Material charging intelligent scheduling and data transmission method based on ai

Abstract

The invention provides an ai-based material charging intelligent scheduling and data transmission method, which comprises the following steps: acquiring post responsibility adjustment data, analyzing responsibility description and historical material use records, determining an update set of common material combinations, grouping materials by adopting a clustering method, and obtaining feature vectors of combination reconstruction changes; obtaining a real-time mapping relation between post responsibility adjustment and material demands according to the feature vectors of the combination reconstruction change, predicting use duration distribution in combination with the use condition of the charging cabinet, and obtaining trend parameters of material use behaviors; and according to the charging cabinet position configuration scheme and the adjusted charging cabinet position distribution logic, a preliminary resource distribution strategy based on the charging cabinets is adjusted in combination with the change trend of the electric quantity residual rule, and a charging cabinet resource grouping scheme is obtained.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to an AI-based intelligent scheduling and data transmission method for material charging. Background Art

[0002] With the acceleration of enterprise digital transformation, material management, a core component of operational efficiency, has a direct impact on production and service quality. Intelligent material classification and optimized counter allocation based on usage scenarios can significantly improve material access efficiency and resource utilization, and has become a key research direction in enterprise management. However, existing material management methods often rely on static allocation or manual experience, making them difficult to adapt to dynamic demand changes brought about by adjustments in work content or organizational structure. These methods often ignore the correlation between material usage scenarios—the dynamic mapping between job responsibilities and required material combinations. This leads to a lack of flexibility in counter allocation and charging strategies that fail to accurately match actual needs, resulting in inconvenient material access and wasted resources. Specifically, current solutions have significant limitations in handling dynamic changes in material combination characteristics. For example, when job responsibilities adjust and frequently used material combinations are restructured, existing systems struggle to update counter allocation logic in real time. Changes in the distribution of material usage duration and the patterns of residual power upon return also lead to inefficient charging resource allocation. These issues stem from insufficient analysis of material usage correlations and a lack of dynamic optimization algorithm support. The core challenges are concentrated in the following technical factors: first, dynamic correlation modeling of material usage scenarios, how to capture the real-time mapping relationship between changes in job responsibilities and material needs; second, dynamic adjustment of cabinet allocation, how to optimize the storage location of materials based on usage frequency and combination characteristics; finally, intelligent grouping of charging strategies, how to achieve efficient resource utilization based on the residual power pattern and usage time distribution. These technical factors have not been resolved, resulting in reduced material access efficiency and waste of charging resources, and innovative algorithms are urgently needed to address them. Therefore, how to design an AI configuration optimization algorithm based on usage correlation to dynamically adjust the storage location of materials in the cabinet and the charging strategy grouping to adapt to the reconstruction of material combinations, changes in usage time distribution, and changes in residual power patterns brought about by work content or organizational structure adjustments has become a key issue in improving material management efficiency and resource utilization. Summary of the Invention

[0003] The present invention provides an AI-based intelligent scheduling and data transmission method for material charging, which mainly includes:

[0004] Obtain job responsibility adjustment data, determine the updated set of commonly used material combinations by analyzing job descriptions and historical material usage records, use clustering methods to group materials, and obtain the characteristic vector of combination reconstruction changes;

[0005] Based on the combined reconstructed characteristic vectors, the real-time mapping relationship between job responsibility adjustment and material demand is obtained. Combined with the usage of the charging cabinet, the usage time distribution is predicted to obtain the trend parameters of material usage behavior.

[0006] Extract the peak period of material use as the key time node from the trend parameters of the usage time distribution, analyze the correlation coefficient between power consumption and usage time, establish a preliminary resource allocation strategy based on charging cabinets, and determine the initial grouping basis for charging resource allocation;

[0007] By combining and reconstructing the changing feature vector and the trend parameter of the usage time distribution, the storage location of materials in the charging smart cabinet is optimized to adjust the preset charging cabinet cabinet allocation logic, and the adjusted charging cabinet cabinet allocation logic is obtained;

[0008] Combined with the charging cabinet return records, the remaining power pattern is analyzed. Based on the initial grouping basis of charging resource allocation, the charging devices are grouped. The charging priority and charging time threshold of each group of charging devices are determined according to the charging cabinet situation, and the charging cabinet position configuration plan is obtained.

[0009] According to the charging cabinet location configuration plan and the adjusted charging cabinet location allocation logic, combined with the changing trend of the residual power law, the initial resource allocation strategy based on the charging cabinet is adjusted to obtain the charging cabinet resource grouping plan;

[0010] Real-time feedback data on the efficiency of charging cabinet material access is obtained, and the modeling parameters of the correlation coefficient between power consumption and usage time are adjusted using an iterative optimization method to determine the optimization results of the charging cabinet position configuration. By predicting the long-term changing trends of usage time distribution and residual power rules, the dynamic correlation modeling of charging cabinet material usage and power consumption is updated to obtain a continuously optimized material management logic based on charging cabinets.

[0011] Furthermore, job responsibility adjustment data is obtained. By analyzing job descriptions and historical material usage records, an updated set of commonly used material combinations is determined. The materials are then grouped using a clustering method to obtain a feature vector representing the combination reconstruction change. This process involves obtaining job responsibility adjustment records from the human resources management database, comparing the text of the job descriptions before and after the adjustment, calculating the magnitude of the change in responsibilities using a text comparison algorithm, and generating a job responsibility change matrix. Based on the responsibility change matrix, a material requisition database is queried to extract the material requisition records corresponding to each position. The frequency and quantity of material requisitions are then calculated to generate a job-material association matrix. The job-material association matrix is then normalized, extracting the material usage interval and requisition quantity as feature variables. Density clustering is then used to group the normalized data to generate a material grouping set. Equipment operation data is retrieved from the charging cabinet equipment database, extracting the charging cabinet number, material storage capacity, and material access records, and establishing a charging cabinet material storage status table. Based on the material grouping set and the charging cabinet material storage status table, the frequency of material combination adjustments is calculated. If the adjustment frequency exceeds a preset threshold, the combination is marked as a reconstruction target. For the marked reconstruction objects, we extract resource access time series data and calculate the characteristic values of material combination changes, including three indicators: combination stability, adjustment range, and correlation strength. We use these characteristic values to construct a feature vector, divide the reconstruction objects into new material combinations, and generate a material combination reconstruction plan.

[0012] Furthermore, based on the characteristic vectors of the combined reconstruction changes, a real-time mapping relationship between job responsibility adjustments and material requirements is obtained. Combined with the charging cabinet usage, the usage duration distribution is predicted to obtain trend parameters for material usage behavior. This process involves obtaining the characteristic vectors of the combined reconstruction changes from the material management database, calculating the ratio of average daily material consumption to inventory to obtain a material shortage warning index, counting the number of consecutive days of material withdrawal to obtain a material withdrawal continuity value, and generating a mapping relationship table between material combinations and job responsibility adjustments using an association rule mining algorithm. Based on this mapping relationship table, the number of responsibility changes is extracted to obtain a responsibility adjustment triggering index. The response time from responsibility adjustments to material requirements is calculated to obtain a job adjustment timeliness index, and a corresponding table between job responsibilities and material requirements is generated. Charging cabinet occupancy time records are read from the charging cabinet management database, combined with material access timestamps to generate material usage time series data. The material access frequency within each time window is calculated to generate a usage intensity curve. The usage intensity curve is sampled according to a fixed time window, and data noise is removed using a sliding average method to generate a charging cabinet usage duration baseline curve. Based on the benchmark curve of charging cabinet usage duration, a time series decomposition method is used to extract periodic and trend characteristics, and a time series prediction model for material usage behavior is constructed. The change rate of material usage frequency and duration is extracted from the output of the time series prediction model, and the material usage behavior trend parameters are calculated.

[0013] Furthermore, peak periods of material usage are extracted from the trend parameters of the usage duration distribution as key time nodes. The correlation coefficient between power consumption and usage duration is analyzed to establish a preliminary resource allocation strategy based on charging cabinets and determine the initial grouping criteria for charging resource allocation. This strategy involves extracting the peak values of the usage duration distribution curve from the material usage behavior trend parameters, dividing the material return records into time windows, and generating a characteristic sequence for peak material usage periods. Material return records are read from the charging cabinet controller, and the device number, return time, and remaining power value are extracted. The power loss rate per unit time is calculated to generate a power consumption characteristic curve. Based on the power consumption characteristic curve, a statistical regression function is constructed between usage duration and power loss, with usage duration as the independent variable and power loss as the dependent variable, to generate a power consumption prediction function. Density clustering is performed on the characteristic sequence of peak material usage periods, and the remaining capacity and occupancy duration of the charging cabinets during each period are calculated to generate a resource occupancy density distribution map. Based on the resource occupancy density distribution map, the remaining power and capacity data of the charging cabinets are extracted, and a resource allocation weight calculation function is constructed to generate a priority sequence for charging cabinet resource allocation. According to the resource allocation priority sequence of charging cabinets and combined with the power consumption prediction function, the resource allocation parameters of charging cabinets in different time periods are calculated to generate the initial grouping plan for charging resource allocation.

[0014] Furthermore, by combining the reconstructed change eigenvectors with the trend parameters of the usage time distribution, the storage location of materials in the smart charging cabinet is optimized to adjust the preset charging cabinet slot allocation logic, resulting in the adjusted charging cabinet slot allocation logic. This method includes: reading the charging cabinet slot allocation logic from the smart cabinet management database; extracting the combined reconstructed change eigenvectors and the usage time distribution trend parameters based on the material number and slot number association table; and constructing a material storage location constraint matrix. The material storage location constraint matrix is numerically quantified, and the spatial size adaptability and location proximity of each slot are calculated to generate a set of slot space utilization indicators. Based on the slot space utilization indicator set, the access frequency and access time interval are extracted from the material access records to construct a material storage location allocation fitness calculation function. Based on the material storage location allocation fitness calculation function, a chromosome coding sequence is generated, in which each gene bit in the chromosome represents the correspondence between materials and slots. The stability parameter of the allocation scheme is set as the termination condition of the genetic algorithm, and crossover and mutation operations are performed on the chromosome sequence to generate a set of optimized material storage location solutions. Based on the set of optimized storage location solutions, the cabinet layout rationality score and adjustment cost score of each solution are calculated, and the solution with the highest overall score is selected. Based on the selected optimization solution, the mapping relationship between the material cabinets is extracted, the charging cabinet cabinet allocation logic database is updated, and a new cabinet allocation rule set is generated.

[0015] Furthermore, the remaining battery capacity patterns are analyzed in conjunction with charging cabinet return records. Based on the initial grouping criteria for charging resource allocation, charging devices are grouped. Charging priorities and charging duration thresholds are determined for each group of charging devices based on the charging cabinet's conditions. This results in a charging cabinet configuration plan. This plan involves reading the latest remaining battery capacity distribution data from the charging cabinet database, extracting the remaining battery percentage, power loss rate per unit time, and average daily usage times for each device, and constructing a feature dataset for the charging devices. The feature dataset is normalized, and the mean and standard deviation of each feature are calculated to generate a standardized device feature vector. Based on the clustering distance calculation results for the device feature vectors, the number of cluster centers is set to one-fifth of the total number of devices, and the devices are grouped using the K-means clustering method. For each clustered device group, the average remaining battery capacity and average power consumption rate are calculated within the group. A device priority ranking table is generated, sorting the remaining battery capacity in descending order and the power consumption rate in ascending order. The charging cabinet's cabinet dimensions and interface type data are obtained from the charging cabinet controller, and a cabinet adaptation matrix is generated based on the matching relationship between device dimensions and interfaces. Based on the cabinet adaptation matrix and the device distribution in each group, the resource allocation ratio for each charging cabinet is calculated to generate a charging resource allocation table. Based on this table, upper and lower charging time thresholds are set for each device group to generate a grouped charging parameter configuration table. Using this table, the charging cabinet spaces are divided according to the device priority table to generate a charging cabinet configuration plan.

[0016] Furthermore, based on the charging cabinet slot configuration plan and the adjusted slot allocation logic, and incorporating the changing trends of residual power, the initial resource allocation strategy for the charging cabinets is adjusted to obtain a charging cabinet resource grouping plan. This involves reading the optimized slot configuration plan and allocation logic from the charging cabinet management database, extracting the slot number, device type, and charging status as state features, and constructing a charging cabinet resource allocation state space. Based on this state space, the slot idle rate and device waiting rate are calculated, and resource utilization scores and waiting time scores are calculated to generate a resource allocation reward metric function. A resource scheduling agent is constructed using deep reinforcement learning, with the slot allocation action space defined as consisting of two dimensions: device selection and slot assignment. This generates a resource allocation strategy network. Based on the resource allocation strategy network, the slot allocation plan within each time segment is evaluated, the reward value is calculated, and the network parameters are updated to obtain an optimized resource allocation decision function. A sliding time window is set to the length of a single work shift, slot usage records are collected within the window, and a dynamic allocation accuracy metric is calculated. Based on the dynamic allocation accuracy metric, the resource allocation decision function is fine-tuned online, and the allocation strategy parameters are optimized using a gradient update method. Based on the optimized allocation strategy parameters, a dynamic grouping rule base for charging cabinet resources is constructed, and a mapping table is established between device types and cabinet groupings. The grouping rule base is continuously updated using online learning, and the grouping scheme is dynamically adjusted based on real-time monitoring of resource utilization, generating a dynamic grouping scheme that efficiently utilizes charging cabinet resources.

[0017] Furthermore, real-time feedback data on charging cabinet material access efficiency is obtained. Modeling parameters for the correlation coefficient between power consumption and usage duration are adjusted using an iterative optimization method to determine the optimized charging cabinet position configuration. By predicting long-term trends in usage duration distribution and residual power patterns, the dynamic correlation model between charging cabinet material usage and power consumption is updated, resulting in a continuously optimized charging cabinet-based material management logic. This includes: reading real-time feedback data on material access efficiency from the charging cabinet controller, extracting material access intervals, access frequencies, and remaining power, and constructing a material usage feature sequence. This feature sequence is then decomposed into time series, and the Pearson correlation coefficient between power consumption rate and usage duration is calculated to generate a parameter matrix correlating power consumption and usage duration. Based on this correlation parameter matrix, the parameter update step size and the upper limit of the number of iterations are set. The parameters are then iteratively optimized using gradient descent over multiple rounds to obtain the optimized cabinet position configuration parameters. Based on the optimized cabinet position configuration parameter results, a material usage behavior prediction model is constructed to extract usage duration distribution characteristics and power consumption patterns. Periodic analysis of the usage duration distribution characteristics is performed, and a time series decomposition method is used to identify long-term trends and seasonal variations. Based on power consumption patterns, the power loss rate per unit time and cumulative usage duration are calculated to establish a power consumption prediction function. Based on the usage duration prediction results and the power consumption prediction function, a dynamic mapping table is constructed between material usage and power consumption. Dynamic programming methods are used to update this mapping table online, generating continuously optimized material management logic for charging cabinets.

[0018] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0019] The present invention discloses an AI-based intelligent scheduling and data transmission method for material charging. By analyzing job responsibility adjustments and material usage records, a characteristic vector of material combination reconstruction changes is established, and the usage time distribution is predicted in combination with the charging cabinet usage. The peak period is extracted and the residual power pattern is analyzed to optimize the charging cabinet position allocation logic. The present invention adopts a clustering method to group charging equipment, determine the charging priority and time threshold, and dynamically adjust the resource allocation strategy using a reinforcement learning algorithm to achieve efficient utilization of charging cabinet resources. By iteratively optimizing the correlation coefficient between power consumption and usage time, predicting long-term change trends, and continuously updating the dynamic correlation modeling of material usage and power consumption, the intelligent and refined management of charging cabinet materials is achieved, and the efficiency of resource utilization is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a flow chart of an AI-based intelligent scheduling and data transmission method for material charging of the present invention. DETAILED DESCRIPTION

[0021] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0022] As the core equipment for material storage and charging, the charging cabinet is responsible for storing and charging various materials, such as tools, equipment or consumables. The material usage scenario is usually highly related to job responsibilities. Adjustments to job responsibilities may lead to changes in the combination of commonly used materials, which in turn affects the access frequency and charging needs of materials. Existing material management methods are mostly static configurations, which are difficult to dynamically adapt to these changes, resulting in inefficient cabinet allocation and waste of charging resources. The embodiment of the present invention uses an artificial intelligence algorithm to dynamically optimize cabinet allocation and charging strategies based on the correlation of material usage to improve management efficiency.

[0023] The following terminology explains: A material package refers to a collection of materials organized according to job responsibilities and usage scenarios, such as wrenches and lubricants commonly used in mechanics. A charging cabinet refers to a smart device with material storage and charging capabilities, containing multiple storage compartments and charging ports. Usage association refers to the dynamic mapping between job responsibilities, material usage frequency, and power consumption.

[0024] The embodiments of the present invention do not impose excessive restrictions on the equipment type or material type of the charging cabinet, which can be determined according to the actual scenario. For example, the charging cabinet can be fixed or mobile, and the materials can include tools, batteries or other consumables.

[0025] Figure 1 The following is a flow chart showing the method for intelligent scheduling and data transmission of material charging based on artificial intelligence provided by an embodiment of the present invention. The specific steps are as follows:

[0026] S101. Obtain job responsibility adjustment data, analyze job descriptions and historical material usage records, determine an updated set of commonly used material combinations, and use a clustering method to generate a combination reconstruction feature vector related to the charging cabinet.

[0027] In an embodiment of the present invention, the charging cabinet has a material management function. This function can be integrated into the control system of the charging cabinet or implemented through an external server. The specific implementation method is determined according to the actual scenario. The activation method of the material management function is also not strictly limited. It can be triggered by the user through the charging cabinet operation interface or activated by remote instructions from the enterprise management system. For example, the user can select the material management mode through the touch screen of the charging cabinet, or send a scheduling instruction to activate the function through the enterprise resource planning system.

[0028] After the material management function is activated, the charging cabinet first obtains the job responsibility adjustment records from the enterprise human resources management database to determine the potential impact of the change in responsibilities on material needs.

[0029] S1011. Obtain job responsibility adjustment records from a human resource management database, compare the job description text content before and after the adjustment using a text analysis algorithm, and obtain a job responsibility change matrix.

[0030] In an embodiment of the present invention, the job responsibility adjustment record includes information such as the job number, adjustment time, and job description text. The charging cabinet extracts the latest record through the database interface, and uses a text analysis algorithm, such as an algorithm based on word segmentation and cosine similarity, to compare the text content of the job description before and after the adjustment, and calculates the magnitude of the job change. For example, when a job responsibility is changed from "equipment maintenance" to "equipment monitoring and fault diagnosis", the algorithm calculates that the job change magnitude is 0.62, indicating that the job has been significantly adjusted. Based on this magnitude, a job responsibility change matrix is generated, in which the matrix elements represent the degree of change in each job responsibility, which is used for subsequent correlation analysis of material requirements.

[0031] S1012. Query the material requisition database based on the job responsibility change matrix, extract the material requisition records corresponding to the job, count the requisition frequency and quantity, and generate a job material association matrix.

[0032] In this embodiment of the present invention, the charging cabinet extracts material requisition records for corresponding positions from the material requisition database based on the job responsibility change matrix. This information includes information such as the requisitioner's position number, material number, time of requisition, and quantity. Statistical analysis is used to calculate the monthly average requisition frequency and quantity of each material. For example, a mechanic's position requisitions 5 barrels of lubricating oil and 10 oil filters per month, generating a position-material association matrix. Matrix elements represent the strength of the association between materials and positions, providing a data foundation for updating material combinations.

[0033] S1013. Preprocess the data of the job material association matrix, use the density clustering method to group the materials, generate a material grouping set, and calculate the combination adjustment frequency based on the storage status of the charging cabinet materials to generate a material combination reconstruction plan.

[0034] To ensure accurate data analysis, in this embodiment of the present invention, the job-specific material association matrix is standardized, extracting the material usage interval and the quantity issued as characteristic variables. Density clustering methods, such as the DBSCAN algorithm, are used to group the standardized data. Materials with similar usage characteristics are grouped together to form a material grouping set. For example, lubricating oil and oil filters are grouped together due to their similar usage intervals and issued quantities.

[0035] At the same time, the charging cabinet reads operational data from the equipment database, including the charging cabinet number, storage compartment capacity, and material access records, and establishes a material storage status table. Based on the material grouping set and the storage status table, the frequency of material combination adjustment is calculated. If the frequency exceeds a preset threshold, such as 10 times per month, the combination is marked as a reconstruction object. For the reconstruction object, the material access time series data of the past 180 days is extracted, and characteristic values such as combination stability, adjustment amplitude, and correlation strength are calculated to construct a three-dimensional feature vector. This feature vector is used to re-divide the material combination, for example, frequently used wrenches and sockets are classified into the same combination, and a material combination reconstruction plan is generated.

[0036] In this embodiment of the present invention, through the above steps, the charging cabinet can dynamically update the material combination based on job responsibilities and generate a combination reconstruction feature vector related to the charging cabinet, providing a basis for subsequent cabinet allocation and charging strategy optimization. This achieves adaptive adjustment of material management, significantly improves access efficiency, and reduces configuration lag caused by changes in responsibilities.

[0037] S102. Based on the combined reconstruction feature vector related to the charging cabinet, a real-time mapping relationship between job responsibility adjustment and material demand is established. Combined with the usage of the charging cabinet, the material usage time distribution is predicted through time series analysis to obtain the trend parameters of the material usage behavior.

[0038] In this embodiment of the present invention, once a material combination reconstruction plan is generated, the charging cabinet uses the combined reconstruction feature vector to analyze the impact of job responsibility adjustments on material demand, establish a real-time mapping relationship, and predict material usage trends through time series analysis, providing data support for cabinet optimization and charging strategy. The material management function is implemented by the charging cabinet control system, which can be implemented through local computing or cloud servers, depending on the actual scenario.

[0039] S1021. Obtain the combined reconstruction feature vector from the material management database, calculate the dynamic ratio of material consumption to inventory, generate a material shortage warning index, and use an association rule mining algorithm to construct a mapping relationship table between job responsibilities and material requirements.

[0040] In an embodiment of the present invention, the charging cabinet extracts a combination reconstruction feature vector from the material management database, which contains indicators such as material combination stability, adjustment range, and correlation strength. Based on the feature vector, the ratio of the average daily consumption of each material to the current inventory is calculated to generate a material shortage warning index. For example, a machine repair tool set has an average daily consumption of 4 screwdrivers and an inventory of 25, resulting in a shortage warning index of 0.16. If the index exceeds the preset threshold of 0.75, a material replenishment warning is triggered, prompting the management system to restock in a timely manner.

[0041] To establish a correlation between responsibilities and requirements, the charging cabinet uses association rule mining algorithms, such as the Apriori algorithm, to process the material shortage warning index and material request records to generate a mapping table between job responsibilities and material combinations. This table records information such as the number of responsibility changes, material demand response time, and the strength of the correlation. For example, if equipment maintenance job responsibilities are adjusted twice per month and the material demand response time is 18 hours, the correlation in the mapping table is 0.78, indicating that the responsibility adjustment significantly affects material demand.

[0042] S1022. Extract the responsibility adjustment trigger degree and timeliness index according to the mapping relationship table, generate a corresponding table of job responsibilities and material requirements, obtain the occupancy time record from the charging cabinet management database, and generate material usage time series data.

[0043] In an embodiment of the present invention, the charging cabinet counts the frequency of responsibility changes based on the mapping relationship table and calculates the trigger degree of responsibility adjustment. For example, a maintenance position has four responsibility adjustments within 60 days, and the trigger degree is 0.067, indicating that the job responsibilities are highly dynamic. At the same time, the time interval from the responsibility change to the material demand response is calculated to generate a timeliness index. For example, if the material configuration is completed within 12 hours after the responsibility adjustment, the timeliness index is 1.2, reflecting a high response efficiency. Based on the trigger degree and timeliness index, a correspondence table between job responsibilities and material demands is generated to guide material allocation.

[0044] The charging cabinet reads the charging cabinet occupancy time records from the management database and combines them with the timestamps of material access to generate material usage time series data. For example, a charging cabinet records 50 material access operations on a weekday, with an average of one interaction every 9.6 minutes. The timestamp contains the material number, access time, and operator information, providing a basis for subsequent time series analysis.

[0045] S1023. Sampling is performed based on the time series data of material usage to generate a usage intensity curve. The time series decomposition method is used to extract periodicity and trend characteristics, and a material usage behavior prediction model is constructed to obtain trend parameters.

[0046] In this embodiment of the present invention, the charging cabinet samples material usage time series data in fixed time windows, such as 30 minutes, and calculates the access frequency within each window to generate a usage intensity curve. The curve shows peak material usage periods, such as when the access frequency reaches 15 times / hour between 8:00 AM and 10:00 AM on weekdays.

[0047] To eliminate data fluctuations, the charging cabinet uses a sliding average method, such as a 60-minute window, to smooth the usage intensity curve and generate a usage duration baseline curve. For example, the baseline usage duration is 7 hours on weekdays, which drops to 3.5 hours on weekends. Based on the baseline curve, the charging cabinet uses time series decomposition methods, such as STL decomposition, to extract periodic features, such as the daily peak hours of 9:00 to 11:00 a.m. and 1:00 to 3:00 p.m., as well as trend features, such as the monthly growth rate of 0.12 in the frequency of machine repair tool usage.

[0048] Based on the decomposition results, the charging cabinet constructs a material usage behavior prediction model, such as an ARIMA model, which inputs periodic and trend characteristics to predict the distribution of usage duration over the next 30 days. The model outputs the rate of change in material usage frequency and duration. For example, the frequency change rate for general-purpose tools is 0.18, and the duration change rate is 0.09; the frequency change rate for specialized equipment is 0.25, and the duration change rate is 0.15. Combining these indicators, trend parameters for material usage behavior are generated for subsequent optimization.

[0049] In this embodiment of the present invention, through feature vector analysis and time series prediction, the charging cabinet can accurately capture the dynamic correlation between job responsibility adjustments and material demand, and predict usage trends. Compared with traditional static methods, this embodiment of the present invention improves the flexibility of material allocation, reduces the problem of material shortages or surpluses caused by demand fluctuations, and provides a reliable basis for the dynamic management of charging cabinets.

[0050] S103. Extract peak usage period characteristics based on material usage behavior trend parameters, analyze power consumption patterns in combination with charging cabinet return records, build a power consumption prediction model, and establish an initial grouping plan for charging resource allocation through resource occupancy analysis.

[0051] In this embodiment of the present invention, the charging cabinet utilizes trend parameters for material usage to identify peak periods, analyze the correlation between power consumption and usage duration, and construct a predictive model. Furthermore, the model optimizes charging resource allocation through resource occupancy density analysis, providing a basis for dynamic scheduling. The material management function is implemented by the charging cabinet control system, which can be implemented through local or cloud computing, depending on the specific scenario.

[0052] S1031. Extract the peak characteristics of the usage duration distribution from the material usage behavior trend parameters to generate a material usage peak period sequence, and calculate the power loss rate in combination with the charging cabinet return record to generate a power consumption characteristic curve.

[0053] In an embodiment of the present invention, the charging cabinet extracts the peak value of the usage duration distribution curve from the material usage behavior trend parameters to identify the peak material usage periods. For example, analysis shows that the main peak hours are 7:00 to 9:00 a.m. and 1:00 to 15:00 p.m. on weekdays, with access frequencies reaching 30 times / hour and 24 times / hour, respectively. By dividing the data into 30-minute time windows, a sequence of peak material usage periods is generated, recording the access frequency and duration of each period, providing a time basis for resource allocation.

[0054] The charging cabinet obtains material return records from the controller database, which include information such as the device number, return time, and remaining power. For example, the charging cabinet numbered CB-201 recorded 85 return operations. An electric screwdriver was returned after 60 minutes of use, and the remaining power dropped from 100% to 40%. The power loss rate per unit time was calculated to be 1.0% / minute. Based on the return records, the power loss of various types of materials is counted to generate a power consumption characteristic curve that reflects the consumption patterns of different materials during peak hours. For example, the average single use time of an electric wrench is 80 minutes, with a power loss of 50%, while the use time of a testing instrument is 150 minutes, with a power loss of 70%.

[0055] S1032. Build a statistical regression model based on the power consumption characteristic curve to generate a power consumption prediction function, and perform density clustering on the material usage peak period sequence to generate a resource occupancy density distribution.

[0056] In an embodiment of the present invention, the charging cabinet utilizes the power consumption characteristic curve and employs a linear regression method to construct a power consumption prediction model. The independent variable is the duration of use of the material, and the dependent variable is the power loss value. For example, regression analysis shows that the power consumption of power tools has a strong linear relationship with the duration of use. The fitting function is y = 0.65x + 5, where y is the percentage of power loss and x is the number of minutes of use, with a correlation coefficient of 0.92. This prediction function can be used to estimate the power requirements of different materials under specific usage durations, providing a basis for charging strategies.

[0057] To analyze resource utilization, charging cabinets perform density clustering on peak usage time series, using, for example, the DBSCAN algorithm, to calculate the remaining capacity and occupancy duration of charging cabinets during each time period. For example, during the morning peak, remaining capacity is 20%, with occupancy concentrated between 70 and 100 minutes. During lunch breaks, remaining capacity rises to 70%, with occupancy decreasing to 15 to 25 minutes. The clustering results generate a resource utilization density distribution map, showing that resource utilization is concentrated during peak hours and more dispersed during lunch breaks and nighttime, providing a spatial basis for optimizing resource allocation.

[0058] S1033. Extract the remaining power and capacity data of the charging cabinet based on the resource occupancy density distribution, construct a resource allocation weight function, generate a resource allocation priority sequence, and generate an initial grouping plan in combination with the power consumption prediction function.

[0059] In an embodiment of the present invention, the charging cabinet extracts the remaining power and capacity data from the resource occupancy density distribution map to construct a resource allocation weight calculation function. For example, the weight function is W = 0.6*(1-E)+0.4*(1-C), where E is the percentage of remaining power and C is the percentage of remaining capacity. When the remaining power of a charging cabinet is less than 25% or the remaining capacity is less than 15%, its weight is significantly increased and its priority is raised to the highest. For example, the charging cabinet numbered CB-202 has a remaining power of 20% and a remaining capacity of 12% during the afternoon peak period, and its priority ranks first.

[0060] Based on the priority sequence and power consumption prediction function, the charging cabinet calculates resource allocation parameters for each time period and generates an initial grouping plan. For example, the charging cabinet is divided into a high-frequency turnover area and a low-frequency storage area. The high-frequency turnover area stores small tools such as screwdrivers that are used for less than 50 minutes and have a power consumption rate of less than 0.6% / minute; the low-frequency storage area stores equipment such as electric drills that are used for more than 100 minutes and have a higher power consumption rate. The grouping plan is optimized based on peak demand during peak hours. For example, charging resources in the high-frequency turnover area are prioritized during the morning shift peak, while charging time in the low-frequency storage area is increased during lunch breaks.

[0061] In an embodiment of the present invention, through peak period analysis, power consumption prediction, and resource occupancy optimization, charging cabinets can accurately allocate charging resources and reduce resource bottlenecks during peak periods. For example, the repair team prefers to use charging cabinets CB-201 to CB-203 near the workstation, with access time less than 1.5 minutes, while remote cabinets such as CB-208 are used as backup storage during maintenance cycles, and their priority is dynamically increased. This adaptive allocation significantly improves the efficiency of material access and charging resource utilization, laying the foundation for dynamic scheduling.

[0062] S104. Based on the existing cabinet allocation logic of the charging cabinet, combined with the material usage time distribution and the combined reconstruction feature vector, a genetic algorithm is used to optimize the material storage location, and the cabinet allocation rules are dynamically updated to obtain the optimized charging cabinet layout plan.

[0063] In this embodiment of the present invention, the charging cabinet utilizes material usage trend parameters and a combined reconstruction feature vector, combined with existing cabinet allocation logic, to optimize the storage location of materials within the charging cabinet using a genetic algorithm. This ensures that the cabinet layout adapts to dynamic demand changes, improving access efficiency and space utilization. This optimization process is executed by the charging cabinet control system and can be implemented through local computing or cloud collaboration, depending on the specific scenario.

[0064] S1041. Extract the counter allocation logic and material feature data from the smart cabinet management database, construct a material storage location constraint matrix, and quantify the counter space adaptability and location proximity to obtain a space utilization index.

[0065] In an embodiment of the present invention, the charging cabinet obtains the existing cabinet allocation logic from the smart cabinet management database, including information such as material number, cabinet number, and storage status. It also extracts the combined reconstruction feature vector and usage time distribution trend parameters to construct a material storage location constraint matrix. This matrix records constraints such as material size, access frequency, and combination correlation. For example, a charging cabinet CB-401 contains 40 cabinets. The feature vector shows that the usage correlation of the screwdriver and wrench combination is 0.85, and they should be placed in adjacent cabinets first.

[0066] The charging cabinet numerically quantifies the constraint matrix and calculates the spatial fit and location proximity of the cabinet slots. Spatial fit measures the degree to which the dimensions of the item match the cabinet slot specifications. For example, an electric drill measuring 400 × 120 × 100 mm must fit into a cabinet with internal dimensions greater than 420 × 130 × 110 mm, with a fit of 0.90. Location proximity is determined based on the relevance of item usage. For example, a caliper and a hammer are often used together, with a proximity of 0.82, and should be stored in adjacent or nearby cabinet slots. Combining fit and proximity, a set of space utilization indicators is generated to reflect the efficiency of each cabinet slot. For example, the utilization rate of cabinet slots 1-10 in the CB-401 is 90%, while that of cabinet slots 31-40 is only 50%, indicating uneven space allocation.

[0067] S1042. Construct a fitness function based on space utilization indicators and access data, use a genetic algorithm to generate a set of optimized solutions for material storage locations, and select the solution with the highest comprehensive score through iterative evaluation.

[0068] In an embodiment of the present invention, the charging cabinet extracts access frequency and time interval data from the material access records. For example, the machine repair team accesses the screwdriver set 28 times a day, with an average interval of 20 minutes. Combined with the space utilization index, a material storage location allocation fitness function is constructed. The function comprehensively considers spatial adaptability, proximity, and access efficiency. For example, the fitness function is F = 0.5A + 0.3N + 0.2E, where A is adaptability, N is proximity, and E is access efficiency. The goal is to maximize the F value.

[0069] Charging cabinets use genetic algorithms to optimize the layout. First, a chromosome coding sequence is generated. Each chromosome represents the mapping relationship between a material and a cabinet position, and the gene position is the cabinet position number. For example, the sequence [3,15,7,29]

[0070] This indicates that four items are stored in counters 3, 15, 7, and 29, respectively. The algorithm generates new solutions through crossover and mutation. For example, the parent sequences [3, 15, 7, 29] and [6, 12, 9, 25] produce the offspring [3, 12, 7, 25] and [6, 15, 9, 29] after crossover. The mutation operation randomly adjusts the gene positions to increase diversity. The termination condition is that the rate of change in the optimal fitness is less than 0.015 for 15 consecutive generations. For example, the algorithm stops when the fitness stabilizes at generation 95.

[0071] After generating the optimized solution set, the charging cabinet calculates each solution's cabinet layout rationality score and adjustment cost score. The rationality score is based on access efficiency and space utilization. For example, a solution's rationality score is 0.87. The adjustment cost score measures the complexity of the layout change. For example, adjusting 10 cabinets would cost 0.22. The solution with the highest overall score, such as a rationality score of 0.89 and a cost of 0.20, is selected.

[0072] S1043. Update the counter allocation logic according to the optimization plan, generate a new allocation rule set, and optimize the material storage and access process to improve operational efficiency.

[0073] In this embodiment of the present invention, the charging cabinet extracts the mapping between items and cabinet locations based on the selected solution and updates the allocation logic in the smart cabinet management database. For example, the new rules allocate frequently used small tools to cabinets 1-15, medium-sized tools to cabinets 16-30, and large equipment to cabinets 31-40. Two buffer cabinets are reserved between the zones to handle unexpected demand.

[0074] The optimized layout significantly improves storage and retrieval efficiency. For example, the locksmith team returns tools collectively at the end of their shift, and the placement of adjacent cabinets reduces the average retrieval time from 3 minutes to 1.2 minutes. The relevance of material usage has also been optimized. For example, hexagonal wrenches and screwdriver sets that are often used together are stored in adjacent cabinets, increasing access frequency by 30% and improving operational fluency. Compared to the traditional static layout, the new rules adapt to the reconfiguration of material combinations and changes in usage demand through historical data analysis and dynamic optimization, ensuring the efficient operation of charging cabinets during peak hours and providing flexible support for material management.

[0075] S105. Based on the latest distribution data of the residual power of the charging cabinet and the basis for initial charging resource allocation, the K-means clustering method is used to group the devices, determine the charging priority and duration threshold, and construct an optimized charging cabinet configuration plan.

[0076] In this embodiment of the present invention, the charging cabinet utilizes residual power data and a K-means clustering method to group charging devices. This, combined with cabinet specifications and interface compatibility, optimizes charging priority and duration allocation, generating an efficient cabinet configuration plan to improve charging resource utilization and device availability. This is executed by the charging cabinet control system, which can be implemented locally or through cloud computing, depending on the scenario.

[0077] S1051. Extract device power data from the charging cabinet database, construct a feature data set of the device to be charged, and generate a feature vector through standardization processing.

[0078] In this embodiment of the present invention, the charging cabinet retrieves data from a database, including the device's remaining battery percentage, power loss rate per unit time, and average daily usage times, to construct a characteristic dataset for the devices to be charged. For example, an electric screwdriver has an average remaining battery of 48%, a loss rate of 0.6% / minute, and is used 10 times per day; an electric drill has a remaining battery of 30%, a loss rate of 0.9% / minute, and is used 15 times per day. These data reflect the differences in device usage intensity and charging requirements.

[0079] To ensure the accuracy of cluster analysis, the charging cabinets normalized the feature dataset and calculated the mean and standard deviation of each feature. For example, the mean remaining power was 42% with a standard deviation of 10%; the mean loss rate was 0.7% / minute with a standard deviation of 0.2% / minute. The normalized feature vectors eliminated dimensional differences, generating a multidimensional feature space suitable for cluster analysis and providing a reliable data foundation for grouping.

[0080] S1052: Use the K-means clustering method to group the devices, calculate the charging priority based on the grouping results, and generate a device priority ranking table.

[0081] In an embodiment of the present invention, the charging cabinet uses the K-means clustering algorithm to group devices based on standardized feature vectors, and sets the number of cluster centers to one-fifth of the total number of devices. For example, in a scenario with 200 devices, the number of cluster centers is 40. Through iterative optimization, the algorithm divides the devices into a low-power high-frequency group, a medium-power medium-frequency group, and a high-power low-frequency group based on the Euclidean distance between feature vectors. The average remaining power of the low-power high-frequency group is 20%, and the power consumption rate is 1.0% / minute; the medium-power medium-frequency group is 45%, and the power consumption rate is 0.65% / minute; the high-power low-frequency group is 70%, and the power consumption rate is 0.35% / minute.

[0082] Based on the grouping results, the charging cabinet calculates the average remaining power and power consumption rate for each group and creates a device priority table, sorting the devices in descending order of remaining power and ascending order of power consumption rate. For example, the low-power, high-frequency group has the highest priority because it is low in power and frequently used, and therefore requires priority charging. The high-power, low-frequency group has the lowest priority and can be charged later. This ranking ensures that charging resources are allocated to the devices most in need, improving overall availability.

[0083] S1053. Based on the specifications and interface types of the charging cabinets, generate a cabinet adaptation matrix, optimize the resource allocation ratio and charging time, and generate a cabinet configuration plan.

[0084] In this embodiment of the present invention, the charging cabinet obtains cabinet dimensions and charging port type data from the controller and generates a cabinet adaptation matrix. For example, a standard cabinet measures 450 × 350 × 250 mm and is equipped with 18V and 24V DC interfaces, suitable for small and medium-sized devices. A large cabinet measures 650 × 450 × 350 mm and is equipped with a 36V interface, suitable for large devices. The matrix records the matching relationship between device dimensions and cabinet specifications. For example, an electric wrench measures 350 × 100 × 80 mm and is compatible with a standard cabinet with a matching degree of 0.95.

[0085] Based on the adaptation matrix and grouping results, the charging cabinet calculates the resource allocation ratio for each group of devices. For example, the low-power, high-frequency group, which accounts for 45% of the total number of devices, is allocated 60% of the standard cabinets; the medium-power, medium-frequency group, which accounts for 30%, is allocated 25% of the standard cabinets and 50% of the large cabinets; and the high-power, low-frequency group, which accounts for 25%, is allocated the remaining cabinets. Charging time is set based on power requirements: 100-130 minutes for the low-power, high-frequency group, 70-100 minutes for the medium-power, medium-frequency group, and 40-70 minutes for the high-power, low-frequency group.

[0086] The charging cabinet divides the cabinet space according to the priority sorting table and allocation ratio. For example, CB-

[0087] The 301 charging cabinet contains 36 slots, including 28 standard slots and 8 large slots. Electric wrenches in the low-power, high-frequency group are assigned to standard slots 1-18, with a maximum charging time of 120 minutes. Electric drills in the medium-power, medium-frequency group are assigned to standard slots 19-28 and large slots 1-5, with a maximum charging time of 90 minutes. Cutting machines in the high-power, low-frequency group are assigned to large slots 6-8, with a maximum charging time of 60 minutes.

[0088] In this embodiment of the present invention, the optimized configuration scheme, through precise grouping and differentiated duration settings, concentrates high-frequency equipment in easily accessible locations, reducing the time charging resources occupy. For example, during the morning peak, the average access time for a repair team to access an electric wrench dropped from 2.5 minutes to 1 minute, improving equipment availability by 20%. By adapting to different sizes of locations and interfaces, the solution meets diverse charging needs, significantly improves resource utilization efficiency, and supports dynamic scheduling.

[0089] S106. Based on the optimized charging cabinet location configuration plan and allocation logic, combined with the dynamic change trend of the residual power law, a reinforcement learning algorithm is used to optimize the charging cabinet resource allocation strategy to generate a dynamic grouping plan.

[0090] In this embodiment of the present invention, the charging cabinet, based on an optimized cabinet configuration and allocation logic, utilizes the changing trends of residual power patterns and dynamically adjusts resource allocation strategies through a deep reinforcement learning algorithm. This generates an efficient dynamic grouping solution, ensuring efficient use of charging resources and convenient device access. This is implemented by the charging cabinet control system, either through local computing or cloud collaboration, depending on the specific scenario.

[0091] S1061. Extract cabinet configuration and status data from the charging cabinet management database, construct a resource allocation state space, and generate a resource allocation reward measurement function through idle rate and waiting rate calculation.

[0092] In this embodiment of the present invention, the charging cabinet retrieves optimized cabinet configuration data from the management database, extracts information such as cabinet number, device type, charging status, and remaining battery level, and constructs a resource allocation state space. For example, the charging cabinet CB-601 contains 56 cabinets, 40 of which are for power tools and 16 for specialized equipment. The state space record shows that cabinet B01 is charging an electric screwdriver SD-108, with a remaining battery level of 42% and a charging time of 50 minutes; cabinet B02 is idle and awaiting allocation.

[0093] The charging cabinet collects state space data and calculates the cabinet idle rate and device waiting rate. The idle rate is calculated as the proportion of empty cabinets per unit time. For example, the idle rate during the morning shift is 20%, which drops to 12% during the afternoon shift. The device waiting rate reflects the queues of devices waiting to be charged. For example, the waiting rate during the morning shift is 15%, which rises to 25% during the afternoon shift. Based on these two indicators, a reward metric function is constructed. The function form is R = 0.6*(1-I) + 0.4*(1-W), where I is the idle rate and W is the waiting rate. When the sum of the idle rate and the waiting rate exceeds 45%, the reward value turns negative, prompting the algorithm to optimize the allocation strategy to reduce resource waste and queue time.

[0094] S1062. Use deep reinforcement learning to build a resource scheduling agent, set the action space and generate a policy network, and optimize the resource allocation decision function through iterative evaluation.

[0095] In this embodiment of the present invention, the charging cabinet uses deep reinforcement learning methods, such as deep Q-networks, to construct a resource scheduling agent. The action space encompasses two dimensions: device selection and cabinet assignment. For example, when assigning a cabinet to a newly returned WR-209 electric wrench, the agent considers its remaining battery life of 18%, its expected next use time being the next morning shift, and the status of neighboring cabinets, and chooses cabinet C04. The policy network, modeled through a neural network, takes state space data as input and outputs a probability distribution of actions.

[0096] The charging cabinet evaluates the allocation plan for each time segment, calculates a reward, and updates the network parameters. For example, within an 8-hour shift, the agent tries different allocation plans, recording an allocation accuracy of 88% for the power tool cabinet and 94% for the specialized equipment cabinet. Through gradient descent optimization, the network parameters gradually converge, generating a resource allocation decision function. This decision function dynamically adjusts allocation priorities based on device type, frequency of use, and charge level. For example, frequently used screwdriver sets are prioritized for cabinets near workstations.

[0097] S1063. Based on the statistical allocation accuracy of the sliding time window, fine-tune the decision function and build a dynamic grouping rule base, and continuously update the grouping plan to adapt to real-time demand changes.

[0098] In this embodiment of the present invention, a sliding time window is set up for the charging cabinet, spanning a single work shift, for example, eight hours. This window compiles usage records for the cabinets and calculates the dynamic allocation accuracy. For example, the window records show that the allocation accuracy for the pliers and screwdriver sets is 90%. Because both are frequently used simultaneously, the agent automatically adjusts the allocation strategy and assigns them to adjacent cabinets C05 and C06. The higher accuracy for dedicated equipment cabinets is due to the agent's recognition of the high usage patterns of equipment such as cutting machines during maintenance cycles, prioritizing their allocation to dedicated areas.

[0099] The charging cabinet uses real-time monitoring data through online fine-tuning to optimize the parameters of its decision-making functions. For example, if a new welding station increases welder usage by 30%, the agent increases the welder's allocation weight, prioritizing it in slots D01-D04. Based on these optimized parameters, the charging cabinet builds a dynamic grouping rule base that defines the mapping between equipment type and slot grouping. For example, large equipment with a usage time exceeding three hours is assigned to slots D05-D16 near the maintenance area, while small inspection tools are assigned to slots B01-B20.

[0100] In an embodiment of the present invention, a dynamic grouping solution, driven by real-time data, significantly improves resource utilization efficiency. For example, when a repair team accesses equipment, the average access distance is shortened from 80 meters to 30 meters, and the charging wait time is reduced from 20 minutes to 12 minutes. The solution adapts to changes in equipment usage patterns. For example, the high-frequency use of inspection tools during daily shifts and the concentrated use of maintenance equipment during specific periods are optimized through dynamic adjustments, ensuring the efficient operation of charging cabinets in diverse scenarios.

[0101] S107. Obtain real-time feedback data on the efficiency of charging cabinet material access, adjust the correlation model between power consumption and usage time through iterative optimization, and update the dynamic correlation between material usage and power consumption in combination with long-term trend prediction to form a continuously optimized charging cabinet material management logic.

[0102] In this embodiment of the present invention, the charging cabinet utilizes real-time feedback data to iteratively optimize the model that correlates power consumption with usage duration. This, combined with long-term trend predictions based on usage duration distribution and residual power patterns, dynamically updates the inventory management logic to ensure efficient operation and resource utilization. This is implemented by the charging cabinet control system, which can be implemented locally or through cloud computing, depending on the specific scenario.

[0103] S1071. Extract real-time feedback data from the charging cabinet controller, construct a usage feature sequence, and generate a correlation parameter matrix between power consumption and usage time through time series decomposition and correlation analysis.

[0104] In this embodiment of the present invention, the charging cabinet obtains real-time feedback data on material usage from the controller, including information such as the material access interval, access frequency, and remaining battery power, to construct a characteristic sequence of material usage. For example, the charging cabinet CB-701 records show that the average access interval for electric screwdrivers is 40 minutes, with an average of 18 times per day, and a remaining battery power of 32% when returned; the access interval for specialized testing equipment is 150 minutes, with an average of 4 times per day, and a remaining battery power of 45%. These data reflect the usage patterns and power consumption characteristics of different devices.

[0105] The charging cabinet performs time series decomposition on the characteristic sequence, using the STL decomposition method to extract periodic, trend, and residual components, identifying regular changes in usage behavior. The Pearson correlation coefficient between power consumption rate and usage duration is then calculated to generate a correlation parameter matrix. For example, the correlation coefficient for the electric drill is 0.90, indicating a strong correlation; the correlation coefficient for the cutting machine is 0.70, indicating a weaker correlation, reflecting differences in the operating characteristics of the devices. The matrix records the correlation coefficient and consumption rate of each device, providing a quantitative basis for optimization.

[0106] S1072. Use the gradient descent method to iteratively optimize the associated parameter matrix, generate the cabinet configuration parameter optimization results, and build a material usage behavior prediction model to extract long-term trend characteristics.

[0107] In this embodiment of the present invention, the charging cabinet uses a gradient descent method to perform multiple rounds of optimization based on the associated parameter matrix, with an update step size of 0.005 and an upper limit of 600 iterations, to adjust the parameters of the power consumption rate prediction model. For example, the initial prediction of the power consumption rate of an electric wrench was 0.80% / minute. After 320 iterations, it converged to 0.68% / minute, reducing the prediction error by 35%. The optimization results generate cabinet configuration parameters, including device priority and charging time allocation, to ensure that resource allocation matches actual needs.

[0108] Based on the optimized parameters, the charging cabinets constructed a predictive model for material usage behavior. The ARIMA model was used to analyze the distribution of usage duration, extracting cyclical characteristics and long-term trends. For example, the model showed that peak usage occurred between 7:00 AM and 9:00 AM on weekdays, and between 1:00 PM and 3:00 PM in the afternoon, reaching 28 and 20 times per hour, respectively. Time series decomposition further revealed that usage fluctuations increased by 20% during maintenance cycles, reflecting the significant impact of cyclical demand.

[0109] S1073. Based on the prediction model and power consumption patterns, a dynamic mapping relationship table is established, and management logic is continuously updated through dynamic planning to adapt to changes in usage patterns.

[0110] In an embodiment of the present invention, the charging cabinet calculates the power loss rate per unit time and the cumulative usage time based on the usage behavior prediction model and power consumption patterns, and constructs a power consumption prediction function. For example, during routine maintenance, an electric wrench is used for 40 minutes at a time, with a power consumption rate of 0.70% / minute; during a maintenance task, the usage time increases to 70 minutes, and the power consumption rate rises to 0.90% / minute. Based on these patterns, the charging cabinet generates a dynamic mapping relationship table between material usage and power consumption, recording the relationship between device type, usage scenario, and charging requirements.

[0111] Dynamic programming is used to update the mapping table online, balancing short-term scheduling and long-term planning. For example, when the addition of night shifts expands the power tool usage period from two peak hours to three, the table parameters are automatically adjusted, expanding the charging period allocation from the morning and evening peak hours to include the nighttime period, optimizing the charging station configuration to avoid resource bottlenecks. Dynamic planning uses real-time monitoring data to identify temporary demand surges, such as a 50% increase in maintenance tool usage due to a sudden failure. It prioritizes the charging needs of high-frequency equipment and allocates resources in advance for periodic peaks to ensure that equipment is fully charged before the peak.

[0112] In this embodiment of the present invention, continuously optimized management logic significantly improves the adaptability of charging cabinets. For example, the equipment access time for repair teams during peak hours has been reduced from 2.8 minutes to 1.5 minutes, and charging resource utilization has increased by 25%. Through dynamic adjustments, the logic adapts to changes in usage patterns, such as increased tool demand during night shifts and concentrated maintenance cycles, ensuring accurate resource allocation and equipment availability, providing efficient support for material management.

[0113] Although the present invention has been described in detail above using general descriptions and specific embodiments, it will be apparent to those skilled in the art that modifications and improvements may be made thereto. Therefore, such modifications and improvements, without departing from the spirit of the present invention, are intended to be within the scope of protection claimed herein.

Claims

1. An AI-based intelligent scheduling and data transmission method for material charging, characterized in that: The method comprises: Obtain job responsibility adjustment data, determine the updated set of commonly used material combinations by analyzing job descriptions and historical material usage records, use clustering methods to group materials, and obtain the characteristic vector of combination reconstruction changes; Based on the combined reconstructed characteristic vectors, the real-time mapping relationship between job responsibility adjustment and material demand is obtained. Combined with the usage of the charging cabinet, the usage time distribution is predicted to obtain the trend parameters of material usage behavior. Extract the peak period of material use as the key time node from the trend parameters of the usage time distribution, analyze the correlation coefficient between power consumption and usage time, establish a preliminary resource allocation strategy based on charging cabinets, and determine the initial grouping basis for charging resource allocation; By combining and reconstructing the changing feature vector and the trend parameter of the usage time distribution, the storage location of materials in the charging smart cabinet is optimized to adjust the preset charging cabinet cabinet allocation logic, and the adjusted charging cabinet cabinet allocation logic is obtained; Combined with the charging cabinet return records, the remaining power pattern is analyzed. Based on the initial grouping basis of charging resource allocation, the charging devices are grouped. The charging priority and charging time threshold of each group of charging devices are determined according to the charging cabinet situation, and the charging cabinet position configuration plan is obtained. According to the charging cabinet position configuration plan and the adjusted charging cabinet position allocation logic, combined with the changing trend of the residual power law, the preliminary resource allocation strategy based on the charging cabinet is adjusted to obtain the charging cabinet resource grouping plan.

2. The method according to claim 1, characterized in that The acquisition of job responsibility adjustment data, by analyzing the job description and historical material usage records, determines the updated set of commonly used material combinations, and uses a clustering method to group the materials to obtain the characteristic vector of the combination reconstruction change, including: Obtain job responsibility adjustment records from the human resources management database, compare the job description texts before and after the adjustment, and obtain the job responsibility change range and job responsibility change matrix; Extracting material requisition records corresponding to positions from the material requisition database according to the position responsibility change matrix, counting the material requisition frequency and quantity, and obtaining a position-material association matrix; Performing data standardization on the position-material association matrix, grouping the standardized data using a density clustering method to obtain a material grouping set; The material combination adjustment frequency is calculated based on the material grouping set and the charging cabinet material storage status table. If the adjustment frequency exceeds a preset threshold, the access time series data is extracted for the material combination to obtain a feature vector of the combination reconstruction change related to the charging cabinet.

3. The method according to claim 1, characterized in that The characteristic vector of the change is reconstructed based on the combination, and the real-time mapping relationship between job responsibility adjustment and material demand is obtained. The usage time distribution is predicted in combination with the usage of the charging cabinet to obtain the trend parameters of the material usage behavior, including: Obtaining the characteristic vector in the material management database, and calculating the ratio of the average daily material consumption to the inventory based on the characteristic vector to obtain a material shortage warning index; An association rule mining algorithm is used to process the material shortage warning index to generate a mapping relationship table between material combinations and job responsibilities adjustments, wherein the mapping relationship table includes the number of responsibilities changes and the material demand response time; According to the responsibility adjustment mapping relationship table, the charging cabinet occupancy time record in the charging cabinet management database is read, and the material usage time series data is generated in combination with the material access timestamp; The material usage time series data is sampled according to a fixed time window to obtain a usage intensity curve, and the periodic and trend characteristics are extracted through the time series decomposition method. A material usage behavior time series prediction model is constructed, and the material usage frequency change rate and duration change rate are extracted from the output results of the time series prediction model, and the trend parameters of the material usage behavior are calculated.

4. The method according to claim 1, wherein The method extracts the peak period of material use from the trend parameters of the usage time distribution as the key time node, analyzes the correlation coefficient between power consumption and usage time, establishes a preliminary resource allocation strategy based on the charging cabinet, and determines the initial grouping basis for charging resource allocation, including: Obtaining material usage behavior trend parameters, extracting peak value points of a usage duration distribution curve from the material usage behavior trend parameters, and obtaining a material usage peak period characteristic sequence; Read the material return record in the charging cabinet controller according to the characteristic sequence of the material usage peak period, calculate the power loss rate per unit time, and obtain the power consumption characteristic curve; Using the power consumption characteristic curve to construct a statistical regression function of usage time and power loss, and generate a power consumption prediction function; Perform density clustering calculation on the characteristic sequence of the peak period of material usage, obtain the remaining capacity and occupancy time data of the charging cabinet, and obtain a resource occupancy density distribution map; The remaining power and remaining capacity data of the charging cabinet are extracted according to the resource occupancy density distribution map, a resource allocation weight calculation function is constructed, the charging cabinet resource allocation priority sequence is obtained, and the initial grouping basis for charging resource allocation is determined.

5. The method according to claim 1, wherein The method optimizes the storage location of materials in the charging smart cabinet by combining and reconstructing the changing feature vector and the trend parameter of the usage time distribution to adjust the preset charging cabinet cabinet allocation logic to obtain the adjusted charging cabinet cabinet allocation logic, including: Read the counter allocation logic information in the smart cabinet management database, obtain the characteristic vector and duration distribution parameters in the counter allocation logic information, and generate a material storage location constraint matrix; Perform numerical quantization calculations based on the material storage location constraint matrix to obtain cabinet space size adaptation parameters and location proximity parameters; Using the cabinet space size adaptability parameter and location proximity parameter, access frequency data and time interval data are extracted from the material access records to obtain the material storage location allocation fitness function; The optimization solution set is evaluated and calculated according to the material storage location allocation fitness function to obtain the cabinet layout rationality score and adjustment cost score, and the adjusted charging cabinet cabinet allocation logic is obtained.

6. The method according to claim 1, characterized in that The method combines the charging cabinet return record to analyze the residual power pattern, groups the charging devices according to the initial grouping basis of charging resource allocation, and determines the charging priority and time threshold of each group of charging devices according to the charging cabinet situation to obtain the charging cabinet position configuration plan, including: Obtain the remaining power percentage and power loss rate per unit time of the device in the charging cabinet database to construct a feature data set of the device to be charged; Calculating the feature mean and standard deviation of the feature data set of the device to be charged, and using the K-means clustering method to obtain the device grouping result, where the number of cluster centers in the K-means clustering method is set according to the total number of devices; Calculate the average remaining power and average power consumption rate of each group of devices based on the device grouping results, and obtain a device priority ranking table in descending order of remaining power and ascending order of power consumption rate; For each group of devices in the device priority table, generate a cabinet adaptation matrix based on the cabinet size data and interface type data obtained by the charging cabinet controller, calculate the resource allocation ratio of each charging cabinet according to the cabinet adaptation matrix, and generate a charging resource allocation table; According to the charging resource allocation table and the equipment priority sorting table, the charging cabinet positions are divided into different spaces to obtain the charging cabinet position configuration plan.

7. The method according to claim 1, characterized in that The charging cabinet position configuration plan and the adjusted charging cabinet position allocation logic are combined with the changing trend of the residual power law to adjust the initial resource allocation strategy based on the charging cabinet to obtain the charging cabinet resource grouping plan, including: Read cabinet configuration data from the charging cabinet management database, extract cabinet number information, device type information, and charging status information from the cabinet configuration data, and obtain resource allocation status characteristic data; Calculating a counter idle rate value and a device waiting rate value according to the resource allocation state characteristic data, and generating a resource allocation reward measurement function through the counter idle rate value and the device waiting rate value; A deep reinforcement learning model is used to process the resource allocation reward measurement function, and a resource allocation strategy network model is generated by setting a counter allocation action space; The cabinet allocation plan is evaluated and scored according to the resource allocation strategy network model, the reward value is calculated based on the evaluated score and the resource allocation strategy network model parameters are updated to obtain the resource allocation decision function and generate the charging cabinet resource grouping plan.

8. The method according to claim 1, characterized in that The method also includes: obtaining real-time feedback data on the efficiency of charging cabinet material access, using an iterative optimization method to adjust the modeling parameters of the correlation coefficient between power consumption and usage time, determining the optimization results of the charging cabinet cabinet configuration, and updating the dynamic correlation modeling of charging cabinet material usage and power consumption by predicting the long-term changing trends of usage time distribution and power residual pattern, thereby obtaining a continuously optimized material management logic based on the charging cabinet.

9. The method according to claim 8, characterized in that The system obtains real-time feedback data on the efficiency of charging cabinet material access, uses an iterative optimization method to adjust the modeling parameters of the correlation coefficient between power consumption and usage time, determines the optimization result of the charging cabinet cabinet configuration, and updates the dynamic correlation modeling between charging cabinet material use and power consumption by predicting the long-term trend of usage time distribution and residual power. This results in a continuously optimized charging cabinet-based material management logic, including: Acquire real-time feedback data on material use from the charging cabinet controller, the real-time feedback data including material use time interval parameters, use frequency parameters, and remaining power parameters, and generate a material use feature sequence based on the real-time feedback data; Performing a time series decomposition operation based on the material usage feature sequence, and obtaining a correlation parameter matrix between power consumption and usage duration by calculating the Pearson correlation coefficient between power consumption rate and usage duration; The gradient descent method is used to perform multiple rounds of iterative operations on the associated parameter matrix to obtain the cabinet configuration parameter optimization result; Establish a material usage behavior prediction model based on the cabinet configuration parameter optimization results, extract the usage time distribution characteristics and power consumption patterns through the material usage behavior prediction model, calculate the power loss rate per unit time and the cumulative usage time, and construct a dynamic mapping relationship table between material usage and power consumption; The dynamic programming method is used to update the mapping relationship table online to generate continuously optimized material management logic based on charging cabinets.

Images