quota.

[0158] The high-speed mode intervention demand analysis unit 520 is coupled to the non-data collection time calculation unit 510, and is used to analyze whether high-speed mode intervention is needed based on the current operating status and pressure indicators, and to determine whether the equipment meets the high-speed mode rule requirements.

[0159] Example 2

[0160] Reference Figure 7 As shown, Figure 7 1 is a flow chart of a device multi-parameter statistical calculation method based on a general task scheduling platform provided by an embodiment of the present invention, including:

[0161] S1. Generate preliminary scheduling strategy: Analyze the historical task data of each device, extract the device task mode, build a multi-device task collaboration network, identify the collaborative device groups and task conflict nodes, generate task pressure indicators, adjust the device task priority, and generate a preliminary scheduling strategy.

[0162] Specifically, extracting the device task mode in this step includes: analyzing the historical task data of each device, identifying the task execution frequency and average execution time, and generating a task mode, which includes a high-frequency task mode and a periodic task mode;

[0163] Building a multi-device task collaboration network includes: analyzing the similarity of parameter types of each device, identifying the task collaboration relationship between devices, building a multi-device task collaboration network diagram, and identifying the collaboration frequency between each device based on the task collaboration network diagram, and evaluating the impact on the overall task completion efficiency; Among them,

[0164] Based on the task collaboration network diagram, the frequency of collaboration between devices is identified and the impact on the overall task completion efficiency is evaluated. Specifically:

[0165] Calculate the collaboration frequency of each device as:

[0166]

[0167] Among them, C D (v) is the degree centrality, which indicates the frequency of collaboration of task node v. D(v) is the number of edges of task node v, that is, the number of devices collaborating with task node v. n is the total number of devices. Among them, node v represents the device. C D The higher the (v), the more active the device is in task collaboration and may become a scheduling bottleneck, requiring priority consideration for load balancing.

[0168] Evaluate the impact on overall task completion efficiency:

[0169]

[0170] Among them, E is the overall task completion efficiency, w i is the edge weight, i.e. the frequency of collaboration between devices. The larger the weight, the more frequent the collaboration. i The efficiency of the equipment in performing its tasks;

[0171] The overall task completion efficiency E can be used to quantify the impact of inter-device collaboration on the overall task execution speed.

[0172] For example, a group of devices that collaborate frequently and have high execution efficiency will increase E, and vice versa. If E is lower than a reasonable value, the system can adjust the collaboration strategy, such as reducing the collaborative tasks of inefficient devices or reallocating tasks to improve overall efficiency.

[0173] Generating task pressure indicators includes: establishing a task pressure prediction model based on the identified task patterns and collaborative relationships, predicting future task pressures, and calculating the pressure indicators of the equipment in different task scenarios;

[0174] The stress indicators of computing equipment in different task scenarios are as follows:

[0175] P i =αF i +βU i +γC i ;

[0176] C i =C D (i)×(n-1);

[0177] Among them, P i is the pressure index of device i in a certain task scenario, F i is the task execution frequency, U i is the resource utilization rate of device i, C i is the collaboration frequency, which indicates the frequency of device i collaborating with other devices in the task collaboration network, and α, β, and γ are weight coefficients.

[0178] By analyzing each device's historical task data, we identify the frequency and average execution time of task execution, thereby generating task patterns that include both high-frequency and periodic task patterns. This pattern recognition helps us understand the task behavior of each device in different time periods, providing basic data support for optimizing scheduling strategies.

[0179] The frequency of collaboration between devices is then evaluated. By analyzing the parameter similarities and task collaboration relationships between devices, a multi-device task collaboration network diagram is constructed. Based on this, the frequency of collaboration between devices is identified and its impact on overall task completion efficiency is evaluated. By calculating the collaboration frequency of each device, the degree centrality metric can be used to measure the collaboration frequency of task nodes. This allows the importance of devices in the task collaboration network to be determined, and the impact of task completion efficiency between devices on overall efficiency can be further evaluated.

[0180] Through in-depth analysis of device task patterns and comprehensive assessment of collaborative relationships, scheduling strategies can factor in multiple factors and improve overall task execution efficiency. Identifying task patterns allows for early prediction of device task loads and collaboration requirements, reducing the likelihood of task conflicts. Furthermore, analysis of collaborative networks enables more refined task prioritization, ensuring that critical tasks are executed under optimal conditions.

[0181] S2. Dynamic Scheduling: Based on the status of each device after the implementation of the preliminary scheduling strategy as a reference baseline, determine the current load level of each device, set the load threshold, read the operating status of nodes that exceed the load threshold, identify nodes whose load level is below the average load level of all nodes, and dynamically schedule tasks.

[0182] The primary goal of this step is to ensure efficient system operation while avoiding equipment overload by monitoring and analyzing equipment operating status in real time. By comparing the collected current equipment status with a reference baseline in real time, the system can accurately calculate the load level of each device and identify potential load issues that may arise during actual operation.

[0183] The system analyzes device load characteristics based on historical performance data and sets appropriate load thresholds. When a device's load exceeds a pre-set threshold, the system marks it as overloaded and immediately initiates countermeasures. This real-time monitoring and rapid response mechanism enables the system to effectively manage device load and prevent performance issues caused by overload.

[0184] To optimize task distribution, the system calculates the average load of all devices and identifies those with lower-than-average loads, marking them as task recipients. This allows the system to identify nodes across the network that can handle more tasks and distribute tasks from overloaded devices among them. This process ensures that the load on task-receiving devices never exceeds a set threshold, maintaining overall system balance.

[0185] Dynamically adjusting device loads enables efficient resource utilization and task scheduling, improving overall system efficiency. Intelligent task load distribution not only reduces the risk of overloading individual devices but also extends the overall equipment lifespan. Furthermore, real-time monitoring and adjustments enable rapid response to load changes, reducing delays and task interruptions, thereby improving user experience and service quality.

[0186] S3. Adjust the sampling frequency: Based on the real-time device data, calculate and mark the change rate of each parameter, set the change fluctuation range, and adjust the sampling frequency according to the fluctuation range of the parameter change rate of different devices.

[0187] In this step, the dynamic change trend of each parameter is determined by calculating the rate of change of the parameter at adjacent time points. This trend is then used to determine the normal fluctuation range for each parameter. Based on the statistical characteristics of the rate of change, parameter changes are classified into three levels: high, stable, and medium. A high level indicates a significant parameter change, potentially impacting the device's status; a stable level indicates a small parameter change, resulting in a relatively stable device status; and a medium level falls somewhere in between.

[0188] At different levels of change, corresponding sampling strategies have been developed for each level. For parameters with high-level changes, the sampling frequency strategy is increased to ensure more frequent acquisition of change data for these key parameters, thereby promptly responding to possible state changes. For parameters with stable levels, the existing sampling frequency is maintained to ensure the stability and reliability of data collection. For parameters with medium levels, especially those with minimal changes, the sampling frequency is appropriately reduced to save resources and improve overall sampling efficiency.

[0189] By accurately analyzing parameter change rates and flexibly adjusting sampling strategies, the efficiency and accuracy of data collection can be effectively improved. High-frequency sampling ensures timely capture of key parameter changes, reducing the risk of instability caused by drastic parameter fluctuations. At the same time, appropriately reducing the sampling frequency of insensitive parameters effectively reduces system resource consumption and optimizes overall data processing performance. This dynamic adjustment mechanism not only improves the precision of data recording but also enhances the system's adaptability to environmental changes, ensuring that the system maintains efficient and stable operation under changing operating conditions.

[0190] In this step, the rate of change of each parameter is calculated, specifically:

[0191]

[0192] Among them, R is the rate of change, which represents the approximate value of the derivative involved in the change, V t+1 and V t-1They represent the parameter values ​​collected at the time point and are the time interval between two adjacent time points.

[0193] Adjusting the sampling frequency includes: dividing the parameters into three stages of change levels according to the change rate of the parameters, and setting sampling strategies for different levels. The change levels include high level, stable level and medium level, among which, the definition of assigning to the high level is to increase the sampling frequency strategy; the definition of assigning to the stable level is to maintain the sampling frequency strategy; the definition of assigning to the medium level is to reduce the sampling frequency strategy.

[0194] S4. Virtual data integration: Continuously collect parameters of the device based on the adjusted sampling frequency. For the collected parameters, analyze whether the data collected at each frequency interval is complete, extract device information with parameter discontinuities, and combine the current status of the device and the historical parameter change rate and change pattern to determine the parameter stability of the device, predict the virtual parameter data at the discontinuity position, and integrate it with the sampling parameters.

[0195] In this step, we first identify devices experiencing data interruptions and record their location and frequency. By analyzing the impact of these interruptions on device health monitoring, we generate an interruption severity rating, which provides an important basis for subsequent processing. The severity rating takes into account the impact of the interruption on the device's key performance indicators and calculates an interruption impact index to measure the overall impact of the interruption on device performance.

[0196] The system then evaluates whether the current parameters are within the normal fluctuation range and detects abnormal patterns. This ensures that the equipment status can be accurately understood and assessed even during periods of discontinuity. For parameters at discontinuous locations, a time series forecasting model is used to predict virtual parameter data for these locations. The figure illustrates the specific calculation method of the forecasting model. By adjusting the baseline level and trend of the parameters, the system can generate virtual data for the corresponding time period.

[0197] Ultimately, the collected data is aligned and integrated with the predicted virtual data to form a complete data stream. This integrated data stream is then updated to the equipment status monitoring unit, ensuring the continuity and integrity of equipment status monitoring and preventing data interruptions from affecting the overall monitoring effect.

[0198] By effectively addressing data gaps, the accuracy of equipment operating status monitoring is significantly improved. Virtual parameter prediction at intermittent locations enables high-quality equipment status assessments even in the absence of data. This not only improves data monitoring continuity but also enhances robustness in the presence of data anomalies. Furthermore, the time series prediction model employed ensures the rationality and reliability of the virtual data, providing solid data support for subsequent equipment management and maintenance.

[0199] Specifically, determining the parameter stability of the device includes: extracting device information with data interruptions, recording the location and frequency of the interruptions, analyzing the impact of the interruptions on the monitoring of the device's operating status, and generating an interruption severity level; evaluating whether the current parameters are within the corresponding normal fluctuation range, and detecting abnormal patterns in parameter fluctuations;

[0200] The interruption impact index in analyzing the impact of interruptions on equipment operation status monitoring is:

[0201]

[0202] Among them, I is the interruption impact index, which represents the comprehensive impact of interruptions on equipment operation status monitoring, ΔKPI k is the rate of change of the kth performance indicator, ω k is the weight of the kth performance indicator, δ is the weight factor of the interruption duration, D is the interruption duration, n is the total number of all performance indicators, KPI k，after is the kth value measured after the interruption occurs, KPI k，before is the kth value measured before the discontinuity occurs;

[0203] The intermittency impact index I is a comprehensive weighted value. The larger I is, the more significant the interference of data intermittency on equipment status monitoring is, and it needs to be handled with priority (such as increasing the sampling frequency or triggering task reallocation).

[0204] Predicting the virtual parameter data of the discontinuous position includes: establishing a time series prediction model, predicting the virtual parameter data of the discontinuous position for the parameters of the discontinuous position, and marking them; specifically:

[0205] S t =ε·X t +(1-ε)·(S t-1 +b t-1 );

[0206] b t =∈·(S t -S t-1 )+(1+∈)·b t-1 ;

[0207]

[0208] in, is the virtual parameter data at the discontinuous position t+m, S t Indicates the basic level state of the parameter at time t, b t is the changing trend of the parameter, X tis the actual observation value at time t, ε is the smoothing coefficient, which is used to control the smoothness of the horizontal component, and 0<ε<1, ∈ is the trend smoothing coefficient, which is used to control the smoothness of the trend component, and 0<∈<1, and m is the prediction duration, that is, the number of time steps predicted from the current time t.

[0209] S5. Adjust task priority: For devices that have not collected parameters and whose collection time exceeds the set time, generate high-speed mode rules to determine whether the device status corresponding to the parameters can meet the high-speed mode rules. If the device status does not meet the requirements of the high-speed mode rules, repeat dynamic scheduling to allocate tasks; if it meets the requirements of the high-speed mode rules, evaluate the priority of the task according to the high-speed mode rules, identify tasks that need to be prioritized and those that can be suspended, and adjust the task execution order.

[0210] This module sets a collection duration quota, which serves as a time limit for device parameter collection. If a device fails to collect parameters within the specified timeframe, the module calculates the duration of the missed period and checks whether this exceeds the pre-set duration quota. For devices that exceed this time limit, the module further analyzes their current operating status and stress indicators. These indicators help assess whether the device is under high load and whether its task execution strategy needs to be adjusted.

[0211] Next, based on the device's current operating status and stress indicators, it analyzes whether high-speed mode intervention is necessary. High-speed mode is a special operating mode suitable for device states that require rapid response. If a device fails to meet the high-speed mode rules, the dynamic scheduling module is invoked to reallocate tasks to alleviate the device load and ensure stable system operation.

[0212] If the device meets the high-speed mode requirements, it then evaluates the task priorities according to the high-speed mode rules. This step aims to identify tasks that require priority and those that can be temporarily suspended, allowing for the proper adjustment of the task execution order. By optimizing task priorities, critical tasks can be executed promptly in high-speed mode while avoiding unnecessary resource waste.

[0213] By comprehensively evaluating and adjusting the scheduling of devices for which parameters have not been collected, we can effectively improve the flexibility and responsiveness of device operations. The introduction of high-speed mode rules enables the system to quickly adapt to changes in special circumstances, ensuring the priority execution of critical tasks and improving the efficiency of overall task scheduling. Furthermore, by dynamically adjusting task priorities and execution order, we can reduce the risks associated with uneven device loads and improve system reliability and stability.

[0214] Among them, the high-speed mode rules include:

[0215] 1. Hardware resource compliance: The device's current CPU resources, cache space, and sampling frequency can be adjusted to meet high-speed mode requirements. Specific indicators are set based on the device type, model, and operating environment (the current CPU idle rate is high enough to support a 2-fold increase in the allocation ratio).

[0216] 2. Task priority matching: The task type currently executed by the device is within the high-speed mode processing range and there is no conflict with higher priority tasks.

[0217] If the device status does not meet the high-speed mode rule requirements, the dynamic scheduling module will be repeated to allocate tasks;

[0218] If the high-speed mode rule requirements are met, the priorities of the tasks are evaluated according to the high-speed mode rules to identify the tasks that need to be processed first and that can be suspended, and the task execution order is adjusted.

[0219] The high-speed mode rule requirements are as follows:

[0220] Increase the sampling frequency to twice the current frequency;

[0221] Increase the CPU resource allocation ratio by 2 times;

[0222] Increase cache space by 1.5 times.

[0223] It can be seen from the above embodiments that the present invention provides at least the following beneficial effects:

[0224] 1. The system significantly improves the efficiency and stability of collaborative work among multiple devices through intelligent task scheduling and efficient resource management. It optimizes resource allocation, reduces task conflicts and device overloads, and implements dynamic task scheduling and priority adjustment by accurately analyzing and identifying device task patterns and collaborative networks. This optimization mechanism ensures excellent performance in device load balancing, improves overall operational efficiency, and reduces resource waste.

[0225] 2. Through real-time analysis of parameter change rates and timely adjustment of sampling frequency, the system not only ensures the timeliness and accuracy of key data, but also optimizes the overall sampling efficiency and effectively reduces unnecessary data processing overhead; at the same time, using the time series prediction model, the system performs virtual data prediction on the parameters at discontinuous positions to ensure that even in the case of data interruptions, the monitoring of equipment operating status remains unaffected, thereby enhancing the continuity and accuracy of data collection and avoiding monitoring blind spots caused by data missing; through the completion and integration of virtual data, the system can provide more comprehensive and reliable equipment status data support, improving the scientificity and accuracy of equipment management and decision-making; overall, this mechanism not only improves the refinement of equipment status monitoring, but also enhances the system's ability to respond to data anomalies and equipment status fluctuations, ensuring efficient and stable operation of equipment in complex environments.

[0226] Although some specific embodiments of the present invention have been described in detail by way of examples, it should be understood by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present invention. It should be understood by those skilled in the art that modifications may be made to the above embodiments without departing from the scope and spirit of the present invention. The scope of the present invention is defined by the appended claims.

Claims

1. A device multi-parameter statistical calculation method based on a general task scheduling platform, characterized in that: include: Generate preliminary scheduling strategy: Analyze historical task data of each device, extract device task patterns, build a multi-device task collaboration network, identify collaborative device groups and task conflict nodes, generate task pressure indicators, adjust device task priorities, and generate preliminary scheduling strategy; Dynamic scheduling: Based on the status of each device after the implementation of the preliminary scheduling strategy as a reference baseline, the current load level of each device is determined, a load threshold is set, the operating status of nodes exceeding the load threshold is read, and nodes with load levels below the average load level of all nodes are identified, and tasks are dynamically scheduled; Adjust the sampling frequency: Based on the real-time device data, calculate and mark the change rate of each parameter, set the change fluctuation range, and adjust the sampling frequency according to the fluctuation range of the parameter change rate of different devices; Virtual data integration: Continuously collect parameters of the device based on the adjusted sampling frequency. For the collected parameters, analyze whether the data collected at each frequency interval is complete, extract information about devices with parameter discontinuities, and combine the current status of the device with the historical parameter change rate and change pattern to determine the parameter stability of the device. Predict virtual parameter data at the discontinuity location and integrate it with the sampled parameters. Adjust task priority: For devices whose parameters are not collected and whose collection time exceeds the set time, a high-speed mode rule is generated to determine whether the device status corresponding to the parameters meets the high-speed mode rule. If the device status does not meet the high-speed mode rule requirements, dynamic scheduling is repeated to assign tasks. If the high-speed mode rule requirements are met, the task priority is evaluated according to the high-speed mode rule, and tasks that need to be prioritized and can be suspended are identified, and the task execution order is adjusted. The extracting device task mode includes: analyzing historical task data of each device, identifying task execution frequency and average execution time, and generating a task mode, wherein the task mode includes a high-frequency task mode and a periodic task mode; The construction of a multi-device task collaboration network includes: analyzing the similarity of parameter types of each device, identifying the task collaboration relationship between devices, constructing a multi-device task collaboration network diagram, and identifying the collaboration frequency between each device based on the task collaboration network diagram, and evaluating the impact on the overall task completion efficiency; wherein, The collaboration frequency of the devices is: Among them, C D (v) is the degree centrality, which indicates the collaboration frequency of task node v, D(v) is the number of edges of task node v, that is, the number of devices collaborating with task node v, and n is the total number of devices; The overall task completion efficiency is: Among them, E is the overall task completion efficiency, w i is the edge weight, i.e. the frequency of collaboration between devices. The larger the weight, the more frequent the collaboration. i The efficiency of the equipment in performing its tasks; Generating the task pressure index includes: establishing a task pressure prediction model based on the identified task mode and collaborative relationship, predicting future task pressure, and calculating the pressure index of the device in different task scenarios; wherein, The pressure indicators of the equipment in different mission scenarios are: P i =αF i +βU i +γC i ; C i =C D (i)×(n-1); Among them, P i is the pressure index of device i in a certain task scenario, F i is the task execution frequency, U i is the resource utilization rate of device i, C i is the collaboration frequency, which indicates the frequency of device i collaborating with other devices in the task collaboration network, and α, β, and γ are weight coefficients.

2. The device multi-parameter statistical calculation method based on the general task scheduling platform according to claim 1 is characterized in that: The rate of change of each parameter is: Among them, R is the rate of change, which represents the approximate value of the derivative involved in the change, V t+1 and V t-1 Respectively represent the parameter values ​​collected at the time point and , which is the time interval between two adjacent time points; The adjustment of the sampling frequency includes: dividing the parameters into three stages of change levels according to the change rate of the parameters, and setting sampling strategies for different levels. The change levels include high level, stable level and medium level, among which, the definition of increasing the sampling frequency strategy is assigned to the high level; the definition of maintaining the sampling frequency strategy is assigned to the stable level; and the definition of reducing the sampling frequency strategy is assigned to the medium level.

3. The device multi-parameter statistical calculation method based on the general task scheduling platform according to claim 1 is characterized in that: The determination of the parameter stability of the device includes: extracting device information with data discontinuities, recording the location and frequency of the discontinuities, analyzing the impact of the discontinuities on the monitoring of the device's operating status, and generating a discontinuity severity level; evaluating whether the current parameters are within the corresponding normal fluctuation range, and detecting abnormal patterns in the parameter fluctuations; wherein, The interruption impact index in the analysis of the impact of interruptions on equipment operation status monitoring is: Among them, I is the interruption impact index, which represents the comprehensive impact of interruptions on equipment operation status monitoring, ΔKPI k is the rate of change of the kth performance indicator, ω k is the weight of the kth performance indicator, δ is the weight factor of the interruption duration, D is the interruption duration, n is the total number of all performance indicators, KPI k，after is the kth value measured after the interruption occurs, KPI k，before is the kth value measured before the discontinuity occurs; The method of predicting the virtual parameter data of the discontinuous position includes: establishing a time series prediction model, predicting the virtual parameter data of the discontinuous position for the parameters of the discontinuous position, and marking the parameters; wherein the method of predicting the virtual parameter data of the discontinuous position includes: S t =e·X t +(1-ε)·(S t-1 +b t-1 ); b t =∈·(S t -S t-1 )+(1+∈)·b t-1 ; in, is the virtual parameter data at the discontinuous position t+m, S t Indicates the basic level state of the parameter at time t, b t is the changing trend of the parameter, X t is the actual observation value at time t, ε is the smoothing coefficient, which is used to control the smoothness of the horizontal component, and 0<ε<1, ∈ is the trend smoothing coefficient, which is used to control the smoothness of the trend component, and 0<∈<1, and m is the prediction duration, that is, the number of time steps predicted from the current time t.

4. The device multi-parameter statistical calculation method based on the general task scheduling platform according to claim 1 is characterized in that: The high-speed mode rules are: Increase the sampling frequency to twice the current frequency; Increase the CPU resource allocation ratio by 2 times; Increase cache space by 1.5 times.

5. A multi-parameter statistical calculation system for equipment based on a general task scheduling platform, characterized in that: The method for performing multi-parameter statistical calculation of devices based on a general task scheduling platform according to any one of claims 1 to 4 comprises: The device task pattern extraction module is used to analyze the historical task data of each device, extract the device task pattern, build a multi-device task collaboration network, identify collaborative device groups and task conflict nodes, generate task pressure indicators, adjust device task priorities, and generate preliminary scheduling strategies; A dynamic scheduling module, coupled to the device task mode extraction module, is used to determine the current load level of each device based on the status of each device after the implementation of the preliminary scheduling strategy as a reference baseline, set a load threshold, read the operating status of nodes that exceed the load threshold, and identify nodes whose load level is below the average load level of all nodes, and dynamically schedule tasks; A sampling frequency adjustment module, coupled to the dynamic scheduling module, is used to calculate and mark the change rate of each parameter based on the real-time sampled device data, set the change fluctuation range, and adjust the sampling frequency according to the fluctuation range of the parameter change rate of different devices; a virtual data integration module, coupled to the sampling frequency adjustment module, configured to continuously collect parameters of the device according to the adjusted sampling frequency, analyze the completeness of the data collected at each frequency interval for the collected parameters, extract information about devices with parameter discontinuities, and determine the parameter stability of the device based on the current state of the device and the historical parameter change rate and change pattern, predict virtual parameter data at the discontinuity location, and integrate the virtual parameter data with the sampled parameters; a task priority adjustment module, coupled to the virtual data integration module, for generating a high-speed mode rule for a device for which no parameters have been collected and the collection time exceeds a set time, and determining whether the device status corresponding to the parameters satisfies the high-speed mode rule; if not, separating and pausing the task based on the task priority; The device task mode extraction module includes: The task pattern recognition unit is used to identify the task pattern of the equipment, analyze the historical task data of each equipment, identify the task execution frequency and average execution time, and generate the task pattern; A collaboration frequency analysis unit, coupled to the task pattern recognition unit, is used to evaluate collaboration frequency, analyze the similarity of parameter types of each device, identify task collaboration relationships between devices, construct a multi-device task collaboration network diagram, and identify the collaboration frequency between each device based on the task collaboration network diagram to evaluate the impact on the overall task completion efficiency; a task conflict point marking unit, coupled to the collaboration frequency analysis unit, for analyzing task conflict points between devices, identifying resource contention during task execution, and marking potential task conflict nodes based on historical data and the collaboration network; a pressure index calculation unit, coupled to the task conflict point marking unit, for establishing a task pressure prediction model based on the identified task patterns and collaborative relationships, predicting future task pressures, and calculating the pressure index of the device under different task scenarios; a task priority division unit, coupled to the pressure index calculation unit, for setting a decision factor affecting the task priority, and prioritizing the tasks executed by the device based on a comparison relationship between the device pressure index and the decision factor; a preliminary scheduling strategy generating unit, coupled to the task priority division unit, for generating a preliminary scheduling strategy based on the task pressure index and the task priority; The collaboration frequency of the devices is: Among them, C D (v) is the degree centrality, which indicates the collaboration frequency of task node v, D(v) is the number of edges of task node v, that is, the number of devices collaborating with task node v, and n is the total number of devices; The overall task completion efficiency is: Among them, E is the overall task completion efficiency, w i is the edge weight, i.e. the frequency of collaboration between devices. The larger the weight, the more frequent the collaboration. i The efficiency of the equipment in performing its tasks; The pressure indicators of the equipment in different mission scenarios are: P i =αF i +βU i +γC i ; C i =C D (i)×(n-1); Among them, P i is the pressure index of device i in a certain task scenario, F i is the task execution frequency, U i is the resource utilization rate of device i, C i is the collaboration frequency, which indicates the frequency of device i collaborating with other devices in the task collaboration network, and α, β, and γ are weight coefficients.

6. The device multi-parameter statistical calculation system based on the general task scheduling platform according to claim 5 is characterized in that: The dynamic scheduling module includes: The equipment operation status collection unit is used to collect the current operation status data of each device after the initial scheduling strategy is implemented, and use the collected status data as a reference baseline; a real-time load level calculation unit, coupled to the device operation status acquisition unit, for comparing the collected current operation status data with a reference baseline in real time to calculate the load level of each device; a load threshold setting unit, coupled to the real-time load level calculation unit, for setting a load threshold by analyzing historical performance data, and marking the device as being in an overloaded state when the load level of a device exceeds the set threshold; an average load calculation unit, coupled to the load threshold setting unit, for calculating the average load of all devices, and extracting all devices with load levels lower than the average load and marking them as task receiving devices; The task allocating unit is coupled to the average load calculating unit and is used to allocate the device tasks in the overload state to the task receiving devices and maintain the load of the task receiving devices within a set threshold.

7. The device multi-parameter statistical calculation system based on the general task scheduling platform according to claim 5 is characterized in that: The sampling frequency adjustment module includes: A parameter change rate calculation unit is used to compare the collected data and calculate the change rate of each parameter; a normal fluctuation interval setting unit, coupled to the parameter change rate calculation unit, for analyzing historical data of the device and setting a normal fluctuation interval for each parameter based on the statistical characteristics of the change rate; a parameter change level classification unit, coupled to the normal fluctuation range setting unit, for classifying the parameter change level into three stages according to the parameter change rate, including a high level, a stable level, and a medium level; The sampling strategy setting unit is coupled to the parameter change level division unit and is used to set sampling strategies for different levels respectively.

8. The device multi-parameter statistical calculation system based on the general task scheduling platform according to claim 5 is characterized in that: The virtual data integration module includes: The device status monitoring unit is used to continuously collect device parameters for the adjusted sampling frequency, check whether the data collected in each sampling frequency interval is complete, and compare whether the data coverage of different frequency intervals is complete; a data interruption analysis unit, coupled to the device status monitoring unit, for extracting information about devices with data interruptions, recording the locations and frequencies of the interruptions, analyzing the impact of the interruptions on device operating status monitoring, and generating an interruption severity level; a parameter fluctuation interval evaluation unit, coupled to the data discontinuity analysis unit, for evaluating whether the current parameter is within the corresponding normal fluctuation interval and performing abnormal pattern detection on the parameter fluctuation; A virtual data generating unit is coupled to the parameter fluctuation interval evaluating unit and is used to establish a time series prediction model, predict virtual parameter data of the discontinuous position for the parameters, and mark them; The state monitoring update unit is coupled to the virtual data generation unit and is used to align and integrate the collected parameter data with the predicted virtual data, and update the integrated complete data stream to the device state monitoring unit.

9. The device multi-parameter statistical calculation system based on the general task scheduling platform according to claim 5, characterized in that: The task priority adjustment module includes: The unit for calculating the duration of non-data collection is used to set a collection duration quota. For devices whose parameters have not been collected, the unit calculates the duration of the non-data collection, compares whether the collection duration quota has been exceeded, and reads the current operating status and pressure index of the device that has exceeded the quota. The high-speed mode intervention demand analysis unit is coupled to the non-data collection time calculation unit and is used to analyze whether high-speed mode intervention is needed based on the current operating status and pressure indicators, and to determine whether the equipment meets the high-speed mode rule requirements.

Images