Underground miner performance management and incentive system based on real-time salary settlement

Abstract

The invention relates to the technical field of mining area intelligent management and salary incentive, in particular to an underground miner performance management and incentive system based on real-time salary settlement, and the system optimizes an income distribution rule through a causal inference model, and dynamically adjusts the weight distribution of basic income, task bonus and reward income; meanwhile, task allocation and resource scheduling are dynamically optimized by adopting a multi-agent reinforcement learning technology, so that the resource utilization efficiency and the task completion rate are ensured to be maximized; finally, by generating a multi-level incentive strategy of personal rewards, team rewards and global rewards, the working enthusiasm of the miner individuals and the teams is stimulated comprehensively; according to the method, fairness of salary distribution, diversity of incentive mechanisms and high efficiency of mining area management are realized, and mining area production benefits and miner satisfaction are effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent management and salary incentive technology for mining areas, and in particular to a performance management and incentive system for underground miners based on real-time salary settlement. Background Art

[0002] In the coal mining environment, production efficiency and safety management are crucial. At the same time, miners' work enthusiasm and fair salary incentive mechanisms also have an important impact on the operation of the enterprise. However, traditional miner management and salary distribution methods often rely on manual records and regular assessments, which cannot achieve comprehensive quantification and accurate feedback on miners' real-time work status, task completion and safe operation behaviors. This lagging management method is not only difficult to adapt to the complex and dynamic mining environment, but may also lead to a decrease in miners' trust in work allocation and salary incentives.

[0003] Existing technologies usually use fixed remuneration or distribution based on simple task completion, such as regularly evaluating miners' task completion or safety records, and then manually calculating remuneration. However, these technologies have significant drawbacks due to the following reasons:

[0004] Existing technologies cannot collect the working status and task completion of miners in real time, resulting in significant data lag, which affects the immediacy and fairness of salary distribution. Existing technologies lack a multi-level incentive mechanism for individual performance, team collaboration and overall resource optimization of the mining area. They only focus on a single indicator and are difficult to stimulate the overall work enthusiasm of miners. Traditional means have poor flexibility in task allocation and resource scheduling, lack dynamic adjustment capabilities, and are prone to waste of resources and task delays. Summary of the invention

[0005] In view of the many problems existing in the above-mentioned prior arts, the present invention provides an underground miner performance management and incentive system based on real-time salary settlement. The present invention collects multimodal data on miner status, working environment and equipment operation, and combines principal component analysis and tensor decomposition technology to realize data feature extraction and performance evaluation; adopts causal inference model to optimize profit distribution rules, and uses multi-agent reinforcement learning to dynamically adjust task allocation and resource scheduling; finally generates multi-level incentive strategies, including personal rewards, team rewards and global rewards, to comprehensively improve the production efficiency of the mining area and the enthusiasm of miners.

[0006] A performance management and incentive system for underground miners based on real-time salary settlement, including:

[0007] The acquisition module is used to obtain the physiological parameters of the miners as the miner status data, obtain the working environment conditions as the environment status data, and obtain the equipment operation status as the equipment status data;

[0008] The quota generation module is used to clean and normalize the miner status data, environmental status data and equipment status data, extract key features through principal component analysis, generate characteristic variable data, and generate dynamic task quota data through nonlinear modeling in combination with the miner's historical task data;

[0009] The task management module is used to collect miners' task status data based on dynamic task quota data, and integrate task progress data and equipment operation data to generate real-time comprehensive data; it is equipped with a task prediction submodule to predict task progress through real-time comprehensive data, generate dynamic task allocation data and adjusted task plan data;

[0010] A performance module is used to generate comprehensive performance data based on dynamic task allocation data and adjusted task plan data, wherein the comprehensive performance data is comprehensively calculated for task completion data, safety operation data and equipment maintenance data through a weight allocation rule, and the daily revenue data is calculated accordingly;

[0011] The incentive module is used to generate detailed revenue data and performance feedback data based on the current day's revenue data, and to optimize incentive rules and generate incentive strategies by analyzing historical performance data and feedback data.

[0012] Preferably, the acquisition module specifically includes:

[0013] The heart rate and body temperature data of miners are obtained in real time through wearable devices to generate data streams of the miners’ current status;

[0014] The temperature, humidity, air pressure and gas concentration data of the working area are collected through environmental monitoring equipment to generate real-time status data of the working environment;

[0015] The equipment monitoring module collects the equipment operation time, power and fault records in real time to generate data input on the current performance of the equipment.

[0016] Preferably, the quota generation module extracts characteristic variables with strong nonlinear correlation from miner status data, environmental status data and equipment status data through characteristic variable extraction technology, uses principal component analysis to reduce data dimensions and improve modeling efficiency, and at the same time generates dynamic task quota data based on the miner's historical task completion status for real-time adjustment of task allocation strategies.

[0017] Preferably, the quota generation module dynamically adjusts the generation process of dynamic task quota data by constructing a multi-objective optimization model, comprehensively considering the current status of miners, operating environment risks and equipment operation reliability, to ensure that the task quota allocation matches the actual operating capacity.

[0018] Preferably, the task management module includes:

[0019] The task status collection submodule is used to collect the task completion amount, operation time and location information of miners during the task execution process;

[0020] The real-time comprehensive data generation submodule is used to integrate the miner's task status data with the task progress data and the equipment operation data to generate real-time comprehensive data that can be used for task allocation optimization;

[0021] The task prediction submodule is used to predict task progress and detect potential task conflicts based on real-time comprehensive data, and generate input data for optimal task allocation.

[0022] Preferably, the task management module dynamically optimizes the task dependencies in the real-time comprehensive data through a multi-agent reinforcement learning method, combines unfinished tasks with the current miner status, and generates task allocation data and adjusted task planning data.

[0023] Preferably, the performance module performs feature decomposition of the task completion data, safety operation data and equipment maintenance data through a tensor decomposition method, and performs weighted comprehensive calculation of each data in combination with dynamically adjusted weight rules to generate daily comprehensive performance data of miners.

[0024] Preferably, when the weight distribution rule of the performance module is dynamically adjusted, calculations are performed based on the difficulty of task completion, the importance of safe operation, and the actual contribution of equipment maintenance, and the weight distribution is corrected in real time.

[0025] Preferably, the incentive module analyzes the relationship between the miner's historical performance data and current feedback data through a causal inference model, and dynamically adjusts the distribution rules of basic income, task bonuses and reward income.

[0026] Preferably, the incentive module further generates a multi-level incentive strategy, including individual rewards based on individual comprehensive performance, team rewards based on team task completion rate, and global rewards based on resource optimization of the entire mining area.

[0027] Compared with the prior art, the advantages and beneficial effects of the present invention are:

[0028] The present invention realizes the dynamic acquisition of miners' task status, safety behavior and equipment maintenance data through real-time data collection and fusion technology, solves the problem of data lag in the prior art, and improves the fairness and timeliness of salary distribution;

[0029] The present invention realizes the comprehensive evaluation of the multi-dimensional performance of miners through dynamic performance evaluation based on tensor decomposition and causal inference model, optimizes the distribution rules of task bonuses, basic income and reward income, and overcomes the problem of single incentive in the prior art;

[0030] The present invention uses multi-agent reinforcement learning technology to dynamically optimize task allocation and resource scheduling, achieving flexibility and efficiency in task allocation and significantly improving the production efficiency of the mining area;

[0031] The present invention fully stimulates the enthusiasm of individual miners and teams by generating a multi-level incentive strategy of personal rewards, team rewards and global rewards, and enhances the overall efficiency of the mining area. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a structural block diagram of the system of the present invention;

[0033] Figure 2 It is a schematic diagram of the comprehensive performance generation process of the performance module in the present invention. DETAILED DESCRIPTION

[0034] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is obvious that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0035] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "include", "comprising", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0036] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0037] like Figure 1 As shown, a performance management and incentive system for underground miners based on real-time salary settlement includes:

[0038] The acquisition module is used to obtain the physiological parameters of the miners as the miner status data, obtain the working environment conditions as the environment status data, and obtain the equipment operation status as the equipment status data;

[0039] Preferably, the acquisition module specifically includes:

[0040] The heart rate and body temperature data of miners are obtained in real time through wearable devices to generate data streams of the miners’ current status;

[0041] The temperature, humidity, air pressure and gas concentration data of the working area are collected through environmental monitoring equipment to generate real-time status data of the working environment;

[0042] The equipment monitoring module collects the equipment operation time, power and fault records in real time to generate data input on the current performance of the equipment.

[0043] The acquisition module consists of three parts: wearable devices for obtaining miners' physiological parameters, environmental monitoring devices for monitoring environmental conditions, and equipment monitoring modules for monitoring equipment operating conditions. These devices are centered on data acquisition and use multimodal sensing technology to integrate real-time information from different sources into structured data input.

[0044] Wearable devices monitor miners' physiological parameters, including heart rate and body temperature, in real time through built-in sensors. The sensing unit in the device (such as photoelectric volumetric pulse wave sensor for heart rate monitoring and thermistor sensor for body temperature monitoring) converts the collected analog signal into a digital signal and transmits it to the data processing end through wireless transmission (such as Bluetooth or LoRa protocol). Combined with the timestamp and miner identity, a continuous miner status data stream is generated to reflect the miner's current physical state. In the system application, the collected heart rate and body temperature data are used to detect the miner's fatigue or stress response through a specific threshold analysis method. For example, when the heart rate exceeds 120 beats / minute and the body temperature is higher than 38°C, the system will mark it as a "high load state" and trigger adaptive adjustments to task allocation.

[0045] Environmental monitoring equipment is deployed at key points in the mining area to collect environmental parameters such as temperature, humidity, air pressure and gas concentration in real time. The equipment filters and corrects the raw data through multi-sensor fusion technology (such as using the Kalman filter algorithm to eliminate sensor errors) to generate an environmental status data stream to reflect the real-time working environment of the mining area. For example, the gas concentration sensor monitors the gas concentration in the mining area in real time based on the principle of catalytic combustion. When the concentration exceeds the set threshold (such as 1.25%), the system will mark the environmental status as "dangerous" and notify the management end to take emergency measures.

[0046] The equipment monitoring module obtains equipment operation data through embedded sensors, including operation time, power consumption, and fault records. The collection of equipment status data mainly relies on vibration sensors (for monitoring abnormal vibration), current sensors (for monitoring power consumption), and embedded diagnostic units (for recording fault logs). For example, when the vibration frequency exceeds the normal range, the module will automatically mark it as "abnormal operation status" and upload the fault record data to the management platform in real time.

[0047] Through multimodal perception and sensor correction technology, the acquisition module ensures the high timeliness and high accuracy of data, providing a reliable basis for task allocation and performance calculation. The module can simultaneously obtain multi-dimensional data on the physiological status of miners, working environment and equipment operating status, supporting the system to perceive the dynamics of the mining area in all directions. By setting reasonable thresholds and anomaly detection mechanisms, the acquisition module can timely warn of potential risks in the mining area and enhance the reliability and safety of the system.

[0048] Example: In a coal mine operation scenario, in order to obtain the working status and environmental conditions of miners in real time, the acquisition module is deployed as follows:

[0049] The wearable devices worn by miners obtain their heart rate and body temperature in real time, with the heart rate ranging from 40 to 150 beats per minute and the body temperature ranging from 35°C to 42°C. After data collection, if the miner's heart rate exceeds 130 beats per minute for 10 consecutive minutes, the system will mark it as "high load status" and send an alert to the management end.

[0050] Environmental monitoring equipment is installed at the main channel entrance and deep well point of the mine operation area to monitor the gas concentration in real time. When the gas concentration exceeds the standard, the system automatically activates the fan exhaust equipment and marks the environment as a "hazardous area" on the system interface.

[0051] The equipment monitoring module is connected to the power unit of the transport vehicle and collects the vehicle's operating power in real time. If the vehicle's operating power deviates from the normal range by more than 10%, the system records the fault event and recommends maintenance.

[0052] Through the effective operation of the acquisition module, the dynamic monitoring data of the coal mine working environment and miners' status were successfully uploaded to the system for subsequent dynamic task allocation and performance management, realizing safe and efficient mine operation management.

[0053] The quota generation module is used to clean and normalize the miner status data, environmental status data and equipment status data, extract key features through principal component analysis, generate characteristic variable data, and generate dynamic task quota data through nonlinear modeling in combination with the miner's historical task data;

[0054] The principle of the quota generation module lies in the collaborative work of data preprocessing and feature extraction, nonlinear modeling and task quota generation. Through the standardization and analysis of multi-source data, the module realizes the precise allocation of dynamic task quotas.

[0055] Data cleaning refers to filtering and correcting the abnormal values ​​in the collected miner status data, environmental status data, and equipment status data, and using the 3σ method based on the mean and standard deviation or the isolation forest algorithm to identify and remove noise points to ensure the accuracy and stability of the data. Normalization is to unify the data of different physical quantities (such as heart rate, temperature, and gas concentration) into the [0,1] interval through the Min-Max normalization method to facilitate subsequent modeling and analysis.

[0056] The cleaned and normalized data is analyzed by principal component analysis (PCA) to extract key features. The core idea of ​​PCA is to extract the principal component that can explain the maximum variance in the data through eigenvalue decomposition, reduce redundant dimensions and improve computational efficiency. For example, suppose the input data matrix is ​​X = [x 1 ,x 2 ,…,x n ], the principal component is Z = XW, where W is the eigenvector matrix, which represents the linear combination relationship of the main features. The extracted characteristic variable data include the principal component reflecting the miner's working ability (such as the comprehensive evaluation index of heart rate and body temperature), the principal component reflecting the environmental conditions (such as the comprehensive index of gas concentration and air pressure), and the principal component reflecting the equipment performance (such as the evaluation index of equipment operating power and failure rate).

[0057] The generation of dynamic task quota is realized through nonlinear modeling. Combining the historical task data of miners, the support vector regression (SVR) method based on radial basis kernel function is used to establish a nonlinear relationship model to predict the task quota value.

[0058] Dynamic task quotas are used to guide subsequent task allocation. The system prioritizes tasks that are suitable for miners' capabilities and environmental conditions based on the size of the quota. Miners with higher quotas are assigned more complex or high-risk tasks, while miners with lower quotas are assigned lightweight tasks to achieve scientific regulation of task loads.

[0059] Preferably, the quota generation module extracts characteristic variables with strong nonlinear correlation from miner status data, environmental status data and equipment status data through characteristic variable extraction technology, uses principal component analysis to reduce data dimensions and improve modeling efficiency, and at the same time generates dynamic task quota data based on the miner's historical task completion status for real-time adjustment of task allocation strategies.

[0060] Feature variable extraction screens out variables that are highly correlated with task completion efficiency by analyzing the nonlinear correlation between miner status data, environmental status data, and equipment status data. For example, by calculating the Pearson correlation coefficient, mutual information value, or kernel function-based correlation score between variables, key parameters such as heart rate, gas concentration, and equipment vibration frequency are extracted as feature variables. These variables can reflect the current working ability of miners, the safety of the working environment, and the reliability of the equipment operating status.

[0061] After extracting the characteristic variables, the module uses principal component analysis to reduce the dimensionality of high-dimensional characteristic data. The role of PCA is to reduce the interference of redundant features on modeling through linear transformation, while retaining the principal components that can explain the maximum variance in the data. For example, suppose the input data matrix is ​​X = [x 1 ,x 2 ,…,x n ], its principal component can be expressed as:

[0062] Z=XW

[0063] Among them, Z is the principal component matrix after dimensionality reduction, and W is the eigenvector matrix. By calculating the variance contribution rate of the principal component, the principal component with a cumulative contribution rate of more than 95% is selected as the input data for subsequent modeling.

[0064] The module combines the extracted principal components and the miners' historical task completion data to construct a dynamic task quota model, using a nonlinear modeling method based on support vector regression (SVR). The model captures the complex nonlinear relationship between feature variables and task quotas through the radial basis kernel function (RBF). The specific calculation formula is:

[0065]

[0066] Among them, Q is the dynamic task quota, K(x i ,x) is the kernel function representing the feature variable x and the support vector x i The nonlinear relationship between i is the support vector weight, and b is the model bias. By training the model, the task quota can reflect the miners' working ability, environmental risks and equipment status in real time and adjust dynamically.

[0067] Dynamic task quota data directly affects the task allocation strategy. Based on the quota, the system will prioritize tasks with higher complexity or higher safety requirements to miners with higher quotas, and assign light-load tasks to miners with lower quotas, so as to achieve balanced workload and maximize production efficiency.

[0068] Through nonlinear correlation analysis of multi-source data, the module can effectively screen key variables directly related to task completion efficiency and improve the pertinence of data modeling. PCA dimensionality reduction reduces data redundancy, improves the efficiency and stability of subsequent nonlinear modeling, and reduces computational complexity. Dynamic task quota modeling achieves accurate matching of miners' capabilities and task requirements by combining real-time data with historical task completion status, and enhances the flexibility and adaptability of task allocation. By dynamically adjusting the quota, the module avoids resource waste and uneven load in task allocation, effectively improving the overall production efficiency of the mining area.

[0069] In an example, in a mining area application, the operation flow of the quota generation module is as follows:

[0070] Feature variable extraction: extract feature variables from the collected data of miners’ heart rate (80 beats / minute), gas concentration (1.2%), and equipment vibration frequency (50Hz). These variables are highly correlated with miners’ physical fitness, environmental safety, and equipment performance, and the core features used as modeling inputs are selected through correlation analysis.

[0071] Dimensionality reduction processing,PCA is used to reduce the dimension of feature variables. The original three-dimensional variable matrix is ​​calculated to generate two principal components, whose variance contribution rates are 60% and 30% respectively, and the cumulative contribution rate reaches 90%, which serve as the input for subsequent modeling.

[0072] Task quota modeling, input the reduced principal components and the miner's historical task completion rate (90%) into the SVR model to generate a dynamic task quota Q = 0.85. This quota reflects the complexity and load capacity of the tasks that the miner can complete in the current state.

[0073] The task strategy is adjusted. According to the task quota Q=0.85, the system assigns high-complexity pipeline maintenance tasks to miners, and assigns low-quota miners (such as Q=0.5) to light-load material transportation tasks.

[0074] Through the above-mentioned embodiments, the quota generation module realizes real-time modeling of dynamic conditions in the mining area and optimization of task strategies, providing key support for the performance management and incentive system for underground miners.

[0075] Preferably, the quota generation module dynamically adjusts the generation process of dynamic task quota data by constructing a multi-objective optimization model, comprehensively considering the current status of miners, operating environment risks and equipment operation reliability, to ensure that the task quota allocation matches the actual operating capacity.

[0076] The quota generation module adopts a multi-objective optimization model, taking the current status of miners, working environment risks and equipment operation reliability as input variables, and dynamically optimizing with the rationality of task quota generation as the goal. Each input variable is weighted by a weight factor to reflect its comprehensive impact on the task quota. The objective function of the model can be expressed as:

[0077] Q = argmax(w 1 S+w 2 E+w 3 R)

[0078] Among them, Q represents the dynamic task quota; S represents the current state of the miner, which is calculated based on physiological parameters such as heart rate and body temperature; E represents the operating environment risk, which is evaluated based on environmental indicators such as gas concentration, temperature and humidity; R represents the equipment operation reliability, which is quantified based on the equipment operation time and fault records; w 1 、w 2 、w 3 They represent the weight factors of the corresponding variables, which are dynamically adjusted by the task characteristics. The model dynamically adjusts the weights through linear weighting or other nonlinear optimization algorithms (such as Pareto frontier optimization method) to ensure that the task quotas generated by miners under different states, environments and equipment conditions have the best adaptability.

[0079] Miner's current status (S): The physical status of miners is calculated by using parameters such as heart rate and body temperature collected by wearable devices and combining them with time series analysis technology. For example, the closer the heart rate is to the physical upper limit, the lower its contribution to the task quota. Working environment risk (E): Environmental monitoring equipment is used to monitor environmental indicators such as gas concentration, temperature and humidity in real time to generate an environmental risk score. If the gas concentration exceeds 1.5%, the environmental risk score drops to a low value. Equipment operation reliability (R): The equipment operation status is monitored through the equipment monitoring module, and the equipment support capability is calculated in combination with the vibration frequency and fault records. For example, if the vibration frequency of the equipment exceeds the standard range, the equipment reliability score is reduced.

[0080] The module adjusts the weight distribution according to the dynamic changes of the input variables. When the task complexity is high (such as high-risk tasks), the environmental risk weight increases; when the miner status is poor, the miner status weight decreases. The dynamic adjustment process is implemented based on a multi-objective optimization algorithm, such as using the particle swarm optimization algorithm (PSO) to quickly converge on the optimal weight distribution scheme.

[0081] After the dynamic task quota is generated, it is directly linked to the task allocation strategy. Miners with higher quotas are given priority to complex tasks, while miners with lower quotas are assigned low-load tasks to ensure scientific allocation of workload.

[0082] Through the multi-objective optimization model, the module can simultaneously consider the three factors of miner status, environmental conditions and equipment support, and the generated task quota has comprehensive adaptability. The dynamic weight adjustment mechanism of the input variables can respond to changes in mining conditions in real time and improve the system's adaptability to complex operating scenarios. The precise generation of task quotas ensures that the task burden is highly matched with the miner's ability, which can not only give full play to the value of high-capacity miners, but also protect the safety of miners with weaker physical fitness. By reasonably allocating task quotas, the module effectively reduces resource waste and efficiency losses caused by uneven task distribution.

[0083] Implementation example, mining area application example: At a coal mine operation site, the quota generation module dynamically adjusts the task quota based on real-time data. The specific operations are as follows:

[0084] (1) Input data collection and evaluation:

[0085] Current status of the miner: The miner's heart rate (85 beats / minute) and body temperature (37.2°C) are collected through wearable devices to generate a status score S=0.8.

[0086] Working environment risk: The gas concentration (1.3%) and temperature and humidity (28°C, 60%) are collected through environmental monitoring equipment to generate an environmental risk score of E=0.6.

[0087] Equipment operation reliability: The vibration frequency (within the normal range) and fault records (none) are monitored through the equipment monitoring module to generate an equipment reliability score of R=0.9.

[0088] (2) Dynamically adjust weights:

[0089] The current task is a high-risk and complex operation (such as pipeline maintenance task), and the weight of environmental risk is w 2 Automatically increase to 0.5, miner status and equipment reliability weights are 0.3 and 0.2 respectively. The weight adjustment formula is:

[0090] w 1 +w 2 +w 3 =1

[0091] The module calculates dynamic task quotas based on a multi-objective optimization model:

[0092] Q=(w 1 S+w 2 E+w 3 R)=0.3×0.8+0.5×0.6+0.2×0.9=0.71

[0093] The calculation result is 0.71, which means that the miner’s current status is suitable for undertaking tasks of medium complexity.

[0094] According to the task quota Q=0.71, the system assigns miners to maintenance operations of medium complexity, and assigns high-quota miners (Q>0.8) to tasks in high-risk areas, while assigning low-quota miners (Q<0.6) to operations in low-risk areas.

[0095] Through this implementation process, the quota generation module achieved task matching optimization based on real-time data, improved the overall operating efficiency of the mining area and reduced safety risks.

[0096] The task management module is used to collect miners' task status data based on dynamic task quota data, and integrate task progress data and equipment operation data to generate real-time comprehensive data; it is equipped with a task prediction submodule to predict task progress through real-time comprehensive data, generate dynamic task allocation data and adjusted task plan data;

[0097] The task management module collects key data in real time during task execution by connecting to the positioning and data recording devices worn by miners. These data include the miner's location information, the amount of tasks completed, and the operation time, forming a task status data stream for evaluating the miner's current task progress. For example, the GPS positioning data is used to determine whether the miner's current location is consistent with the designated task area; the completion progress is evaluated in real time by comparing the completion amount record with the task target value.

[0098] The module integrates task status data with task progress data (such as the staged completion rate of the overall task) and equipment operation data (such as the current load and performance indicators of the equipment) to generate real-time comprehensive data. The integration process uses weighted data fusion technology to improve data consistency and eliminate noise. For example, the equipment load information and the miner's task completion rate are corrected through time synchronization to ensure accurate reflection of the task status. The integrated real-time comprehensive data can fully reflect the miner's execution ability, task progress and equipment support.

[0099] The task prediction submodule is based on real-time comprehensive data and uses time series analysis and task dependency modeling technology to accurately predict task progress. This module uses the Variational Bayesian Factorization (VBF) algorithm to extract the potential patterns of miners' work efficiency and task dependencies from comprehensive data, and predict task completion time and possible task conflicts. For example, by analyzing the miners' historical work efficiency and equipment reliability data, the expected completion time range of the current task is predicted, and task progress prediction data is generated.

[0100] Based on the task progress prediction data, the module uses the Graph Attention Network (GAT) to optimize the dynamic task allocation strategy. The GAT algorithm generates dynamic task allocation data and adjusted task planning data by constructing a task dependency graph, identifying the priorities and resource conflicts between tasks. For example, when a task is delayed due to equipment failure, the module automatically reallocates tasks and prioritizes resources to task nodes with higher completion value to ensure that the overall task objectives are not affected.

[0101] Preferably, the task management module includes:

[0102] The task status collection submodule is used to collect the task completion amount, operation time and location information of miners during the task execution process;

[0103] The real-time comprehensive data generation submodule is used to integrate the miner's task status data with the task progress data and the equipment operation data to generate real-time comprehensive data that can be used for task allocation optimization;

[0104] The task prediction submodule is used to predict task progress and detect potential task conflicts based on real-time comprehensive data, and generate input data for optimal task allocation.

[0105] The task status collection submodule collects key data of miners in real time during task execution by connecting to the sensor equipment, positioning system and work recorder worn by miners. Task completion data is used to record the amount of work completed by miners, such as the weight of ore mined or the amount of materials transported; operation time data is used to record the length of time miners work on a task; and positioning information is used to track the current working area of ​​the miners. All data is synchronized and uploaded to the system database through timestamps to generate task status data streams, providing a basis for subsequent comprehensive data generation and task optimization.

[0106] The real-time comprehensive data generation submodule combines the task status data with the task progress data and equipment operation data through data integration technology to generate real-time comprehensive data that reflects the global status of task execution.

[0107] Task status data includes the miner's current task completion amount, operation time and location information. Task progress data records the overall stage completion rate of the task (such as the completion progress after being decomposed into several subtasks). Equipment operation data monitors the current load, operation time and fault records of the equipment. Data integration uses a weighted fusion algorithm to assign different weights to each data source. For example, the impact of equipment operation status on task completion is greater than the miner's location information. During the integration process, the module will perform denoising processing on the data (such as Kalman filtering) to ensure the accuracy and consistency of the comprehensive data. The generated real-time comprehensive data fully reflects the miner's task execution efficiency, current task status and equipment support capabilities.

[0108] The task prediction submodule predicts task progress and detects potential task conflicts through real-time comprehensive data analysis and modeling. Time series analysis technology is combined with machine learning models (such as LSTM, long short-term memory network) to extract the dynamic pattern of task execution, so as to predict the time range for task completion. For example, based on the current task completion rate and historical efficiency of the miner, the estimated completion time is calculated. Based on the task dependency graph, analyze whether there is a conflict in the current resource allocation. For example, detect resource competition problems when the load of a certain device may be requested by multiple tasks at the same time. After detecting a conflict, the module generates a conflict warning and provides adjustment suggestions to the task allocation subsystem.

[0109] By collecting task completion, operation time and location information, the task management module can fully understand the execution of miners' current tasks and provide accurate input for task optimization. Real-time comprehensive data integrates miners' task status, task progress and equipment operation status, fully reflects the execution efficiency and bottlenecks of tasks, and provides high-quality data support for subsequent prediction and optimization. The task prediction submodule can timely discover potential problems in task execution through dynamic pattern extraction and conflict detection, enhancing the flexibility and intelligence of task allocation. Through conflict detection and optimization suggestions, the module ensures the rationality and efficiency of task allocation resources and reduces delays caused by resource conflicts.

[0110] Embodiment, mining area application example: In a stone transportation task in a mining area, the specific operation process of the task management module is as follows:

[0111] Task status data collection: The miners’ completed transportation volume (e.g., 10 tons) is collected through the work recording equipment they wear, the miners’ current location (e.g., the middle of the transportation route) is recorded through the positioning equipment, and the working time (e.g., 3 hours) is collected through the operation time record. All data generates a task status data stream and is uploaded to the task management system.

[0112] Real-time comprehensive data generation: Combine task progress data (such as the overall transportation task completion rate is 40%) and equipment operation data (such as the vehicle load rate is 90% and the vibration frequency is within the normal range) to generate comprehensive data that reflects the miners' task execution efficiency and equipment support stability.

[0113] Task progress prediction and conflict detection: Through the task prediction submodule, based on the miner's current transportation speed (3 tons per hour) and the vehicle's operating status, it is predicted that the current task is expected to be completed in 2 hours, and it is detected that the high load state of the transport vehicle may cause delays to subsequent transport tasks.

[0114] Task adjustment and optimization suggestions: The task management module generates optimization suggestions to the scheduling system, reallocates transportation tasks, and allocates other miners' light-load tasks near the current vehicle to reduce the load on the single vehicle. At the same time, the current miner's task plan is adjusted to give priority to completing high-value transportation tasks to ensure the priority execution of tasks. The optimization suggestions are pushed in real time through the miner's terminal device, and the adjusted plan takes effect quickly.

[0115] Preferably, the task management module dynamically optimizes the task dependencies in the real-time comprehensive data through a multi-agent reinforcement learning method, combines unfinished tasks with the current miner status, and generates task allocation data and adjusted task planning data.

[0116] The task management module first constructs a task dependency graph through real-time integrated data. Each node in the graph represents a specific task, and the weight of the edge represents the strength of the dependency between tasks. For example, the completion of a task is a prerequisite for the start of another task. The weight of the dependency can be calculated by the overlap of resource requirements, task priority, and completion time urgency. For example, assuming that task A and task B need to share the same device, and the completion of task A is a prerequisite for the start of task B, the dependency weight between A and B is higher.

[0117] The application of Multi-Agent Reinforcement Learning (MARL) in task management aims to optimize tasks through the collaboration of multiple agents. Each agent represents a task, and its state includes the current progress of the task, resource requirements, and expected completion time. The action space includes task reallocation and priority adjustment.

[0118] The module uses the Deep Q-Network (DQN) algorithm, and the agent learns the optimal allocation strategy through interaction with the environment. The reward function R adjusted for each task is defined according to the following objective function:

[0119] R=w 1 ·P+w 2 ·C+w3 ·T

[0120] Among them, P is the task completion rate; C is the resource utilization rate; T is the total task delay time; w ... 1 、w 2 、w 3 The weight factors reflect the priorities of task completion, resource utilization, and delay, respectively. Through training, the agent gradually optimizes the allocation strategy to ensure that the task completion rate and resource utilization are maximized while minimizing task delay.

[0121] The module dynamically adjusts the action strategy of the agent by continuously updating the task dependency graph and task status. For example, when a device fails and cannot continue to support the current task, the task status is updated to "delayed". The agent reallocates task resources and adjusts priorities according to the new status, generating new task allocation data and adjusted task plans.

[0122] Through multi-agent reinforcement learning, the module can respond to changes in the dynamic environment of the mine in real time, ensuring the efficiency and flexibility of task allocation. Collaborative learning between agents can reasonably allocate limited resources in the mine and maximize the utilization of equipment and manpower. Based on the modeling and optimization of task dependencies, the module can quickly identify key tasks and prioritize resource allocation to ensure the efficient execution of the overall task chain. The dynamic adjustment mechanism can respond quickly to emergencies (such as equipment failures or task delays) to reduce the impact on overall production.

[0123] Embodiment, mining area application example: In a pipeline maintenance task in a mining area, the task management module realizes task optimization through the following steps:

[0124] Task dependency modeling, the pipeline maintenance task is decomposed into three subtasks: inspection task (A), maintenance task (B) and test task (C). The task dependency is A→B→C, where the completion of A is the premise of B, and the completion of B is the premise of C. At the same time, the maintenance task (B) and the inspection task (A) need to share a detection device, and their dependency graph is expressed as:

[0125] G=(N,E),N={A,B,C},E={(A,B),(B,C)}

[0126] In the weight calculation, the weight of A→B is 0.8, and the weight of B→C is 0.5.

[0127] Task optimization and dynamic adjustment: When the detection device fails and causes Task A to be delayed, the module updates the task status in real time, marks Task A as "delayed", and adjusts the task plan. In the multi-agent reinforcement learning module, the reward function of Agent A is recalculated in real time, its task completion rate P is reduced to 0.6, and the delay time T is increased to 10 minutes. Based on the optimization strategy, the module reallocates the resources of Task A to devices with more stable status and generates an adjusted task plan.

[0128] Based on the adjustment results, the system generates new task allocation data, lowers the priority of the maintenance task (B), and allocates key resources to the test task (C) to ensure the stability and optimal execution of the overall task chain.

[0129] like Figure 2 As shown, the performance module is used to generate comprehensive performance data based on dynamic task allocation data and adjusted task plan data, wherein the comprehensive performance data is comprehensively calculated for task completion data, safety operation data and equipment maintenance data through a weight allocation rule, and the daily revenue data is calculated accordingly;

[0130] The generation of comprehensive performance data relies on dynamic task allocation data and adjusted task plan data, which provide an objective record of the miners' workload, completion quality, and execution safety in the current task.

[0131] Task completion data includes the amount of tasks completed by miners within a specified time (such as the total amount of materials transported or the number of equipment repaired), which is used to evaluate miners' work efficiency. Safety operation data records the safety operation of miners in performing tasks through the equipment monitoring module (such as whether protective equipment is worn as required and whether operating specifications are followed). Equipment maintenance data includes equipment maintenance and troubleshooting records in which miners participate, which is used to measure miners' contribution to equipment use and maintenance.

[0132] The comprehensive performance data is calculated by weighting the above three types of data through the weight allocation rules. The weight allocation rules are dynamically adjusted according to the task type, task difficulty and safety requirements. For example:

[0133] For high-risk tasks, the weight of safe operation data is w 2 will be adjusted upwards appropriately;

[0134] For equipment-intensive tasks, the weight of equipment maintenance data is w 3 will account for a larger proportion.

[0135] The calculation formula of the comprehensive performance score E is:

[0136] E=w 1 ·F+w 2 ·S+w 3 ·M

[0137] Among them, F represents the task completion data, which indicates the miner's task completion efficiency; S represents the safety operation data, which indicates the miner's safety operation score in the task; M represents the equipment maintenance data, which indicates the miner's equipment maintenance contribution; w 1 、w 2 、w 3 Represents the weight factor, which reflects the influence of each data on the overall performance.

[0138] After the comprehensive performance data is generated, it is used to calculate the miner's daily income data. The income consists of three parts: basic income, task bonus and reward income: basic income is calculated according to the basic salary standard of the miner's position and is proportional to the comprehensive performance score. Task bonus is rewarded according to the task completion data and the part with the largest weight in the comprehensive performance score. Reward income is rewarded based on the miner's safety operation data and equipment maintenance data to encourage miners to work safely and actively maintain equipment.

[0139] The weight distribution rules and profit calculation formula can be dynamically adjusted according to the miners' actual work performance and task complexity. For example, if the completion volume of a task is low but the safety operation score is high, the system will adjust the profit distribution ratio and increase the reward profit part to encourage miners to maintain safety awareness.

[0140] Preferably, the performance module performs feature decomposition of the task completion data, safety operation data and equipment maintenance data through a tensor decomposition method, and performs weighted comprehensive calculation of each data in combination with dynamically adjusted weight rules to generate daily comprehensive performance data of miners.

[0141] The performance module receives data input from other system modules, including: Task completion data, generated by the task management module, contains quantitative indicators of miners completing tasks, such as the amount of work completed, work efficiency, and task complexity. Safety operation data, obtained through safety monitoring equipment, reflects the safety operation performance of miners during task execution, such as whether protective equipment is used correctly or whether operating specifications are followed. Equipment maintenance data, generated by the equipment monitoring module, records the frequency and effectiveness of miners' participation in equipment maintenance, including routine inspections, fault repairs, etc. These data are aligned by timestamp and used as input for tensor decomposition.

[0142] The module uses the tensor decomposition method to decompose the above data. The core of tensor decomposition is to decompose high-dimensional and multi-dimensional data into low-dimensional structures and extract the potential features of each dimension.

[0143] The input data is represented as a tensor χ, whose dimensions correspond to task completion data, safety operation data, and equipment maintenance data: X(i,j,k), where i represents the miner number, j represents the task type, and k represents the time step. Tensor decomposition is achieved by minimizing the following objective function:

[0144]

[0145] Among them, R is the rank of the potential feature, a i,r , b j,r 、c k,r They are low-dimensional representations of miner characteristics, task characteristics, and time characteristics respectively.

[0146] To ensure that the comprehensive calculation results of performance data reflect the real-time needs of the mine, the module introduces dynamic weighting rules. The weighting rules are dynamically adjusted according to the difficulty of the task, the importance of safe operation and the actual contribution of equipment maintenance. For example, the weight of completion data of more difficult tasks is increased, while the weight of safe operation data accounts for a larger proportion in tasks in high-risk areas. The weight calculation formula is:

[0147]

[0148] Among them, w i is the weight of the i-th data, the contribution factor i Indicates the importance of data under current conditions.

[0149] The comprehensive performance data is generated by combining the tensor decomposition results with the weighting rule, and the formula is:

[0150]

[0151] Among them, S is the comprehensive performance score, X i It is the score of each feature after tensor decomposition.

[0152] The tensor decomposition method ensures accurate feature extraction of performance data, while integrating three-dimensional data of task completion, safe operation and equipment maintenance, improving the comprehensiveness and applicability of the data. Through dynamic weight rules, the performance module can flexibly adjust the focus of performance calculation according to the real-time needs of the mining area, making the performance evaluation closer to the actual production scenario. The comprehensive performance data is directly related to the daily income of miners, which can motivate miners to focus on safe operation and equipment maintenance while completing tasks, thereby improving overall production efficiency and safety.

[0153] Embodiment, mining area application example: In the daily operation of a mining area, the operation process of the performance module is as follows:

[0154] Data collection and processing: Task completion data: miner number 101 completed a 50-ton material transportation task, with a task complexity of 0.8. Safety operation data: miners wore protective equipment throughout the operation, with a safety operation score of 0.95. Equipment maintenance data: miners completed 3 equipment inspections and 1 simple fault repair, with an equipment maintenance score of 0.85.

[0155] Tensor decomposition: The input tensor is χ(i,j,k), and the characteristic decomposition result of miner 101 is:

[0156] a 101 =0.9,b 任务复杂度 =0.8,c 时间段 =0.95

[0157] The comprehensive score is:

[0158] χ 101 =0.9×0.8×0.95=0.684

[0159] The current operation area is a high-risk area, the weight of safe operation is increased to 0.5, the weight of task completion is 0.3, and the weight of equipment maintenance is 0.2.

[0160] The comprehensive score is calculated as: S = (0.3 × 0.684) + (0.5 × 0.95) + (0.2 × 0.85) = 0.831

[0161] The results show that the overall performance score of miner 101 is 0.831, which is used for subsequent profit calculations.

[0162] Preferably, when the weight distribution rule of the performance module is dynamically adjusted, calculations are performed based on the difficulty of task completion, the importance of safe operation, and the actual contribution of equipment maintenance, and the weight distribution is corrected in real time.

[0163] The weight allocation rules define the relative importance of task completion data, safety operation data, and equipment maintenance data in the comprehensive performance calculation. Specifically:

[0164] Task completion data weight (w t ): reflects the importance of task completion efficiency, and its size is determined by the complexity of the task and the priority of the mining area. Tasks with high complexity (such as high-risk and high-intensity tasks) correspond to higher weights.

[0165] Security operation data weight (w s ): reflects the importance of miners' safety behavior, and its size is proportional to the task risk level. For example, when working in areas with high gas concentration or equipment prone to failure, the weight of safe operation is significantly increased.

[0166] Equipment maintenance data weight (wm ): reflects the importance of the miner's equipment maintenance participation, and its size is determined by the proportion of the equipment's contribution to the task completion. For example, in equipment-intensive tasks, maintenance data has a higher weight.

[0167] The weight calculation formula is:

[0168]

[0169] Among them, w i Indicates the weight of the i-th data, contribution factor i It is the contribution degree of each data under the current conditions.

[0170] The dynamic adjustment of weight allocation rules is based on real-time data analysis and environmental assessment. For example, when the difficulty of a task increases (such as the complexity increases from 0.6 to 0.9), the weight of the task completion data automatically increases, while the weight of the equipment maintenance data decreases accordingly to ensure that the performance evaluation matches the task requirements. In high-risk operation scenarios (such as gas concentration increases from 1.0% to 1.5%), the weight of safe operations is dynamically adjusted to the highest, highlighting the core role of safe behavior in performance calculations.

[0171] To ensure the accuracy of weight adjustment, the module monitors the task status and mining conditions in real time and dynamically corrects the weight. The correction logic includes: optimizing weights in real time according to the progress of task completion. For example, when the task completion rate is high but the safety risk increases, the weight of task completion data is automatically reduced and the weight of safe operation data is increased. Dynamically adjust the weight of maintenance data according to the equipment operation status. For example, when the vibration frequency of the equipment is abnormal, increase the weight of equipment maintenance data.

[0172] The comprehensive performance calculation is combined with the dynamically adjusted weight rules, and the formula is:

[0173] S=w t ·T+w s ·S o +w m ·M

[0174] Among them, S represents the comprehensive performance score; T represents the task completion data; S o represents safety operation data; M represents equipment maintenance data; w t 、w s 、w m Represents the dynamically adjusted weight.

[0175] Through dynamic adjustment and real-time correction mechanisms, the weight distribution rules can flexibly adapt to the real-time needs of the mining area, making the performance evaluation closer to the actual operation scenario. The module dynamically adjusts the weights based on task difficulty, safety risks and equipment contribution to ensure that high-difficulty tasks, high-risk operations and equipment-intensive work are reasonably evaluated, thereby improving the fairness of performance calculations. The weight distribution rules highlight the importance of safe operation and equipment maintenance, guiding miners to pay attention to safe behavior and equipment maintenance while completing tasks, further improving the overall efficiency and safety of mining production. Through the real-time correction mechanism, the performance module can dynamically optimize the weight distribution to ensure the accuracy and timeliness of performance calculation results.

[0176] Embodiment, mining area application example: In the daily operation of a mining area, the specific operation of the performance module is as follows:

[0177] Data collection and evaluation, task completion data, miner number 101, completed 50 tons of material transportation task, task complexity is 0.8. Safety operation data, during the operation, the miner wore protective equipment and followed safety regulations, the safety operation score was 0.95. Equipment maintenance data, the miner completed 1 equipment inspection and 1 fault repair, the equipment maintenance score was 0.85.

[0178] Weight calculation and dynamic adjustment: the current operation area is a high-risk area (gas concentration reaches 1.5%), and the weight adjustment is as follows: Task completion data weight (w t ): 0.4 (higher task complexity); safety operation data weight (w s ): 0.5 (high-risk operation area); Equipment maintenance data weight (w m ): 0.1 (the contribution of equipment maintenance tasks is low).

[0179] During the operation, the task completion progress exceeded 80%, but the environmental risk did not decrease. The weight of safe operation data remained unchanged, the weight of task completion data slightly decreased to 0.35, and the weight of equipment maintenance data was adjusted to 0.15.

[0180] Comprehensive performance score calculation, based on dynamically adjusted weights:

[0181] S=(0.35·0.8)+(0.5·0.95)+(0.15~0.85)=0.905

[0182] The overall performance score was 0.905, indicating that the miners performed very well in task completion, safe operation and equipment maintenance.

[0183] The incentive module is used to generate detailed revenue data and performance feedback data based on the current day's revenue data, and to optimize incentive rules and generate incentive strategies by analyzing historical performance data and feedback data.

[0184] The incentive module first displays the miners’ income in three parts based on the daily revenue data, including basic salary, task bonus and additional reward:

[0185] Basic salary: generated according to job salary standards and attendance records, reflecting the fixed income of miners; task bonus: calculated based on the miners' task completion on the day, reflecting the miners' task contribution; additional reward: calculated based on the miners' safety operation scores and equipment maintenance performance, encouraging miners to pay attention to safety and equipment maintenance behaviors. The income details are pushed to the miners' personal terminals through the system port, showing the miners' daily income sources and distribution ratios in a clear way.

[0186] Performance feedback data is a multi-dimensional analysis of the miner's overall performance, including indicators such as task completion efficiency, safe operation performance, and equipment maintenance contribution rate. By comparing these indicators with the miner's historical performance and the team's average level, the module generates intuitive charts or text descriptions to help miners understand the strengths and weaknesses of their own work performance. For example, performance feedback data can be displayed in the form of a bar chart to show the comparison between the miner and the team average, providing direction for the miner's self-improvement.

[0187] The incentive module mines the correlation between miners' behavior and performance by analyzing miners' historical performance data (such as the time series trend of comprehensive performance scores) and miners' feedback data (such as task satisfaction evaluation or safety suggestions). This analysis is based on a causal inference model, which identifies the key factors that affect miners' enthusiasm by constructing a causal network between performance and incentive rules. For example, the analysis may find that an increase in task bonuses can more significantly motivate miners, while additional rewards have a weaker effect on specific task types.

[0188] Based on the results of causal inference, the module dynamically optimizes the incentive rules, including adjusting the weight distribution of benefits and designing new incentives: weight optimization, such as adjusting the weight of task bonuses from 50% to 60%, to strengthen the incentive effect of miners in task completion. Incentives are set to add additional rewards for team collaboration tasks to improve the overall performance of the team.

[0189] The incentive strategy is generated based on the optimized incentive rules, covering the three-layer structure of miner individual incentives, team incentives and global incentives:

[0190] Individual incentives provide direct rewards to miners who complete tasks efficiently or perform outstandingly; team incentives provide collaboration rewards to teams with a higher overall task completion rate; global incentives provide additional rewards when the overall production efficiency of the mining area is significantly improved to stimulate the enthusiasm of all miners.

[0191] Preferably, the incentive module analyzes the relationship between the miner's historical performance data and current feedback data through a causal inference model, and dynamically adjusts the distribution rules of basic income, task bonuses and reward income.

[0192] The incentive module receives and integrates the miners' historical performance data and current feedback data to analyze the impact of the revenue distribution rules on miners' behavior: historical performance data records key indicators such as the miners' comprehensive performance scores, task completion rates, and safety operation scores in different time periods. Current feedback data is feedback information collected through the miner port, such as evaluation of the fairness of task allocation, satisfaction with incentive measures, or suggestions for current revenue rules. After timestamp alignment and preprocessing, these data form an input data stream for causal inference analysis.

[0193] The incentive module analyzes the causal relationship between historical performance data and current feedback data based on the causal inference model, and identifies the key factors affecting the behavior of miners by the profit distribution rules. The core of the model is to simulate the dependency between variables by building a causal relationship diagram, and weigh the impact of different distribution rules on miners' performance.

[0194] Construct a causal relationship graph, where nodes represent the main factors affecting the distribution of benefits (such as task completion rate, safety operation score, equipment maintenance score), and edge weights represent the causal strength between variables. For example, the task completion rate has a higher causal weight on the distribution of task bonuses, while the safety operation score has a higher causal weight on the distribution of bonus benefits.

[0195] Through intervention analysis of the causal path, the impact of the adjustment of a certain profit distribution rule on the overall performance of miners can be quantified. For example, by adjusting the task bonus weight, the improvement of the task completion rate can be predicted. The quantitative formula of the causal effect is:

[0196] ΔY=E[Y|intervention]-E[Y|baseline]

[0197] Among them, ΔY represents the causal effect size, E[Y|intervention] represents the expected value of performance after intervention, and E[Y]baseline] represents the baseline value before intervention.

[0198] According to the results of the causal effect analysis, the module dynamically adjusts the distribution weights of basic income, task bonuses, and reward income: Basic income adjustment, based on the feedback of miners on the fairness of task distribution, appropriately adjust the fixed proportion of basic income. For example, for miners who complete high-intensity tasks for many consecutive days, increase the proportion of basic income. Task bonus adjustment, based on the causal weight of the task completion rate, increase the bonus weight to encourage task efficiency improvement. Reward income adjustment, in high-risk operating environments, increase the weight of reward income on safe operation scores to encourage miners to comply with safety regulations.

[0199] The module optimizes the dynamically adjusted profit distribution rules in real time to ensure the fairness and effectiveness of incentives. For example, when the task bonus weight is too high, causing the enthusiasm of inefficient miners to decrease, the system automatically adjusts the weight to increase the proportion of reward income, thereby balancing the incentive effect.

[0200] The causal inference model can identify the key causal relationship between the profit distribution rules and the miners' behavior, and realize the precise incentives for the miners' behavior. The module can dynamically adjust the profit distribution rules based on historical data and real-time feedback to ensure that the incentive mechanism always adapts to the real-time needs of the mining area. By optimizing the distribution rules, the incentive module can balance the profit distribution between efficient miners and other miners, and improve the miners' recognition of the profit system. The optimized incentive mechanism can stimulate the miners' work enthusiasm and improve the overall task completion rate and safety level of the mining area.

[0201] Embodiment, mining area application example: In the daily operation of a mining area, the operation process of the incentive module is as follows:

[0202] Data collection and causal analysis, historical performance data: The historical task completion rate of miner No. 101 is 85%, the safety operation score is 0.9, and the equipment maintenance score is 0.8. Current feedback data: Miners reported that the task bonus ratio is too low, and it is recommended to increase the reward for efficient task completion. Causal analysis results: Through causal relationship modeling, it is found that the causal weight of task completion rate to task bonus is 0.6, the causal weight of safety operation score to reward income is 0.8, and the causal weight of equipment maintenance score to basic income is 0.4.

[0203] The profit distribution rules are adjusted dynamically. According to the results of causal analysis, the module adjusts the profit distribution weights: the task bonus weight is increased from 50% to 60% to strengthen the incentive for task completion; the reward profit weight is increased from 20% to 30% to highlight the incentive for high-risk operational safety behaviors; the basic profit weight is reduced from 30% to 10% to reduce reliance on fixed income.

[0204] Optimize and apply incentive rules. The adjusted incentive rules generate new profit distribution strategies:

[0205] Task bonus: The bonus for the miner's completed 50-ton material transportation task is calculated according to the new weight, with a total bonus of 600 yuan (12 yuan per ton).

[0206] Reward income: Based on the miner’s safety operation score (0.9), the reward income is calculated to be 135 yuan (150 yuan per unit of reward based on the score).

[0207] Basic income: Calculated according to the new weights, the basic income is 100 yuan. Comprehensive income distribution: 600 yuan (task bonus) + 135 yuan (reward income) + 100 yuan (basic income) = 835 yuan.

[0208] Preferably, the incentive module further generates a multi-level incentive strategy, including individual rewards based on individual comprehensive performance, team rewards based on team task completion rate, and global rewards based on resource optimization of the entire mining area.

[0209] Individual rewards are calculated based on the miners’ comprehensive performance data, aiming to directly motivate miners’ task completion efficiency, safe operation performance and equipment maintenance contribution rate. i ) and weight distribution rule (w t ,w s ,w m ) Calculate personal rewards:

[0210] R i =S i ·B

[0211] Among them, R i represents the personal reward of miner i; S i It represents the comprehensive performance score, which is calculated based on the three-dimensional data of task completion, safe operation and equipment maintenance; B represents the reward benchmark value, which is determined by the daily revenue budget of the mining area and the difficulty of the task.

[0212] Example: The comprehensive performance score of miner No. 101 is 0.9, and the reward base value is set to 500 yuan. Then his personal reward is:

[0213] R 101 =0.9·500=450 yuan

[0214] Team rewards are generated based on the team task completion rate, focusing on motivating team members to collaborate and achieve overall task goals. r ) and team reward benchmark (B t ) calculate the total amount of team rewards and distribute it to team members based on individual contribution rates (C i ) for weighted allocation:

[0215]

[0216] Example: The task completion rate of a transportation team is 85%, the team reward base is 1,000 yuan, the team members are numbered 101, 102 and 103, and their contribution rates are 0.4, 0.35 and 0.25 respectively. The team reward distribution is:

[0217] Miner 101 Team Rewards:

[0218]

[0219] The global reward is generated based on the resource utilization rate and task completion rate of the entire mining area, aiming to improve the overall production efficiency of the mining area. The module calculates the global reward pool based on the resource utilization rate (U) and task completion rate (C) of the entire mining area, and distributes it to each team based on the team's task completion contribution. The global reward formula is:

[0220] R g =U·C·B g

[0221] Among them, R g represents the total amount of the global reward pool; U represents the resource utilization rate of the entire mining area; C represents the task completion rate of the entire mining area; B g It represents the base value of rewards for the entire mining area, which is determined by the production budget of the mining area.

[0222] Example: The resource utilization rate of a mining area is 0.8, the task completion rate is 0.9, and the reward base for the entire mining area is 5,000 yuan. Then the global reward pool is:

[0223] R g =0.8·0.9·5000=3600 yuan

[0224] The global reward is distributed according to the task completion contribution rate of each team. The contribution rates are 50%, 30% and 20% respectively. The distribution results are:

[0225] Global rewards for Team 1:

[0226] R 1,g =0.5·3600=1800 yuan

[0227] The multi-level incentive strategy covers the individual, team and global aspects, which can effectively stimulate the individual enthusiasm of miners, promote team collaboration, and improve the overall production efficiency of the mining area. The module ensures that the incentive measures accurately match the actual work performance through the analysis of comprehensive performance data, team task completion rate and mining area resource utilization rate. The module dynamically adjusts the incentive strategy according to the real-time production situation of the mining area to ensure that the incentive rules always meet the needs of the mining area. By making the incentive strategy public and the allocation rules clear, the miners' recognition and trust in the incentive mechanism are improved.

[0228] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware.

[0229] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A performance management and incentive system for underground miners based on real-time salary settlement, characterized in that: include: The acquisition module is used to obtain the physiological parameters of the miners as the miner status data, obtain the working environment conditions as the environment status data, and obtain the equipment operation status as the equipment status data; The quota generation module is used to clean and normalize the miner status data, environmental status data and equipment status data, extract key features through principal component analysis, generate characteristic variable data, and generate dynamic task quota data through nonlinear modeling in combination with the miner's historical task data; The task management module is used to collect miners' task status data based on dynamic task quota data, and integrate task progress data and equipment operation data to generate real-time comprehensive data; it is equipped with a task prediction submodule to predict task progress through real-time comprehensive data, generate dynamic task allocation data and adjusted task plan data; A performance module is used to generate comprehensive performance data based on dynamic task allocation data and adjusted task plan data, wherein the comprehensive performance data is comprehensively calculated for task completion data, safety operation data and equipment maintenance data through a weight allocation rule, and the daily revenue data is calculated accordingly; The incentive module is used to generate detailed revenue data and performance feedback data based on the current day's revenue data, and to optimize incentive rules and generate incentive strategies by analyzing historical performance data and feedback data.

2. The underground miner performance management and incentive system based on real-time salary settlement according to claim 1 is characterized in that: The acquisition module specifically includes: The heart rate and body temperature data of miners are obtained in real time through wearable devices to generate data streams of the miners’ current status; The temperature, humidity, air pressure and gas concentration data of the working area are collected through environmental monitoring equipment to generate real-time status data of the working environment; The equipment monitoring module collects the equipment operation time, power and fault records in real time to generate data input on the current performance of the equipment.

3. The underground miner performance management and incentive system based on real-time salary settlement according to claim 1 is characterized in that: The quota generation module extracts characteristic variables with strong nonlinear correlation from miner status data, environmental status data and equipment status data through characteristic variable extraction technology, uses principal component analysis to reduce data dimensions and improve modeling efficiency, and at the same time generates dynamic task quota data based on the miners' historical task completion status for real-time adjustment of task allocation strategies.

4. The underground miner performance management and incentive system based on real-time salary settlement according to claim 3 is characterized in that: The quota generation module dynamically adjusts the generation process of dynamic task quota data by constructing a multi-objective optimization model, comprehensively considering the current status of miners, operating environment risks and equipment operation reliability, and ensuring that the task quota allocation matches the actual operating capacity.

5. The underground miner performance management and incentive system based on real-time salary settlement according to claim 1 is characterized in that: The task management module includes: The task status collection submodule is used to collect the task completion amount, operation time and location information of miners during the task execution process; The real-time comprehensive data generation submodule is used to integrate the miner's task status data with the task progress data and the equipment operation data to generate real-time comprehensive data that can be used for task allocation optimization; The task prediction submodule is used to predict task progress and detect potential task conflicts based on real-time comprehensive data, and generate input data for optimal task allocation.

6. The underground miner performance management and incentive system based on real-time salary settlement according to claim 5 is characterized in that: The task management module dynamically optimizes the task dependencies in the real-time comprehensive data through a multi-agent reinforcement learning method, combines unfinished tasks with the current miner status, and generates task allocation data and adjusted task planning data.

7. The underground miner performance management and incentive system based on real-time salary settlement according to claim 1 is characterized in that: The performance module performs feature decomposition on the task completion data, safety operation data and equipment maintenance data through the tensor decomposition method, and performs weighted comprehensive calculation on each data in combination with dynamically adjusted weight rules to generate the miners' daily comprehensive performance data.

8. The underground miner performance management and incentive system based on real-time salary settlement according to claim 7 is characterized in that: When the weight distribution rule of the performance module is dynamically adjusted, calculations are performed based on the difficulty of task completion, the importance of safe operation, and the actual contribution of equipment maintenance, and the weight distribution is corrected in real time.

9. The underground miner performance management and incentive system based on real-time salary settlement according to claim 1 is characterized in that: The incentive module analyzes the relationship between miners' historical performance data and current feedback data through a causal inference model, and dynamically adjusts the distribution rules of basic income, task bonuses and reward income.

10. The underground miner performance management and incentive system based on real-time salary settlement according to claim 9 is characterized in that: The incentive module further generates a multi-level incentive strategy, including individual rewards based on individual comprehensive performance, team rewards based on team task completion rate, and global rewards based on resource optimization of the entire mining area.

Images