Intelligent management method and system for first-aid integrated box

Abstract

The invention relates to the technical field of intelligent terminal control, in particular to an intelligent management method and system for an integrated first-aid kit, and the method comprises the steps: setting a step length through a long-short-term memory network, synchronously generating a trend change curve through forward rolling and backward rolling, extracting a trend difference, and marking a periodic fluctuation material type under a sensitivity threshold value; fusion of bidirectional time sequence association and sensitivity screening is realized in trend dynamic capture and change identification, the sensitivity of material shortage prediction is improved, and the sensitivity of material shortage prediction is improved by combining resource weight priority allocation and pointer network coding, setting urgency parameters to cut low-frequency categories and reconstructing a material category sending sequence. And hierarchical scheduling and sequence planning of instruction sending are optimized. Based on remote feedback recording of continuous failure categories and reorganization of inventory synchronous update sending instructions, the actual state of the materials is dynamically incorporated into a cyclic supplement process, the continuous supply guarantee capability of emergency material use is enhanced, and the abnormal response speed and the inventory health management level are improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent terminal control technology, and in particular to an intelligent management method and system for an integrated first aid box. Background Art

[0002] The field of intelligent terminal control technology includes the use of built-in microprocessors, sensing devices, communication modules and other hardware facilities, combined with software control programs, to conduct real-time monitoring, data collection, information processing and remote transmission of the status of materials in the first aid kit, and complete information collection, status judgment, control execution and feedback interaction for specific application objects.

[0003] An intelligent management method for first aid integrated boxes aims to improve the timeliness and accuracy of first aid material management, ensure that first aid supplies are in good, sufficient and usable condition when needed, avoid the risk of first aid failure due to missing, expired or misplaced materials, realize the information management of the entire life cycle of first aid materials, and improve the efficiency of first aid response and the safety of material use.

[0004] Existing technologies lack the ability to analyze historical data trends and extract cyclical change patterns, resulting in insufficient construction of material usage behavior models, reliance on single-time detection results for material inventory status updates, lack of dynamic trend tracking and sensitivity difference screening, and difficulty in timely capturing potential shortage risks caused by fluctuations in material consumption. There is a lag in replenishment response, and existing technologies use a fixed batch push method, without hierarchical sorting and dynamic adjustment of instructions based on the urgency of materials, resulting in disordered replenishment scheduling priorities, increased risks of delayed replenishment of important materials, and failure to predict demand growth in advance through trend fluctuation analysis, resulting in material shortages not being replenished in a timely manner, which directly increases the risk of failed emergency response. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an intelligent management method and system for an integrated first aid box.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: a method for intelligent management of an integrated first aid box, comprising the following steps: Step 1: Read and sort the data stream based on the material category identification number and access timestamp, group and lock the records by category, calculate the continuous access time interval and accumulate the quantity, classify the period interval and calculate the average, and generate a period usage characteristic set; Step 2: Based on the set of periodic usage characteristics, a long short-term memory network is used to call the usage monitoring module to set the step size. The indicator vector is rolled forward to generate a cumulative trend and record the changes. The indicator vector is rolled backward to generate a reverse trend and record the changes. The trend difference is calculated and the offset category is extracted. The sensitivity screening is performed to mark the periodic fluctuation material category to form a periodic fluctuation material category set. Step 3: Based on the set of cyclically fluctuating material categories, the material identification module is used to screen sensitivity categories, extract remaining inventory, calculate the product of usage frequency, screen out-of-stock categories, archive priority materials, and generate a replenishment priority material list; Step 4: Based on the replenishment priority material list, resource weights are set in conjunction with the replenishment management unit. Material categories are prioritized and allocated to the scheduling pool and the hierarchy is locked. Using a pointer network, urgency parameters are set to filter the sending categories and cut low-frequency categories. Combined coding is used to generate a hierarchical sending order, generating a remote replenishment instruction sequence. Step 5: Based on the remote replenishment instruction sequence, the remote delivery feedback module is called in the intelligent execution window to record the delivery status and response time, count the consecutive failure categories and queue them for retransmission, update the inventory status of the successful categories and reorganize the delivery plan, and generate a rapid replenishment status table for emergency supplies.

[0007] As a further solution of the present invention, the specific steps of generating the periodic usage characteristic set are: According to the material category identification number and the material access timestamp, the material category identification number and the material access timestamp data are extracted in chronological order and invalid records are cleaned. The material category identification number and the timestamp are sorted in ascending order according to the material category number. An independent index identifier is generated for the records under each category number, and a set of access records of the same category is output; Based on the same category access record set, extract the time difference between adjacent records according to the index identifier, calculate the continuous time interval under each category and store it as a time difference set, count the number of accesses within each set of time differences and classify them into the material category index, and generate a material access interval statistics table; Based on the material usage interval statistics table, a fixed time period length is set and the usage quantity in each time period is extracted. The usage quantity is accumulated and the average value of each time period is calculated. At the same time, the material category number is marked to generate a period usage characteristic set.

[0008] As a further solution of the present invention, the specific steps of generating the periodically fluctuating material category set are: Based on the periodic usage characteristic set, a long short-term memory network is used, a forward step size parameter is set, and a periodic indicator vector of the resource category is extracted. Subsequences are extracted by sliding on the indicator vector according to the step size, and the numerical changes of each subsequence are accumulated and the direction of change is recorded. The forward trend change curves of each category are generated by superposition, and the forward trend change sequence is output; Based on the forward trend change sequence, a backward step size parameter is set, and subsequences are extracted by sliding in reverse order on the indicator vector according to the step size, and the numerical changes of each subsequence are accumulated and the direction of change is recorded. The backward trend change curves of each category are superimposed to generate the backward trend change curves, and the backward trend change sequence is output; Based on the backward trend change sequence, the forward trend value and the backward trend value of the corresponding time step are extracted, the difference between the two is calculated and the difference category is extracted, the categories whose differences exceed the sensitivity threshold are screened and assigned periodic fluctuation labels, and the periodic fluctuation material category set is output.

[0009] As a further solution of the present invention, the long short-term memory network is according to the formula: ; in: represents the hidden state at the current time step t, represents the activation function, Represents the weight matrix from the hidden state to the current hidden state, represents the weight matrix input to the hidden state, represents the input vector at the current time step, represents the bias term, represents the periodic trend attenuation coefficient, represents the weighted sum of historical input vectors, represents the weight coefficient of the historical input vector, is the time step First aid supplies management data input, represents the external interference correction coefficient, Indicates external influencing factors.

[0010] As a further solution of the present invention, the subsequence is extracted by sliding on the indicator vector according to the step size to obtain a subsequence set. The sliding operation is performed on the periodic indicator vector corresponding to each material category with the set forward step size parameter as the unit, and continuous data sub-segments are intercepted according to the step size length. After one round of sliding, the starting position is moved forward by one data point, and the subsequence interception is repeated until the end of the indicator vector to form a subsequence set.

[0011] As a further solution of the present invention, the specific steps of generating the replenishment priority material list are: Based on the periodic fluctuation material category set, the material category number is extracted and matched with the data stored in the material identification module, the categories with a sensitivity greater than the set sensitivity threshold are screened, and the corresponding remaining inventory quantity is extracted, and a material category inventory corresponding table is established, and the sensitivity inventory data set is output; Based on the sensitivity inventory data set, historical usage records of each material category are extracted, the usage frequency is extracted and classified into the inventory quantity corresponding table, the product of inventory quantity and usage frequency is calculated and compared with the threshold, and the material categories whose product is lower than the threshold are screened, and a list of materials in short supply is output; Based on the inventory shortage material list, the screened material category numbers are extracted, classified and archived and sorted according to the inventory shortage level, a material category priority processing index table is generated and integrated into the inventory management list, and a replenishment priority material list is output.

[0012] As a further solution of the present invention, the specific steps of generating the remote replenishment instruction sequence are: Based on the replenishment priority material list, extract the material category number and set the resource allocation weight value, arrange the material categories according to priority, allocate the material categories to different scheduling pools and set corresponding hierarchical identifiers, and output the scheduling pool material hierarchical data; Based on the material hierarchical data of the scheduling pool, extract the material category number in each scheduling pool, set the urgency judgment parameters and filter the urgency categories, cut the low-frequency category records, retain the urgent materials and establish a sending index table, and output the material instruction sending classification table; Based on the material instruction sending classification table, a pointer network is used to extract the material category numbers and combine the category codes in priority order, generate a material instruction code sequence and arrange it in sequence, encapsulate it into a remote control sending data packet, and output a remote replenishment instruction sequence.

[0013] As a further solution of the present invention, according to the formula: ; in: represents the probability of selecting material category m at the kth position, represents the similarity score between the k-th position output by the encoder and the m-th emergency supplies category, is the exponential function of the similarity score, m is the total number of emergency supplies categories, Indicates the weight coefficient of priority between first aid material categories, represents the relative distance measure between location k and material category m, is the external factor correction coefficient, It is the correction value of material category m affected by external factors.

[0014] As a further solution of the present invention, the specific steps of generating the above-mentioned steps are: Based on the remote replenishment instruction sequence, the sending instruction data corresponding to each material category is extracted, the instruction content is pushed one by one by setting the sending window time parameter, and the sending status and response time are recorded. Each sending record is archived and classified and summarized by material category number and instruction status to generate a sending feedback record table; Based on the delivery feedback record table, extract the delivery failure category numbers and sort them by the number of failures, filter the categories of materials with consecutive failures and establish a resending priority index, update the remaining inventory quantity by extracting the delivery success categories and generate a corresponding material status list, and integrate the resending and update lists to generate an inventory reorganization task list; Based on the inventory reorganization task list, the remaining inventory status records are extracted and updated to the material inventory database. The sending order list is recompiled by extracting the resend category number, the resend instructions are pushed in sequence and the replenishment status is synchronized, the replenishment operation records and inventory changes are summarized, and a rapid replenishment status table for emergency supplies is generated.

[0015] An intelligent management system for an integrated first aid box, the intelligent management system for an integrated first aid box being used to execute the intelligent management method for an integrated first aid box, the system comprising: Material characteristics collection module: Based on the material category identification number and access timestamp, read the data stream, sort it in ascending order according to the material category identification number, group it by the material category identification number and lock the records, calculate the time difference between the access timestamps of two consecutive records with the same category identification number, count the number of accesses corresponding to the time difference, extract the time difference value set, classify and attribute the time difference according to the preset period interval, count the number of time differences in each period interval, calculate the average time difference in the period interval, and establish a period usage characteristics set; Cyclic Fluctuation Screening Module: Based on the set of cyclic usage characteristics, the access monitoring module is called to set the step size. Based on the long-short-term memory network structure, the cyclic mean index vector corresponding to the material category identification number is constructed. The sum of the index vectors within the step size interval is rolled forward along the time series and recorded to generate a cumulative trend. The sum of the index vectors within the step size interval is rolled backward along the time series and recorded to generate a reverse trend. The corresponding point difference between the positive cumulative trend and the reverse cumulative trend is calculated. The category identification numbers whose absolute value of the difference exceeds the set sensitivity threshold are screened, and the category identification number set with significant fluctuations is extracted to form a cyclically fluctuating material category set. Inventory priority filing module: Based on the periodic fluctuation material category set, the material identification module is called to screen the category identification numbers within the sensitivity range, extract the remaining inventory value corresponding to each category identification number, extract the corresponding usage frequency value, calculate the product of the remaining inventory value and the usage frequency value, screen the category identification numbers whose product value is lower than the preset inventory safety threshold, and establish a replenishment priority material list; A hierarchical instruction generation module extracts resource weight values ​​set by the replenishment management unit based on the replenishment priority material list, sorts the resource weight values ​​corresponding to the category identification numbers in ascending order, selects high-weight category identification numbers and adds them to the scheduling pool, sets a hierarchical lock flag for each category identification number in the scheduling pool, sets an urgency value based on the pointer network structure, filters category identification numbers with urgency values ​​above a threshold, deletes category identification numbers in the scheduling pool with urgency values ​​below the threshold, combines the category identification numbers and urgency values ​​to generate a coding vector, determines the sending order based on the arrangement order of the coding vectors, and establishes a remote replenishment instruction sequence; Remote scheduling execution module: Based on the remote replenishment instruction sequence, the remote sending feedback module is called to send the replenishment instruction to the external execution terminal, the sending status and response time of each replenishment instruction are recorded, the category identification numbers of the three consecutive failed sending are screened and added to the resending queue, the category identification numbers of the successful sending are extracted and the inventory value is updated, the remote replenishment instruction sequence is reorganized, and a rapid replenishment status table of emergency supplies is generated.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are: 1. This invention uses material category identification numbers and access timestamps to read, sort, and group data, accurately locks category records, and calculates continuous access intervals. It then combines classification period intervals with mean calculations to generate a set of periodic usage characteristics. This optimizes the data granularity level and enhances the accuracy of modeling periodic material usage behavior. 2. This invention uses a long short-term memory network to set the step size, synchronously rolling forward and backward to generate trend change curves, extract trend differences, and mark the categories of periodic fluctuating materials below the sensitivity threshold. This achieves the integration of bidirectional time series correlation and sensitivity screening in trend dynamic capture and change identification, improving the sensitivity of material shortage prediction. 3. This invention combines resource weighted priority allocation with pointer network coding, sets urgency parameters to trim low-frequency categories, and restructures the order in which material categories are sent, optimizing the hierarchical scheduling and sequential planning of command transmission. Based on remote feedback, it records consecutive failure categories, reorganizes inventory and synchronizes updates to send commands, and dynamically incorporates the actual status of materials into the replenishment cycle. This enhances the continuous supply guarantee capability of emergency materials, improves abnormal response speed, and improves inventory health management. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the workflow of the present invention; Figure 2 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0019] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0020] See also Figure 1 The present invention provides a technical solution: a method for intelligent management of an integrated first aid box, comprising the following steps: Step 1: Read and sort the data stream based on the material category identification number and access timestamp, group and lock the records by category, calculate the continuous access time interval and accumulate the quantity, classify the period interval and calculate the average, and generate a period usage characteristic set; Step 2: Based on the set of periodic usage characteristics, a long short-term memory network is used to call the usage monitoring module to set the step size. The indicator vector is rolled forward to generate a cumulative trend and record the changes. The indicator vector is rolled backward to generate a reverse trend and record the changes. The trend difference is calculated and the offset category is extracted. The sensitivity screening marks the cyclically fluctuating material categories to form a cyclically fluctuating material category set. Step 3: Based on the cyclically fluctuating material category set, the material identification module is used to screen sensitivity categories, extract remaining inventory, calculate the product of usage frequency, and screen out-of-stock categories. Prioritized materials are archived and a replenishment priority list is generated. Step 4: Based on the replenishment priority material list, resource weights are set in conjunction with the replenishment management unit. Material categories are prioritized and allocated to the scheduling pool and the hierarchy is locked. Using a pointer network, urgency parameters are set to filter the delivery categories and cut low-frequency categories. Combined coding is used to generate a hierarchical delivery order, generating a remote replenishment instruction sequence. Step 5: Based on the remote replenishment instruction sequence, call the remote delivery feedback module in the intelligent execution window, record the delivery status and response time, count the consecutive failure categories and queue them for resending, update the inventory status of the successful categories and reorganize the delivery plan, and generate a rapid replenishment status table for emergency supplies.

[0021] The specific steps to generate a periodic usage feature set are: According to the material category identification number and the access timestamp, extract the material category identification number and material access timestamp data in chronological order and clean invalid records, and sort the timestamps in ascending order according to the material category number. Generate an independent index identifier for the records under each category number and output the access record set of the same category; Based on the same category of access record sets, the time difference between adjacent records is extracted according to the index identifier, the continuous time interval under each category is calculated and stored as a time difference set, the number of access times within each set of time differences is counted and classified into the material category index, and a material access interval statistics table is generated; Based on the material usage interval statistics table, a fixed time period length is set and the usage quantity in each time period is extracted. The usage quantity is accumulated and the average value of each time period is calculated. At the same time, the material category number is marked to generate a period usage characteristic set. Based on the material category identification number and access timestamp data, the missing value cleaning method is used to eliminate invalid records. The dropna function in the pandas library is used, the parameter subset is set to the material category identification number and access timestamp field, and inplace is set to True. The data is sorted in chronological order using the data sorting method. The sort_values ​​function in the pandas library is used, the parameter by is set to the access timestamp field, and ascending is set to True. For the arranged data, the grouping index method is used to generate an independent index identifier under each material category identification number. The groupby function in the pandas library is combined with the cumcount function. The groupby parameter is set to the material category identification number field, and cumcount is used to generate each group of internal continuous indexes. The sorted material category identification number, access timestamp, and independent index identifier are extracted to establish a set of access records of the same category. Based on the same category access record set, the time difference calculation method is used to extract the difference of the timestamps corresponding to adjacent index identifiers. The diff function in the pandas library is used to first group by the material category identification number field, and then the diff function is executed on the access timestamp field. The result is returned as a timedelta object. The unit conversion method is used to convert the time difference to hours. The total_seconds method of the timedelta object is divided by 3600 to convert the difference unit from seconds to hours. The continuous time intervals under each category identification number are stored. The statistical grouping method is used to calculate the number of occurrences of each time difference under each category identification number and classify them. The material category identification number field is grouped by the pandas library and agg aggregated. The count function is used to count the number of accesses and establish a material access interval statistics table. Based on the material usage interval statistics table, the interval partitioning method is used to divide the fixed cycle length. The arange function in the numpy library is used to generate a fixed step time period sequence. The parameter start is set to 0, the stop is set to the maximum time difference, and the step is set to the fixed cycle length. The interval attribution method is used to assign each time difference to the corresponding time period. The cut function in the pandas library is used, the parameter bins is set to the generated time period sequence, the labels is set to the time period number, and include_lowest is set to True. The period accumulation method is used to accumulate the usage quantity in each time period. The groupby time period number field in the pandas library is used and the sum aggregation of the usage count field is used. The mean calculation method is used to extract the mean usage quantity in each group of time periods. The groupby time period number field in the pandas library is used and the agg aggregation mean function is used. The material category identification number is added to each record to establish a period usage feature set.

[0022] The specific steps for generating a cyclically fluctuating material category set are as follows: Based on the periodic usage characteristic set, a long short-term memory network is used to set the forward step size parameter to extract the periodic indicator vector of the resource category. Subsequences are extracted by sliding on the indicator vector according to the step size. The numerical changes of each subsequence are accumulated and the direction of change is recorded. The forward trend change curves of each category are superimposed and output as the forward trend change sequence. Based on the forward trend change sequence, set the backward step size parameter, slide the indicator vector in reverse order according to the step size to extract the subsequence, accumulate the numerical changes of each subsequence and record the change direction, superimpose and generate the backward trend change curve of each category, and output the backward trend change sequence; Based on the backward trend change sequence, the forward trend value and the backward trend value of the corresponding time step are extracted, the difference between the two is calculated and the difference category is extracted. The categories with differences exceeding the sensitivity threshold are screened and assigned periodic fluctuation labels, and the periodic fluctuation material category set is output; Based on the periodic usage feature set, a long short-term memory network is used, and the forward step parameter is set. The step value is set to a fixed integer of 3. The periodic indicator vector data corresponding to the resource category is extracted. The Sequential model in the TensorFlow framework is called, the number of units is set to 64, the activation function is tanh, and the input shape is set to the step length and the number of indicator vector dimensions. The time step expansion method is used to perform a sliding window on the indicator vector according to the step length to extract subsequences. Each subsequence consists of consecutive step time points. The numerical changes of consecutive time points within each subsequence are accumulated. The diff function in the numpy library is used to calculate the adjacent numerical changes. The sum function in the numpy library is used to accumulate and sum the changes. The positive and negative signs of the total changes are recorded as the change direction. The change directions of all subsequences are superimposed to generate the forward trend change curve corresponding to the resource category, and the forward trend change sequence is output. Based on the forward trend change sequence, the long short-term memory network modeling method is adopted, the backward step parameter is set, and the step value is set to a fixed integer of 3. The Sequential model in the TensorFlow framework is called, and LSTM layers are added in sequence. The number of LSTM units is set to 64, the activation function is tanh, and the input shape is set to the step length and the number of indicator vector dimensions. The periodic indicator vector is reversed. The flip function in the numpy library is used to flip the indicator vector according to the time step dimension. The subsequence is extracted by sliding the window in reverse order according to the step size. Each subsequence consists of consecutive step-length reverse time points. The numerical changes of consecutive time points within each reverse subsequence are accumulated. The diff function in the numpy library is used to calculate the adjacent numerical changes. The sum function in the numpy library is used to accumulate and sum the changes. The positive and negative signs of the total changes are recorded as the change direction. The change directions of all reverse subsequences are superimposed to generate the backward trend change curve corresponding to the material category, and the backward trend change sequence is output. Based on the backward trend change sequence, the trend difference analysis method is adopted to extract the forward trend value and backward trend value corresponding to each time step. The subtract function in the numpy library is used to calculate the trend value difference according to the element position to obtain the difference vector. The data points corresponding to the material category identification number in the difference vector are extracted, and the categories with the absolute value of the difference greater than the set sensitivity threshold are screened. The sensitivity threshold is set to a fixed value of 5. The where function in the numpy library is used for screening. The condition is that the absolute value of the difference vector is greater than the sensitivity threshold. The screened categories are assigned periodic fluctuation labels, and a periodic fluctuation material category set is established.

[0023] Long short-term memory network, according to the formula: ; in: represents the hidden state at the current time step t, represents the activation function, Represents the weight matrix from the hidden state to the current hidden state, represents the weight matrix input to the hidden state, represents the input vector at the current time step, represents the bias term, represents the periodic trend attenuation coefficient, represents the weighted sum of historical input vectors, represents the weight coefficient of the historical input vector, is the time step First aid supplies management data input, represents the external interference correction coefficient, Indicates external influencing factors; Execution process: First, by inputting vector Represents the current management status information and sensor data of the first aid box, including box temperature, inventory status and usage of first aid items, and then according to the hidden status and the current input , through the weight matrix and Calculate the current hidden state , and through the activation function To output the current management decision and forecast results, use weighted summation , considering the impact of past management state changes on current decisions, is the weight of historical data corresponding to the time step, Is the periodic trend attenuation coefficient, which is used to control the influence of historical data on the current state. External factors are corrected by the term Adjustment by coefficient Control ensures that the impact of external factors can be fully reflected in current management decisions, and ultimately generates accurate first aid box management decisions, thereby optimizing the management and replenishment of items in the box, monitoring the status of the system, and improving first aid response capabilities.

[0024] The subsequence is extracted by sliding on the indicator vector according to the step size to obtain a subsequence set. The sliding operation is performed on the periodic indicator vector corresponding to each material category with the set forward step size parameter as the unit. The continuous data sub-segments are intercepted according to the step size. After one round of sliding, the starting position is moved forward to advance one data point. The subsequence interception is repeated until the end of the indicator vector to form a subsequence set.

[0025] The specific steps to generate a replenishment priority material list are: Based on the periodic fluctuation material category set, the material category number is extracted and matched with the material identification module storage data. The categories with sensitivity greater than the set sensitivity threshold are screened, and the corresponding remaining inventory quantity is extracted. The material category inventory correspondence table is established and the sensitivity inventory data set is output; Based on the sensitive inventory data set, the historical usage records of each material category are extracted, the usage frequency is extracted and classified into the inventory quantity corresponding table, the product of inventory quantity and usage frequency is calculated and compared with the threshold, the material categories with the product below the threshold are screened, and a list of materials in short supply is output; Based on the inventory shortage list, the selected material category numbers are extracted, classified and filed, and sorted according to the degree of inventory shortage. A material category priority processing index table is generated and integrated into the inventory management list, and a replenishment priority material list is output; Based on the cyclically fluctuating material category set, the data matching retrieval method is used to extract the material category number. The merge function in the pandas library is used, and the parameter left_on is set to the category number field in the cyclically fluctuating material category set, right_on is set to the category number field in the material identification module storage data, and how is set to inner. Two-way matching is performed to extract the detailed attribute data of the material category. The sensitivity screening method is used to extract the sensitivity value field. The where function in the numpy library is used to set the screening condition as the sensitivity value field is greater than the set sensitivity threshold. The sensitivity threshold is fixed to a value of 0.75. The remaining inventory quantity corresponding to the filtered category number is extracted. The set_index function in the pandas library is used with the category number as the index and the remaining inventory quantity as the corresponding value to establish a material category inventory corresponding table and output the sensitivity inventory data set. Based on the sensitive inventory dataset, a usage record analysis method is used to extract the historical usage records of each material category number. The SELECT command in the database query language SQL is called, and the fields are the category number and the access timestamp. The WHERE condition is limited to a set historical period of 180 days. The access frequency calculation method is used. The category number field in the pandas library is grouped by and the access records are counted. The number of times each material category is used is extracted and classified into the inventory quantity corresponding table. The inventory frequency product calculation method is used to extract the inventory quantity field and the access frequency field. The multiply function in the numpy library is called to multiply each element to generate the product field. The low product screening method is used. The where function in the numpy library is used to filter category numbers whose product field values ​​are less than a fixed threshold of 100. The screening results are extracted by category number to establish a list of scarce materials in stock. Based on the inventory of scarce materials, the classified archiving and sorting method is used to extract the selected material category numbers. The assign function in the pandas library is called to create a new scarcity field. The scarcity field value is set to the ratio of the remaining inventory quantity to the access frequency. The sort_values ​​function in the pandas library is called with the by parameter set to the scarcity field and ascending set to True to sort the category numbers in ascending order of scarcity. The reset_index function in the pandas library is used to rebuild the index and generate an index table for material category priority processing. The concat function in the pandas library is called to splice the index table with the original inventory management table by column, integrate them to generate an inventory management list, and output a replenishment priority material list.

[0026] The specific steps to generate a remote replenishment instruction sequence are: Based on the replenishment priority material list, extract the material category number and set the resource allocation weight value, sort the material categories according to priority, assign the material categories to different scheduling pools and set the corresponding hierarchical identifiers, and output the scheduling pool material hierarchical data; Based on the hierarchical data of materials in the scheduling pool, the material category numbers in each scheduling pool are extracted, the urgency judgment parameters are set and the urgency categories are filtered, low-frequency category records are trimmed, urgent materials are retained and a sending index table is established, and a material instruction sending classification table is output; Based on the material instruction sending classification table, a pointer network is used to extract the material category numbers and combine the category codes in priority order to generate a material instruction code sequence and arrange them in sequence. The sequence is encapsulated into a remote control sending data packet and outputs a remote replenishment instruction sequence. Based on the replenishment priority material list, the resource allocation sorting method is used to extract the material category number. The DataFrame object in the pandas library is called to extract the material category number field and add a resource weight field. The resource weight value is set to the weighted average of the category priority parameter and the historical usage frequency parameter. The assign function in the pandas library is used to create a new field. Using the priority sorting method, the sort_values ​​function in the pandas library is used, the by parameter is set to the resource weight field, and ascending is set to False. The material category numbers are sorted in descending order. The hierarchical allocation method is used, and the cut function in the pandas library is used. The bins parameter is set to the preset hierarchical segmentation threshold list and the labels is set to the hierarchical identifier set. The material category numbers are assigned to different scheduling pools and labeled with the corresponding hierarchies to establish hierarchical material data in the scheduling pool. Based on the stratified material data of the scheduling pool, the urgency screening method is used to extract the material category numbers in each scheduling pool. The urgency judgment parameter value is set to 0.8. The query function in the pandas library is called to filter the category numbers whose urgency score field is greater than the urgency judgment parameter value. The low-frequency clipping method is used. The category number field in the pandas library is grouped by and the access frequency is counted. The lower limit of the access frequency is set to 3 times. The where function in the numpy library is called to filter and remove category numbers with an access frequency less than the lower limit. The urgency category number records are retained. The reset_index function in the pandas library is used to rebuild the index, establish the sending index table corresponding to the material category, and output the material instruction sending classification table. Based on the material instruction sending classification table, the pointer network coding method is used to extract the material category numbers. The PointerNet network structure in the TensorFlow framework is called, the input sequence is set to the material category number array, the number of LSTM units used in the pointer generation layer is set to 128, the activation function is relu, and the output layer uses softmax activation. The network coding module is used to combine the material category numbers in priority order. The argsort function in the numpy library is used to sort the category numbers in ascending order according to the priority field to generate the arranged coding sequence. The stack function in the numpy library is called to combine the category number and the priority code into a material instruction coding sequence. The coding sequence is encapsulated into the standard remote control sending format to generate a remote replenishment instruction sequence.

[0027] According to the formula: ; in: represents the probability of selecting material category m at the kth position, represents the similarity score between the k-th position output by the encoder and the m-th emergency supplies category, is the exponential function of the similarity score, m is the total number of emergency supplies categories, Indicates the weight coefficient of priority between first aid material categories, represents the relative distance measure between location k and material category m, is the external factor correction coefficient, is the correction value of material category m affected by external factors; Implementation process: First, calculate the similarity score between each category based on the different categories of supplies in the first aid kit , the score reflects the relationship between location k and material category m, followed by the similarity score Through the exponential function Convert it into a probability value so that the priorities of different material categories can be compared in subsequent calculations, and then introduce category priority, relative distance between categories and external environmental factors, and weight coefficient and relative distance metrics Helps adjust the priority of material categories in terms of space or demand, ensuring that more urgent and relevant material categories are given priority, and external factor correction coefficient and external correction values It is used to introduce the impact of the environment or emergencies on the selection of supplies, ensuring that emergency supplies can be dynamically adjusted according to changes in external conditions. Ultimately, the combined effect of all factors is calculated by calculating the selection probability of each supply category. , determines the priority of material categories, generates corresponding replenishment instructions, and ensures that the materials in the first aid box can meet first aid needs and respond to possible emergencies.

[0028] The specific steps for generation are: Based on the remote replenishment instruction sequence, the corresponding sending instruction data for each material category is extracted. By setting the sending window time parameters, the instruction content is pushed one by one, and the sending status and response time are recorded. Each sending record is archived and classified and summarized by material category number and instruction status to generate a sending feedback record table; Based on the delivery feedback record table, extract the delivery failure category numbers and sort them by the number of failures. Filter the categories of materials with consecutive failures and establish a resend priority index. Update the remaining inventory quantity by extracting the successful delivery categories and generate a corresponding material status list. Combine the resend and update lists to generate an inventory reorganization task list. Based on the inventory reorganization task list, the remaining inventory status records are extracted and updated to the material inventory database. The sending order list is reorganized by extracting the reissue category number, and the reissue instructions are pushed in sequence and the replenishment status is synchronized. The replenishment operation records and inventory changes are summarized to generate a rapid replenishment status table for emergency supplies. Based on the remote replenishment instruction sequence, the window sending and pushing method is used to extract the sending instruction data corresponding to each material category. The Python standard library sched module is called to establish a scheduler object, and the sending window time parameter is set to 10 seconds. The enter method in the sched module is used, and the parameter delay is set to the sending window time parameter, priority is set to 1, action is set to the sending instruction push function, and argument is set to the instruction data content. The instruction content is pushed one by one. The time function in the Python standard library time module is called to record the sending start timestamp and the receiving response timestamp. The sending status marking method is used to set the sending status field value according to the response return content. The successful return value is set to 1, and the failed return value is set to 0. The pandas library DataFrame object is used, and the fields include material category number, sending time, response time, and sending status. Each sending record is archived. The pandas library groupby function is called to classify and summarize the instruction status according to the material category number field to generate a sending feedback record table. Based on the delivery feedback record table, the failure count sorting and filtering method is used to extract the delivery failure category numbers. The query function in the pandas library is used to filter data with a delivery status field value of 0. The groupby function in the pandas library is called to group by category number and count the number of failures. The sort_values ​​function in the pandas library is called with the by parameter set to the failure count field and ascending set to False to sort in descending order. Category numbers with three or more consecutive failures are filtered out. The successful delivery inventory update method is used to extract the delivery success category numbers. The query function in the pandas library is used to filter data with a delivery status field value of 1. A list of successful category numbers is extracted. The inventory database update interface is called, and the parameter category number and success flag are used to synchronize the inventory quantity. The updated material category number and inventory quantity are archived. The filtered consecutive failure category numbers are integrated with the successfully updated category numbers and inventory quantity records to generate an inventory reorganization task list. Based on the inventory reorganization task list, the inventory synchronization update and resending scheduling method is used to extract the remaining inventory status records. The inventory database connection module is called, and the SQL UPDATE command is executed. The parameter field category number matches the updated inventory quantity. The category number that needs to be resent is extracted, and the sending order list is recompiled. The sort_values ​​function in the pandas library is called with the parameter by set to the inventory urgency field in ascending order to generate a new sending list. The Python standard library sched module is used to create a new scheduler object. The enter method is used to set the time window for re-pushing each instruction to 5 seconds. The resent instructions are pushed in sequence, and the time recording module is used to synchronize the resending instruction status. The concat function in the pandas library is called to merge the replenishment sending record with the inventory change record to generate a complete form. After summarizing, a rapid replenishment status table of emergency supplies is generated.

[0029] See also Figure 2 , an intelligent management system for first aid integrated box, the system includes: Material characteristics collection module: Based on the material category identification number and access timestamp, read the data stream, sort it in ascending order according to the material category identification number, group it by the material category identification number and lock the records, calculate the time difference between the access timestamps of two consecutive records with the same category identification number, count the number of accesses corresponding to the time difference, extract the time difference value set, classify and attribute the time difference according to the preset period interval, count the number of time differences in each period interval, calculate the average time difference in the period interval, and establish a period usage characteristics set; Cyclic Fluctuation Screening Module: Based on the set of cyclical usage characteristics, the access monitoring module is called to set the step size. Based on the long-short-term memory network structure, the cyclic mean index vector corresponding to the material category identification number is constructed. The sum of the index vectors within the cumulative step interval is rolled forward along the time series and recorded to generate a cumulative trend. The sum of the index vectors within the cumulative step interval is rolled backward along the time series and recorded to generate a reverse trend. The corresponding point difference between the positive cumulative trend and the reverse cumulative trend is calculated. The category identification numbers whose absolute value of the difference exceeds the set sensitivity threshold are screened, and the category identification number set with significant fluctuations is extracted to form a cyclically fluctuating material category set. Inventory priority filing module: Based on the cyclically fluctuating material category set, the material identification module is called to screen the category identification numbers within the sensitivity range, extract the remaining inventory value corresponding to each category identification number, extract the corresponding usage frequency value, calculate the product of the remaining inventory value and the usage frequency value, screen the category identification numbers whose product value is lower than the preset inventory safety threshold, and establish a replenishment priority material list; Hierarchical instruction generation module: Based on the replenishment priority material list, the resource weight values ​​set by the replenishment management unit are extracted, the resource weight values ​​corresponding to the category identification numbers are sorted in ascending order, the high-weight category identification numbers are selected and added to the scheduling pool, the hierarchical lock flag of each category identification number in the scheduling pool is set, the urgency value is set according to the pointer network structure, the category identification numbers with urgency values ​​above the threshold are filtered, and the category identification numbers with urgency values ​​below the threshold are deleted from the scheduling pool. The category identification numbers and urgency values ​​are combined to generate a coding vector. The sending order is determined according to the order of the coding vector arrangement to establish a remote replenishment instruction sequence; Remote scheduling execution module: Based on the remote replenishment instruction sequence, the remote sending feedback module is called to send the replenishment instruction to the external execution terminal, the sending status and response time of each replenishment instruction are recorded, the category identification numbers of three consecutive failed sending are filtered out and added to the resending queue, the category identification numbers of successful sending are extracted and the inventory value is updated, the remote replenishment instruction sequence is reorganized, and a rapid replenishment status table of emergency supplies is generated.

[0030] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A method for intelligent management of an integrated first aid box, characterized in that: The following steps are involved: Step 1: Read and sort the data stream based on the material category identification number and access timestamp, group and lock the records by category, calculate the continuous access time interval and accumulate the quantity, classify the period interval and calculate the average, and generate a period usage characteristic set; Step 2: Based on the set of periodic usage characteristics, a long short-term memory network is used to call the usage monitoring module to set the step size. The indicator vector is rolled forward to generate a cumulative trend and record the changes. The indicator vector is rolled backward to generate a reverse trend and record the changes. The trend difference is calculated and the offset category is extracted. The sensitivity screening is performed to mark the periodic fluctuation material category to form a periodic fluctuation material category set. Step 3: Based on the set of cyclically fluctuating material categories, the material identification module is used to screen sensitivity categories, extract remaining inventory, calculate the product of usage frequency, screen out-of-stock categories, archive priority materials, and generate a replenishment priority material list; Step 4: Based on the replenishment priority material list, resource weights are set in conjunction with the replenishment management unit. Material categories are prioritized and allocated to the scheduling pool and the hierarchy is locked. Using a pointer network, urgency parameters are set to filter the sending categories and cut low-frequency categories. Combined coding is used to generate a hierarchical sending order, generating a remote replenishment instruction sequence. Step 5: Based on the remote replenishment instruction sequence, the remote delivery feedback module is called in the intelligent execution window to record the delivery status and response time, count the consecutive failure categories and queue them for retransmission, update the inventory status of the successful categories and reorganize the delivery plan, and generate a rapid replenishment status table for emergency supplies.

2. The intelligent management method of first aid integrated box according to claim 1, characterized in that: The specific steps of generating the periodic usage characteristic set are: According to the material category identification number and the material access timestamp, the material category identification number and the material access timestamp data are extracted in chronological order and invalid records are cleaned. The material category identification number and the timestamp are sorted in ascending order according to the material category number. An independent index identifier is generated for the records under each category number, and a set of access records of the same category is output; Based on the same category access record set, extract the time difference between adjacent records according to the index identifier, calculate the continuous time interval under each category and store it as a time difference set, count the number of accesses within each set of time differences and classify them into the material category index, and generate a material access interval statistics table; Based on the material usage interval statistics table, a fixed time period length is set and the usage quantity in each time period is extracted. The usage quantity is accumulated and the average value of each time period is calculated. At the same time, the material category number is marked to generate a period usage characteristic set.

3. The intelligent management method of first aid integrated box according to claim 1, characterized in that: The specific steps for generating the cyclically fluctuating material category set are as follows: Based on the periodic usage characteristic set, a long short-term memory network is used, a forward step size parameter is set, and a periodic indicator vector of the resource category is extracted. Subsequences are extracted by sliding on the indicator vector according to the step size to obtain a subsequence set. The numerical changes of each subsequence are accumulated and the direction of change is recorded. The forward trend change curves of each category are generated by superposition, and the forward trend change sequence is output; Based on the forward trend change sequence, a backward step size parameter is set, and subsequences are extracted by sliding in reverse order on the indicator vector according to the step size, and the numerical changes of each subsequence are accumulated and the direction of change is recorded. The backward trend change curves of each category are superimposed to generate the backward trend change curves, and the backward trend change sequence is output; Based on the backward trend change sequence, the forward trend value and the backward trend value of the corresponding time step are extracted, the difference between the two is calculated and the difference category is extracted, the categories whose differences exceed the sensitivity threshold are screened and assigned periodic fluctuation labels, and the periodic fluctuation material category set is output.

4. The intelligent management method for first aid integrated box according to claim 3, characterized in that: The long short-term memory network is based on the formula: ； in: represents the hidden state at the current time step t, represents the activation function, Represents the weight matrix from the hidden state to the current hidden state, represents the weight matrix input to the hidden state, represents the input vector at the current time step, represents the bias term, represents the periodic trend attenuation coefficient, represents the weighted sum of historical input vectors, represents the weight coefficient of the historical input vector, is the time step First aid supplies management data input, represents the external interference correction coefficient, Indicates external influencing factors.

5. The intelligent management method for first aid integrated box according to claim 3, characterized in that: The subsequence is extracted by sliding on the indicator vector according to the step size to obtain a subsequence set. The sliding operation is performed on the periodic indicator vector corresponding to each material category with the set forward step size parameter as the unit, and continuous data sub-segments are intercepted according to the step size. After one round of sliding, the starting position is moved forward to advance one data point, and the subsequence interception is repeated until the end of the indicator vector to form a subsequence set.

6. The intelligent management method for first aid integrated box according to claim 1, characterized in that: The specific steps for generating the replenishment priority material list are as follows: Based on the periodic fluctuation material category set, the material category number is extracted and matched with the data stored in the material identification module, the categories with a sensitivity greater than the set sensitivity threshold are screened, and the corresponding remaining inventory quantity is extracted, and a material category inventory corresponding table is established, and the sensitivity inventory data set is output; Based on the sensitivity inventory data set, historical usage records of each material category are extracted, the usage frequency is extracted and classified into the inventory quantity corresponding table, the product of inventory quantity and usage frequency is calculated and compared with the threshold, and the material categories with the product below the threshold are screened, and a list of materials in short supply is output; Based on the inventory shortage material list, the screened material category numbers are extracted, classified and archived and sorted according to the inventory shortage level, a material category priority processing index table is generated and integrated into the inventory management list, and a replenishment priority material list is output.

7. The intelligent management method of first aid integrated box according to claim 1, characterized in that: The specific steps for generating the remote replenishment instruction sequence are: Based on the replenishment priority material list, extract the material category number and set the resource allocation weight value, arrange the material categories according to priority, allocate the material categories to different scheduling pools and set corresponding hierarchical identifiers, and output the scheduling pool material hierarchical data; Based on the material hierarchical data of the scheduling pool, extract the material category number in each scheduling pool, set the urgency judgment parameters and filter the urgency categories, cut the low-frequency category records, retain the urgent materials and establish a sending index table, and output the material instruction sending classification table; Based on the material instruction sending classification table, a pointer network is used to extract the material category numbers and combine the category codes in priority order, generate a material instruction code sequence and arrange it in sequence, encapsulate it into a remote control sending data packet, and output a remote replenishment instruction sequence.

8. The intelligent management method for first aid integrated box according to claim 7, characterized in that: According to the formula: ； in: represents the probability of selecting material category m at the kth position, represents the similarity score between the k-th position output by the encoder and the m-th emergency supplies category, is the exponential function of the similarity score, m is the total number of emergency supplies categories, Indicates the weight coefficient of priority between first aid material categories, represents the relative distance measure between location k and material category m, is the external factor correction coefficient, It is the correction value of material category m affected by external factors.

9. The intelligent management method for first aid integrated box according to claim 1, characterized in that: The specific steps for generating the above are: Based on the remote replenishment instruction sequence, the sending instruction data corresponding to each material category is extracted, the instruction content is pushed one by one by setting the sending window time parameter, and the sending status and response time are recorded. Each sending record is archived and classified and summarized by material category number and instruction status to generate a sending feedback record table; Based on the delivery feedback record table, extract the delivery failure category numbers and sort them by the number of failures, filter the categories of materials with consecutive failures and establish a resending priority index, update the remaining inventory quantity by extracting the delivery success categories and generate a corresponding material status list, integrate the resending and update lists to generate an inventory reorganization task list; Based on the inventory reorganization task list, the remaining inventory status records are extracted and updated to the material inventory database. The sending order list is recompiled by extracting the resend category number, the resend instructions are pushed in sequence and the replenishment status is synchronized, the replenishment operation records and inventory changes are summarized, and a rapid replenishment status table for emergency supplies is generated.

10. An intelligent management system for first aid integrated box, characterized in that: The intelligent management method for an integrated first aid box according to any one of claims 1 to 9, wherein the system comprises: Material characteristics collection module: Based on the material category identification number and access timestamp, read the data stream, sort it in ascending order according to the material category identification number, group it by the material category identification number and lock the records, calculate the time difference between the access timestamps of two consecutive records with the same category identification number, count the number of accesses corresponding to the time difference, extract the time difference value set, classify and attribute the time difference according to the preset period interval, count the number of time differences in each period interval, calculate the average time difference in the period interval, and establish a period usage characteristics set; Cyclic Fluctuation Screening Module: Based on the set of cyclic usage characteristics, the access monitoring module is called to set the step size. Based on the long-short-term memory network structure, the cyclic mean index vector corresponding to the material category identification number is constructed. The sum of the index vectors within the step size interval is rolled forward along the time series and recorded to generate a cumulative trend. The sum of the index vectors within the step size interval is rolled backward along the time series and recorded to generate a reverse trend. The corresponding point difference between the positive cumulative trend and the reverse cumulative trend is calculated. The category identification numbers whose absolute value of the difference exceeds the set sensitivity threshold are screened, and the category identification number set with significant fluctuations is extracted to form a cyclically fluctuating material category set. Inventory priority filing module: Based on the periodic fluctuation material category set, the material identification module is called to screen the category identification numbers within the sensitivity range, extract the remaining inventory value corresponding to each category identification number, extract the corresponding usage frequency value, calculate the product of the remaining inventory value and the usage frequency value, screen the category identification numbers whose product value is lower than the preset inventory safety threshold, and establish a replenishment priority material list; A hierarchical instruction generation module extracts resource weight values ​​set by the replenishment management unit based on the replenishment priority material list, sorts the resource weight values ​​corresponding to the category identification numbers in ascending order, selects high-weight category identification numbers and adds them to the scheduling pool, sets a hierarchical lock flag for each category identification number in the scheduling pool, sets an urgency value based on the pointer network structure, filters category identification numbers with urgency values ​​above a threshold, deletes category identification numbers in the scheduling pool with urgency values ​​below the threshold, combines the category identification numbers and urgency values ​​to generate a coding vector, determines the sending order based on the arrangement order of the coding vectors, and establishes a remote replenishment instruction sequence; Remote scheduling execution module: Based on the remote replenishment instruction sequence, the remote sending feedback module is called to send the replenishment instruction to the external execution terminal, the sending status and response time of each replenishment instruction are recorded, the category identification numbers of the three consecutive failed sending are screened and added to the resending queue, the category identification numbers of the successful sending are extracted and the inventory value is updated, the remote replenishment instruction sequence is reorganized, and a rapid replenishment status table of emergency supplies is generated.

Images