Intelligent distribution system and method for dairy product production line

Abstract

The invention discloses an intelligent distribution system and method for a dairy product production line, and relates to the technical field of production process regulation and control. The multi-target genetic algorithm is improved by introducing adaptive variation changing along with the number of iterations and dynamically adjusting the weight, and the optimal production distribution scheme of the day is obtained through the improved multi-target genetic algorithm based on the estimated order data time sequence of the day; setting a judgment standard of each flow sub-sequence according to the current-day optimal production distribution scheme, and performing intra-day monitoring on each flow sub-sequence according to the judgment standard; carrying out dynamic weight adjustment and redistribution operation according to the intra-day monitoring result, and obtaining a new optimal production distribution scheme on the current day; the whole stability estimation operation is carried out according to the intraday monitoring result, the early warning process sub-sequence is generated, the temporary alternative distribution operation is carried out on the early warning process sub-sequence, and the real-time performance and efficiency of intelligent distribution of the dairy product production line are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of production process control, and in particular to an intelligent distribution system and method for a dairy product production line. Background Art

[0002] A Chinese patent with publication number CN115619171A discloses a method and system for allocating production and maintenance coupled tasks that take into account the operating status of equipment, including the following steps: a digital twin platform, which simulates the allocation plan obtained by the task allocation decision module to verify its feasibility; an equipment model management module, which establishes a reliability model for the equipment status and establishes a numerical model for the equipment production process; an order receiving module, which decomposes the order into production tasks for each equipment; a maintenance strategy management module, which manages the maintenance strategy and selects the corresponding maintenance strategy for the equipment based on the current equipment reliability; a task allocation decision module, which unifies the planning of production tasks and maintenance tasks and gives an allocation plan; an allocation result interaction module, which outputs the final allocation plan and interacts with the physical production line to provide production guidance for the physical production line.

[0003] In the process of dairy production, the reasonable allocation of production lines is crucial to improving production efficiency, reducing costs and ensuring product quality. The traditional allocation method of dairy production lines often relies on manual experience, and there are problems such as unreasonable allocation, low efficiency, and difficulty in adapting to order changes. Manual allocation is difficult to take into account various complex factors such as equipment status, raw material supply, order urgency, etc. in real time, which can easily lead to production bottlenecks, idle or overused equipment, etc., thus affecting the production cycle of dairy products and the economic benefits of the enterprise. With the intensification of market competition and the improvement of consumers' requirements for the quality and timeliness of dairy products, traditional allocation methods can no longer meet the needs of modern dairy production. Therefore, an intelligent allocation method for dairy production lines is urgently needed. Summary of the invention

[0004] In order to solve the above technical problems, the object of the present invention is to provide an intelligent distribution method for a dairy product production line, comprising the following steps: Step s1: Build a day-ahead demand forecasting model and output the day-ahead estimated order data time series; Step s2: Obtain the current dairy product production process information, set IoT points according to the process information, collect actual production data at each point, build a production directed graph, and obtain the preferred equipment and alternative equipment for each process subsequence; Step s3: Introduce adaptive mutation that changes with the number of iterations and dynamically adjust weights to improve the multi-objective genetic algorithm, and obtain the best production allocation plan for the day based on the time series of the estimated order data of the day through the improved multi-objective genetic algorithm; Step s4: setting the judgment criteria for each process subsequence according to the optimal production allocation plan for the day, and performing intraday monitoring on each process subsequence according to the judgment criteria; Step s5: Perform dynamic weight adjustment and reallocation operations based on the intraday monitoring results to obtain a new optimal production allocation plan for the day; Step s6: Perform an overall stability estimation operation based on the intraday monitoring results, generate an early warning process subsequence based on the overall stability estimation operation results, and perform a temporary alternative allocation operation on the early warning process subsequence.

[0005] Furthermore, the process of building a day-ahead demand forecasting model and outputting the time series of the day's estimated order data includes: Collect historical multi-source data, including historical order data, meteorological data, and raw material supply data, etc., extract time features from the historical multi-source data, obtain a time series sequence of changes in the historical multi-source data, perform STL decomposition on the time series sequence of changes, obtain trend components and seasonal components, build a day-ahead demand prediction model based on deep learning, use the trend components and seasonal components as training data, use the training data to train the day-ahead demand prediction model, and output the trained day-ahead demand prediction model;

[0006] The calculation formula for decomposing the changing time series into trend component and seasonal component by STL decomposition method is: ; in, is the value of the changing time series at time point t, is the value of the trend component at time point t, is the value of the seasonal component at time point t, is the value of the random component at time point t. The random component represents the unpredictable part remaining after removing the trend component and the periodic component. This part of the data reflects the random fluctuations that cannot be explained by the model, including outliers, measurement errors, unidentified patterns and other unknown factors. Therefore, the random component is removed before training the day-ahead demand forecasting model. Output the estimated order data time series for the day based on the day-ahead demand forecasting model.

[0007] Furthermore, the process of obtaining the current dairy production process information, setting the IoT points according to the process information, and constructing the production directed graph includes: Obtain the process flow characteristics of the current dairy product production equipment, obtain process information according to the process flow characteristics, divide the dairy product production line according to the process information, divide it into several process sub-sequences, and deploy an Internet of Things node in each process sub-sequence. The Internet of Things node is used to collect actual production data, and the actual production data includes an actual equipment state feature time series sequence and a standard equipment state feature time series sequence; The dairy production equipment in each process subsequence and the assembly order and assembly relationship between each dairy production equipment are extracted, each process subsequence is used as a node of the production directed graph, and the assembly order and assembly relationship between each dairy production equipment is used as the connection relationship between the nodes to construct a production directed graph.

[0008] Furthermore, the process of obtaining the preferred equipment and the candidate equipment for each process subsequence includes: Extract the production equipment data of each dairy product production equipment in each process subsequence, the production equipment data includes equipment operation status data: whether the equipment is in operation, shutdown, standby, fault, etc.; equipment operation parameters, such as temperature, pressure, speed, flow, etc.; equipment maintenance data: equipment maintenance records, including the last maintenance time, maintenance content, next maintenance time estimate, etc.; equipment failure records, such as failure time, failure phenomenon, failure cause, maintenance time, etc.; equipment capacity data: actual equipment capacity, capacity utilization rate, etc. The production equipment data of each dairy product production equipment is used as an evaluation index, and the index weight of the evaluation index is set. The membership matrix of each dairy product production equipment to the preset reliability level is obtained through fuzzy comprehensive evaluation; The reliability level of each dairy production equipment is obtained according to the membership matrix and indicator weight, and the reliability level of the dairy production equipment is compared with the preset reliability level threshold. The dairy production equipment with a reliability level greater than or equal to the reliability level threshold is marked as the preferred equipment, and the dairy production equipment with a reliability level less than the reliability level threshold is marked as the alternative equipment.

[0009] Furthermore, the process of obtaining the reliability level of each dairy production equipment according to the membership matrix and the index weight includes: The indicator weights and the membership matrix of the evaluation indicators are integrated through a formula to obtain a fuzzy comprehensive evaluation matrix of the evaluation indicators, the membership of each dairy product production equipment to different reliability levels is obtained according to the fuzzy comprehensive evaluation matrix, the reliability level with the highest membership corresponding to each dairy product production equipment is screened out, and the reliability level with the highest membership corresponding to each dairy product production equipment is used as the reliability level of each dairy product production equipment; Wherein, the formula is: ; in, is the fuzzy comprehensive evaluation matrix of the evaluation index, is the indicator weight of the evaluation indicator, is the membership matrix, " represents the multiplication of the weight matrix of the evaluation index and the elements at the corresponding positions of the membership matrix, It is a weighted parameter for controlling the balance between the weight matrix and the membership matrix in the fuzzy comprehensive evaluation matrix of the evaluation index.

[0010] Furthermore, the multi-objective genetic algorithm is improved by introducing adaptive mutation that changes with the number of iterations and dynamically adjusting weights. The process of obtaining the optimal production allocation plan for the day through the improved multi-objective genetic algorithm based on the time series of the estimated order data for the day includes: Construct several objective functions, construct constraint condition formulas based on the time series of the estimated order data of the day and the production equipment data of the preferred equipment, preset the initial weight coefficients of several objective functions, construct fitness functions based on several objective functions and initial weight coefficients, preset several production allocation plans, perform chromosome encoding and initialize populations for several production allocation plans, generate initial populations, introduce adaptive mutation that changes with the number of iterations and dynamically adjust weights to improve the multi-objective genetic algorithm; Based on the initial population, fitness function and constraint formula, the optimal production allocation plan for the day is obtained through an improved multi-objective genetic algorithm.

[0011] Furthermore, the fitness function is specifically: ; in, represents the total cost of the current production allocation plan, Indicates the preset reference value, represents the equipment utilization rate of the current production allocation plan, Indicates the theoretical maximum utilization of the equipment, The production cycle of the current production allocation plan, Indicates the minimum cycle required by the order, Quality indicators of the current production allocation plan, Indicates the indicator threshold, , , and Represents the initial weight coefficient.

[0012] Several objective functions are specifically: ; Where C is the total cost, is the unit cost of the ith raw material, is its usage, n1 represents the total number of raw materials, is the unit time operating cost of the j-th dairy production equipment, is the running time, n2 is the total number of devices, is the hourly wage of the employee in position z, is the working hours, n3 represents the total number of original positions, is the unit time holding cost of the v-th inventory, is the inventory quantity, n4 represents the total number of inventory types, It's inventory time.

[0013] ; is the equipment utilization rate, is the actual operating time of the j-th dairy production equipment, is its runnable time.

[0014] ; Where T is the production cycle, is the process time of the pth process subsequence, is the device switching time of the pth process subsequence, is the waiting time of the pth process subsequence.

[0015] ; in, It represents the quality index, and is measured by the fluctuation range of the product quality index, such as the standard deviation of key indicators such as the protein and fat content of dairy products, so as to minimize the fluctuation of the quality index. is the rth product quality index value, is the average value and n5 represents the total number of products.

[0016] The constraint formula is as follows: ; ; ; ; ; in, is the production quantity of the wth type of dairy product, is the order quantity required, is the production completion time of the wth dairy product, is the delivery time of the wth dairy product order, is the production quantity of the jth equipment, is its maximum capacity, is the inventory quantity of the i-th raw material, is the safety stock quantity, is the inventory quantity of the wth type of dairy product, is the warehouse storage capacity.

[0017] The calculation formula of adaptive mutation that changes with the number of iterations is: ; Among them, D is the population diversity index, is the mutation rate, , are preset parameters.

[0018] The specific process of obtaining the optimal production allocation plan for the day through the improved multi-objective genetic algorithm includes: The production plan is encoded as a chromosome. For example, a string of numbers is used to represent the type, quantity and production order of dairy products produced by each device. A certain number of chromosomes are randomly generated to form an initial population. Each chromosome represents a possible production allocation plan. The fitness function is obtained, and the initial weight coefficient of the fitness function is dynamically adjusted according to the current actual production data. Then, the tournament selection method is used to select chromosomes with higher fitness from the current population as parents. The parent chromosomes are crossed and some genes are exchanged to generate new daughter chromosomes. This simulates the exchange of biological genetic genes and generates new production plans. The daughter chromosomes are mutated and some genes are randomly changed to increase population diversity and avoid falling into local optimality. The above steps are repeated and iterated continuously until the termination conditions are met, such as reaching the maximum number of iterations or the fitness value is no longer significantly improved, and the best production allocation plan for the day is output.

[0019] Furthermore, the judgment criteria of each process subsequence are set according to the optimal production allocation plan of the day, and the process of performing intraday monitoring on each process subsequence according to the judgment criteria includes: The judgment criteria include standard order feature timing sequence, standard equipment status feature timing sequence, order feature deviation threshold interval and equipment status feature deviation threshold interval. The order features include order quantity, product type, priority, etc. The equipment status features include energy consumption change, temperature, pressure and other parameters. The timing sequence interval is one hour. The actual order feature timing sequence and the actual equipment status feature timing sequence of each process subsequence are collected and acquired according to the IoT node. Compare the actual order feature timing sequence of each process subsequence with the standard order feature timing sequence to obtain the order feature deviation, and compare the actual equipment status feature timing sequence of each process subsequence with the standard equipment status feature timing sequence to obtain the equipment status feature deviation; Compare the order feature deviation and equipment status feature deviation of each process subsequence with the corresponding order feature deviation threshold interval and equipment status feature deviation threshold interval respectively. If there is a process subsequence whose order feature deviation is not within the order feature deviation threshold interval or whose equipment status feature deviation is not within the equipment status feature deviation threshold, perform dynamic weight adjustment and reallocation operations; If the order feature deviations of each process subsequence are within the order feature deviation threshold and the equipment state feature deviations are within the equipment state feature deviation threshold, then threshold points are selected within the equipment state feature deviation threshold to divide sub-intervals of different stability levels, and the stability level of the process subsequence is obtained according to the equipment state feature deviation of the process subsequence.

[0020] Furthermore, the process of dynamically adjusting and reallocating weights based on the intraday monitoring results to obtain a new optimal production allocation plan for the day includes: According to the actual order feature time series of the process subsequence and the actual equipment status feature time series, the demand urgency coefficient and equipment health coefficient of the process subsequence are obtained; The calculation formula of demand urgency coefficient is: ; in, represents the demand urgency coefficient, Indicates the order quantity, Indicates the remaining time of the production cycle. represents the historical average order volume, Indicates the production cycle; The calculation formula of equipment health factor is: ; in, represents the equipment health factor, represents the historical failure rate, Indicates energy consumption, Indicates rated capacity; Obtain the historical production data of each process subsequence, and obtain the sensitivity coefficients of various types of indicators in order characteristics and equipment status characteristics corresponding to several objective functions based on the historical production data; The process of obtaining the sensitivity coefficients of various types of indicators corresponding to several objective functions includes: Taking the cost-related daily order information variable as an example, assuming that the cost objective function is C=f( , ..., ),in (i=1,2,...,n) is the order information variable of the day related to cost, such as order quantity, raw material price, production time, etc. Sensitivity coefficient for cost target C The calculation formula is: ; in, is a variable A tiny increment of Construct weight adjustment rules based on the sensitivity coefficients of various types of indicators corresponding to several objective functions; The calculation formula of the weight adjustment rule is:

[0021] in, represents the initial weight coefficient of the i-th objective function, Indicator For the sensitivity coefficient of objective function i, represents the dynamic adjustment factor of indicator j (set according to expert experience), m represents the total number of objective functions, and n represents the total number of indicator types; A dynamic weight calculation model is constructed based on the weight adjustment rules. The dynamic weight calculation model includes a data perception layer: real-time collection of demand urgency coefficient, equipment health coefficient, order feature deviation and equipment status feature deviation; a weight decision layer: dynamically adjust the weight based on the weight adjustment rules; a feedback optimization layer: verify the rationality of the weight through actual production results and iterate the model, input the demand urgency coefficient, equipment health coefficient, order feature deviation and equipment status feature deviation of the process subsequence into the dynamic weight calculation model, and update the initial weight coefficients corresponding to several objective functions in the fitness function according to the dynamic weight calculation model; Set a feedback cycle, input the order feature deviation and equipment status feature deviation of the process subsequence into the dynamic weight calculation model at the end timestamp of the feedback cycle, and update the weight adjustment rules of the dynamic weight calculation model; The specific process of updating the weight adjustment rules of the dynamic weight calculation model is as follows: using exponentially weighted moving average (EWMA) to update the weight adjustment rules: ;in, represents the forgetting coefficient, represents the order characteristic deviation, Indicates the characteristic deviation of the equipment status; Based on the fitness function that completes the dynamic weight adjustment and the updated constraint formula, the improved multi-objective genetic algorithm is used to obtain a new optimal production allocation plan for the day. According to the new optimal production allocation plan for the day, the standard order feature timing sequence, standard equipment status feature timing sequence, order feature deviation threshold interval and equipment status feature deviation threshold interval of each process subsequence are reset.

[0022] Furthermore, the process of performing overall stability estimation based on the intraday monitoring results includes: Import the stability level of each process subsequence into the nodes of each process subsequence in the production directed graph, construct the adjacency matrix of the production directed graph, obtain the stability influence coefficient of each node in the production directed graph on other nodes according to the adjacency matrix, and set the weight label of each node according to the stability influence coefficient of each node on other nodes; A full-process stability prediction model is built based on a graph convolutional neural network, which learns and represents the production directed graph and outputs the overall estimated stability.

[0023] Furthermore, the calculation formula for obtaining the stability influence coefficient of each node on other nodes in the production directed graph is:

[0024] in, , They represent the stability influence coefficient of the process subsequence on other process subsequences after the i+1th and ith iterations respectively; = , whose specific value is obtained through the node adjacency matrix; represents the damping factor; represents the state transition matrix; E represents an N×1 matrix whose elements are all 1; NUM represents the number of process subsequences; The production directed graph is represented and learned through graph convolutional neural network. For a pair of nodes (a, b) connected in the production directed graph, the stability coefficient of node a is obtained. and the stability coefficient of node b ,Will , As the initial feature vector, the neighbor aggregation mechanism of node a is represented as follows: ; Among them, the final update of node a is expressed as, represents the activation function, G represents the parameter matrix of feature transformation, represents the influence coefficient of node a on the stability of node b, Represents the set of adjacent nodes of node a; According to the above formula, the vector representation of node a is generated Then the vector representation of the next connected node is performed, and so on. After all the nodes in the encapsulated directed graph are represented, the vector inner product of the nodes is calculated to obtain the overall estimated stability.

[0025] Furthermore, the process of generating a warning process subsequence according to the overall stability estimation operation result and performing a temporary alternative allocation operation on the warning process subsequence includes: The overall estimated stability is compared with the preset overall stability threshold. If the overall estimated stability is less than the overall stability threshold, the overall vulnerability coefficient of each process subsequence is obtained according to the stability level of each process subsequence and the stability influence coefficient of each process subsequence on other process subsequences, and the overall vulnerability coefficient threshold is preset. The overall vulnerability coefficient of each process subsequence is compared with the overall vulnerability coefficient threshold, and the process subsequence with an overall vulnerability coefficient greater than the overall vulnerability coefficient threshold is marked as a warning process subsequence. A temporary alternative allocation operation is performed on the warning process subsequence, and the operation of the preferred equipment in the warning process subsequence is suspended. The alternative equipment temporarily takes over the optimal production allocation plan of the day for the preferred equipment, and arranges relevant personnel to inspect and repair the preferred equipment.

[0026] Furthermore, the specific calculation formula for obtaining the overall vulnerability coefficient of each process subsequence is:

[0027] in, represents the overall vulnerability coefficient of process subsequence j; represents the stability influence coefficient of process subsequence j; It represents the influence coefficient of process subsequence j on the stability of other process subsequences; Stability influence coefficients of several other process subsequences.

[0028] An intelligent distribution system for a dairy product production line includes a monitoring center, wherein the monitoring center is communicatively connected to a day-ahead demand estimation module, a data processing module, a distribution plan planning module, an intraday monitoring module, a dynamic adjustment module, and a temporary alternative distribution module; The day-ahead demand prediction module is used to build a day-ahead demand prediction model and output the time series of the day's estimated order data; The data processing module is used to obtain the current dairy production process information, set the IoT points according to the process information, collect the actual production data of each point, build a production directed graph, and obtain the preferred equipment and alternative equipment for each process subsequence; The allocation plan planning module is used to improve the multi-objective genetic algorithm by introducing adaptive mutations that change with the number of iterations and dynamically adjusting weights. The optimal production allocation plan for the day is obtained through the improved multi-objective genetic algorithm based on the time series of the estimated order data for the day. The intraday monitoring module is used to set the judgment criteria for each process subsequence according to the optimal production allocation plan of the day, and perform intraday monitoring on each process subsequence according to the judgment criteria; The dynamic adjustment module is used to perform dynamic weight adjustment and reallocation operations based on the intraday monitoring results to obtain the new optimal production allocation plan for the day; The temporary alternative allocation module is used to perform overall stability estimation operations according to the intraday monitoring results, generate early warning process subsequences according to the overall stability estimation operation results, and perform temporary alternative allocation operations on the early warning process subsequences.

[0029] Compared with the prior art, the present invention has the following beneficial effects: 1. Integrating historical multi-source data can more comprehensively capture the factors that affect the demand for dairy products compared to traditional forecasting methods that rely solely on single sales data. For example, combined with high temperature information in meteorological data, the peak demand for cold dairy products can be predicted in advance, which greatly improves the forecast accuracy and reduces overproduction or underproduction caused by demand forecast deviations. Through time feature extraction and STL decomposition, the trend, seasonality and cyclical components in historical data can be effectively separated. The trend component can clearly show the long-term demand trend of the market, and the seasonal component accurately reflects the demand fluctuation pattern in different periods, providing more valuable data features for model training. Compared with forecasting models that do not use this technology, seasonal demand fluctuations are captured more accurately, and the error rate can be reduced by 15%-20%.

[0030] 2. Fuzzy comprehensive evaluation to achieve equipment classification: Use fuzzy comprehensive evaluation methods to evaluate equipment reliability based on production equipment data, and divide preferred equipment and alternative equipment. This refined management method enables a more reasonable allocation of equipment maintenance resources. For preferred equipment, preventive maintenance strategies can be adopted to ensure its efficient and stable operation; for alternative equipment, maintenance or update plans can be planned in advance, and the overall equipment failure rate can be reduced by 30%-40%. In production allocation, equipment with high reliability is given priority to ensure the stability of the production process and product quality. At the same time, as an emergency reserve, alternative equipment can be quickly switched when the preferred equipment fails, reducing production interruption time and improving the resilience of the production system.

[0031] 3. Improve the multi-objective genetic algorithm to improve performance: Introduce adaptive mutation and dynamic weight adjustment mechanism that changes with the number of iterations to improve the multi-objective genetic algorithm. Adaptive mutation can maintain population diversity in the early stage of the algorithm and accelerate convergence to the optimal solution in the later stage, which can improve the algorithm search efficiency by 30%-40%. Dynamic weight adjustment can flexibly balance multiple objectives such as cost, efficiency, and quality according to real-time production conditions. Compared with the traditional fixed weight algorithm, the comprehensive production efficiency is improved by 15%-20%.

[0032] 4. Determination criteria for refined monitoring: Set the determination criteria for each process subsequence, covering the standard sequence and deviation threshold range of order characteristics and equipment status characteristics, and conduct real-time monitoring of the production process within the day. By comparing the actual order feature timing sequence with the standard sequence, and the actual equipment status feature timing sequence with the standard sequence, production deviations can be accurately identified. Compared with manual inspections and extensive monitoring, the accuracy of deviation identification is significantly improved.

[0033] 5. Dynamic weight adjustment and solution reconstruction: According to the monitoring results, the objective function weight is adjusted in real time and the production tasks are reallocated to obtain the new optimal production allocation plan. For example, when the urgency coefficient of order demand changes, the efficiency target weight is adjusted in time to give priority to the delivery of urgent orders. This dynamic adjustment mechanism enables the production system to respond quickly to market and production changes, and the production plan adjustment time is shortened from several hours to within half an hour.

[0034] 6. Graph convolutional neural network to evaluate overall stability: The stability level of each process subsequence is imported into the production directed graph, and the stability of the entire production process is evaluated based on the graph convolutional neural network by constructing an adjacency matrix and a full-process stability estimation model. Compared with traditional empirical judgment or local stability assessment methods, it can predict potential risks more comprehensively and accurately.

[0035] 7. Early warning and emergency allocation mechanism: Based on the overall stability estimation results, a warning process subsequence is generated and a temporary alternative allocation operation is performed on it. When abnormal fluctuations occur in production, the emergency plan can be quickly activated to ensure production continuity and reduce economic losses caused by production interruptions.

[0036] 8. Multi-module collaborative optimization: The day-ahead demand forecast module, data processing module, allocation plan planning module, intraday monitoring module, dynamic adjustment module and temporary alternative allocation module work together to achieve digital management of the entire chain from demand forecasting to production execution. Data is shared in real time and work collaboratively between modules, significantly improving overall operational efficiency compared to traditional production management systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 This is a schematic diagram of an intelligent distribution method for a dairy product production line according to an embodiment of the present application.

[0038] Figure 2 This is a schematic diagram of an intelligent distribution system for a dairy product production line according to an embodiment of the present application. DETAILED DESCRIPTION

[0039] The following is a clear and complete description of the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0040] like Figure 1 As shown, a method for intelligent distribution of a dairy product production line comprises the following steps: Step s1: Build a day-ahead demand forecasting model and output the day-ahead estimated order data time series; Step s2: Obtain the current dairy product production process information, set IoT points according to the process information, collect actual production data at each point, build a production directed graph, and obtain the preferred equipment and alternative equipment for each process subsequence; Step s3: Introduce adaptive mutation that changes with the number of iterations and dynamically adjust weights to improve the multi-objective genetic algorithm, and obtain the best production allocation plan for the day based on the time series of the estimated order data of the day through the improved multi-objective genetic algorithm; Step s4: setting the judgment criteria for each process subsequence according to the optimal production allocation plan for the day, and performing intraday monitoring on each process subsequence according to the judgment criteria; Step s5: Perform dynamic weight adjustment and reallocation operations based on the intraday monitoring results to obtain a new optimal production allocation plan for the day; Step s6: Perform an overall stability estimation operation based on the intraday monitoring results, generate an early warning process subsequence based on the overall stability estimation operation results, and perform a temporary alternative allocation operation on the early warning process subsequence.

[0041] It should be further explained that, in the specific implementation process, the process of building a day-ahead demand forecasting model and outputting the time series of the day's estimated order data includes: Collect historical multi-source data, including historical order data, meteorological data, and raw material supply data, etc., extract time features from the historical multi-source data, obtain a time series sequence of changes in the historical multi-source data, perform STL decomposition on the time series sequence of changes, obtain trend components and seasonal components, build a day-ahead demand prediction model based on deep learning, use the trend components and seasonal components as training data, use the training data to train the day-ahead demand prediction model, and output the trained day-ahead demand prediction model; The calculation formula for decomposing the changing time series into trend component and seasonal component by STL decomposition method is: ; in, is the value of the changing time series at time point t, is the value of the trend component at time point t, is the value of the seasonal component at time point t, is the value of the random component at time point t. The random component represents the unpredictable part remaining after removing the trend component and the periodic component. This part of the data reflects the random fluctuations that cannot be explained by the model, including outliers, measurement errors, unidentified patterns and other unknown factors. Therefore, the random component is removed before training the day-ahead demand forecasting model. Output the estimated order data time series for the day based on the day-ahead demand forecasting model.

[0042] It should be further explained that, in the specific implementation process, the process of obtaining the current dairy production process information, setting the IoT points according to the process information, and constructing the production directed graph includes: Obtain the process flow characteristics of the current dairy product production equipment, obtain process information according to the process flow characteristics, divide the dairy product production line according to the process information, divide it into several process sub-sequences, and deploy an Internet of Things node in each process sub-sequence. The Internet of Things node is used to collect actual production data, and the actual production data includes an actual equipment state feature time series sequence and a standard equipment state feature time series sequence; The dairy production equipment in each process subsequence and the assembly order and assembly relationship between each dairy production equipment are extracted, each process subsequence is used as a node of the production directed graph, and the assembly order and assembly relationship between each dairy production equipment is used as the connection relationship between the nodes to construct a production directed graph.

[0043] It should be further explained that, in the specific implementation process, the process of obtaining the preferred equipment and alternative equipment for each process subsequence includes: Extract the production equipment data of each dairy product production equipment in each process subsequence, the production equipment data includes equipment operation status data: whether the equipment is in operation, shutdown, standby, fault, etc.; equipment operation parameters, such as temperature, pressure, speed, flow, etc.; equipment maintenance data: equipment maintenance records, including the last maintenance time, maintenance content, next maintenance time estimate, etc.; equipment failure records, such as failure time, failure phenomenon, failure cause, maintenance time, etc.; equipment capacity data: actual equipment capacity, capacity utilization rate, etc. The production equipment data of each dairy product production equipment is used as an evaluation index, and the index weight of the evaluation index is set. The membership matrix of each dairy product production equipment to the preset reliability level is obtained through fuzzy comprehensive evaluation; The reliability level of each dairy production equipment is obtained according to the membership matrix and indicator weight, and the reliability level of the dairy production equipment is compared with the preset reliability level threshold. The dairy production equipment with a reliability level greater than or equal to the reliability level threshold is marked as the preferred equipment, and the dairy production equipment with a reliability level less than the reliability level threshold is marked as the alternative equipment.

[0044] It should be further explained that, in the specific implementation process, the process of obtaining the reliability level of each dairy production equipment according to the membership matrix and the index weight includes: The indicator weights and the membership matrix of the evaluation indicators are integrated through a formula to obtain a fuzzy comprehensive evaluation matrix of the evaluation indicators, the membership of each dairy product production equipment to different reliability levels is obtained according to the fuzzy comprehensive evaluation matrix, the reliability level with the highest membership corresponding to each dairy product production equipment is screened out, and the reliability level with the highest membership corresponding to each dairy product production equipment is used as the reliability level of each dairy product production equipment; Wherein, the formula is: ; in, is the fuzzy comprehensive evaluation matrix of the evaluation index, is the indicator weight of the evaluation indicator, is the membership matrix, " represents the multiplication of the weight matrix of the evaluation index and the elements at the corresponding positions of the membership matrix, It is a weighted parameter for controlling the balance between the weight matrix and the membership matrix in the fuzzy comprehensive evaluation matrix of the evaluation index.

[0045] It should be further explained that, in the specific implementation process, the adaptive mutation and dynamic weight adjustment that change with the number of iterations are introduced to improve the multi-objective genetic algorithm. The process of obtaining the best production allocation plan for the day through the improved multi-objective genetic algorithm based on the time series of the estimated order data of the day includes: Construct several objective functions, construct constraint condition formulas based on the time series of the estimated order data of the day and the production equipment data of the preferred equipment, preset the initial weight coefficients of several objective functions, construct fitness functions based on several objective functions and initial weight coefficients, preset several production allocation plans, perform chromosome encoding and initialize populations for several production allocation plans, generate initial populations, introduce adaptive mutation that changes with the number of iterations and dynamically adjust weights to improve the multi-objective genetic algorithm; Based on the initial population, fitness function and constraint formula, the optimal production allocation plan for the day is obtained through an improved multi-objective genetic algorithm.

[0046] It should be further explained that, in the specific implementation process, the fitness function is specifically: ; in, represents the total cost of the current production allocation plan, Indicates the preset reference value, represents the equipment utilization rate of the current production allocation plan, Indicates the theoretical maximum utilization of the equipment, The production cycle of the current production allocation plan, Indicates the minimum cycle required by the order, Quality indicators of the current production allocation plan, Indicates the indicator threshold, , , and Represents the initial weight coefficient.

[0047] Several objective functions are specifically: ; Where C is the total cost, is the unit cost of the ith raw material, is its usage, n1 represents the total number of raw materials, is the unit time operating cost of the j-th dairy production equipment, is the running time, n2 is the total number of devices, is the hourly wage of the employee in position z, is the working hours, n3 represents the total number of original positions, is the unit time holding cost of the v-th inventory, is the inventory quantity, n4 represents the total number of inventory types, It's inventory time.

[0048] ; is the equipment utilization rate, is the actual operating time of the j-th dairy production equipment, is its runnable time.

[0049] ; Where T is the production cycle, is the process time of the pth process subsequence, is the device switching time of the pth process subsequence, is the waiting time of the pth process subsequence.

[0050] ; in, It represents the quality index, and is measured by the fluctuation range of the product quality index, such as the standard deviation of key indicators such as the protein and fat content of dairy products, so as to minimize the fluctuation of the quality index. is the rth product quality index value, is the average value and n5 represents the total number of products.

[0051] The constraint formula is as follows: ; ; ; ; ; in, is the production quantity of the wth type of dairy product, is the order quantity required, is the production completion time of the wth dairy product, is the delivery time of the wth dairy product order, is the production quantity of the jth equipment, is its maximum capacity, is the inventory quantity of the i-th raw material, is the safety stock quantity, is the inventory quantity of the wth type of dairy product, is the warehouse storage capacity.

[0052] The calculation formula of adaptive mutation that changes with the number of iterations is: ; Among them, D is the population diversity index, is the mutation rate, , are preset parameters.

[0053] The specific process of obtaining the optimal production allocation plan for the day through the improved multi-objective genetic algorithm includes: The production plan is encoded as a chromosome. For example, a string of numbers is used to represent the type, quantity and production order of dairy products produced by each device. A certain number of chromosomes are randomly generated to form an initial population. Each chromosome represents a possible production allocation plan. The fitness function is obtained, and the initial weight coefficient of the fitness function is dynamically adjusted according to the current actual production data. Then, the tournament selection method is used to select chromosomes with higher fitness from the current population as parents. The parent chromosomes are crossed and some genes are exchanged to generate new daughter chromosomes. This simulates the exchange of biological genetic genes and generates new production plans. The daughter chromosomes are mutated and some genes are randomly changed to increase population diversity and avoid falling into local optimality. The above steps are repeated and iterated continuously until the termination conditions are met, such as reaching the maximum number of iterations or the fitness value is no longer significantly improved, and the best production allocation plan for the day is output.

[0054] It should be further explained that, in the specific implementation process, the judgment criteria of each process subsequence are set according to the best production allocation plan of the day, and the process of intraday monitoring of each process subsequence according to the judgment criteria includes: The judgment criteria include standard order feature timing sequence, standard equipment status feature timing sequence, order feature deviation threshold interval and equipment status feature deviation threshold interval. The order features include order quantity, product type, priority, etc. The equipment status features include energy consumption change, temperature, pressure and other parameters. The timing sequence interval is one hour. The actual order feature timing sequence and the actual equipment status feature timing sequence of each process subsequence are collected and acquired according to the IoT node. Compare the actual order feature timing sequence of each process subsequence with the standard order feature timing sequence to obtain the order feature deviation, and compare the actual equipment status feature timing sequence of each process subsequence with the standard equipment status feature timing sequence to obtain the equipment status feature deviation; Compare the order feature deviation and equipment status feature deviation of each process subsequence with the corresponding order feature deviation threshold interval and equipment status feature deviation threshold interval respectively. If there is a process subsequence whose order feature deviation is not within the order feature deviation threshold interval or whose equipment status feature deviation is not within the equipment status feature deviation threshold, perform dynamic weight adjustment and reallocation operations; If the order feature deviations of each process subsequence are within the order feature deviation threshold and the equipment state feature deviations are within the equipment state feature deviation threshold, then threshold points are selected within the equipment state feature deviation threshold to divide sub-intervals of different stability levels, and the stability level of the process subsequence is obtained according to the equipment state feature deviation of the process subsequence.

[0055] It should be further explained that, in the specific implementation process, the dynamic weight adjustment and reallocation operation is performed according to the intraday monitoring results, and the process of obtaining a new optimal production allocation plan for the day includes: According to the actual order feature time series of the process subsequence and the actual equipment status feature time series, the demand urgency coefficient and equipment health coefficient of the process subsequence are obtained; The calculation formula of demand urgency coefficient is: ; in, represents the demand urgency coefficient, Indicates the order quantity, Indicates the remaining time of the production cycle. represents the historical average order volume, Indicates the production cycle; The calculation formula of equipment health factor is: ; in, represents the equipment health factor, represents the historical failure rate, Indicates energy consumption, Indicates rated capacity; Obtain the historical production data of each process subsequence, and obtain the sensitivity coefficients of various types of indicators in order characteristics and equipment status characteristics corresponding to several objective functions based on the historical production data; The process of obtaining the sensitivity coefficients of various types of indicators corresponding to several objective functions includes: Taking the cost-related daily order information variable as an example, assuming that the cost objective function is C=f( , ..., ),in (i=1,2,...,n) is the order information variable of the day related to cost, such as order quantity, raw material price, production time, etc. Sensitivity coefficient for cost target C The calculation formula is: ; in, is a variable A tiny increment of Construct weight adjustment rules based on the sensitivity coefficients of various types of indicators corresponding to several objective functions; The calculation formula of the weight adjustment rule is:

[0056] in, represents the initial weight coefficient of the i-th objective function, Indicator For the sensitivity coefficient of objective function i, represents the dynamic adjustment factor of indicator j (set according to expert experience), m represents the total number of objective functions, and n represents the total number of indicator types; A dynamic weight calculation model is constructed based on the weight adjustment rules. The dynamic weight calculation model includes a data perception layer: real-time collection of demand urgency coefficient, equipment health coefficient, order feature deviation and equipment status feature deviation; a weight decision layer: dynamically adjust the weight based on the weight adjustment rules; a feedback optimization layer: verify the rationality of the weight through actual production results and iterate the model, input the demand urgency coefficient, equipment health coefficient, order feature deviation and equipment status feature deviation of the process subsequence into the dynamic weight calculation model, and update the initial weight coefficients corresponding to several objective functions in the fitness function according to the dynamic weight calculation model; Set a feedback cycle, input the order feature deviation and equipment status feature deviation of the process subsequence into the dynamic weight calculation model at the end timestamp of the feedback cycle, and update the weight adjustment rules of the dynamic weight calculation model; The specific process of updating the weight adjustment rules of the dynamic weight calculation model is as follows: using exponentially weighted moving average (EWMA) to update the weight adjustment rules: ;in, represents the forgetting coefficient, represents the order characteristic deviation, Indicates the characteristic deviation of the equipment status; Based on the fitness function that completes the dynamic weight adjustment and the updated constraint formula, the improved multi-objective genetic algorithm is used to obtain a new optimal production allocation plan for the day. According to the new optimal production allocation plan for the day, the standard order feature timing sequence, standard equipment status feature timing sequence, order feature deviation threshold interval and equipment status feature deviation threshold interval of each process subsequence are reset.

[0057] It should be further explained that, in the specific implementation process, the process of conducting overall stability estimation based on the intraday monitoring results includes: Import the stability level of each process subsequence into the nodes of each process subsequence in the production directed graph, construct the adjacency matrix of the production directed graph, obtain the stability influence coefficient of each node in the production directed graph on other nodes according to the adjacency matrix, and set the weight label of each node according to the stability influence coefficient of each node on other nodes; A full-process stability prediction model is built based on a graph convolutional neural network, which learns and represents the production directed graph and outputs the overall estimated stability.

[0058] It should be further explained that in the specific implementation process, the calculation formula for obtaining the stability influence coefficient of each node in the production directed graph on other nodes is:

[0059] in, , They represent the stability influence coefficient of the process subsequence on other process subsequences after the i+1th and ith iterations respectively; = , whose specific value is obtained through the node adjacency matrix; represents the damping factor; represents the state transition matrix; E represents an N×1 matrix whose elements are all 1; NUM represents the number of process subsequences; The production directed graph is represented and learned through graph convolutional neural network. For a pair of nodes (a, b) connected in the production directed graph, the stability coefficient of node a is obtained. and the stability coefficient of node b ,Will , As the initial feature vector, the neighbor aggregation mechanism of node a is represented as follows: ; Among them, the final update of node a is expressed as, represents the activation function, G represents the parameter matrix of feature transformation, represents the influence coefficient of node a on the stability of node b, Represents the set of adjacent nodes of node a; According to the above formula, the vector representation of node a is generated Then the vector representation of the next connected node is performed, and so on. After all the nodes in the encapsulated directed graph are represented, the vector inner product of the nodes is calculated to obtain the overall estimated stability.

[0060] It should be further explained that, in the specific implementation process, the early warning process subsequence is generated according to the overall stability estimation operation result, and the process of performing temporary alternative allocation operation on the early warning process subsequence includes: The overall estimated stability is compared with the preset overall stability threshold. If the overall estimated stability is less than the overall stability threshold, the overall vulnerability coefficient of each process subsequence is obtained according to the stability level of each process subsequence and the stability influence coefficient of each process subsequence on other process subsequences, and the overall vulnerability coefficient threshold is preset. The overall vulnerability coefficient of each process subsequence is compared with the overall vulnerability coefficient threshold, and the process subsequence with an overall vulnerability coefficient greater than the overall vulnerability coefficient threshold is marked as a warning process subsequence. A temporary alternative allocation operation is performed on the warning process subsequence, and the operation of the preferred equipment in the warning process subsequence is suspended. The alternative equipment temporarily takes over the optimal production allocation plan of the day for the preferred equipment, and arranges relevant personnel to inspect and repair the preferred equipment.

[0061] It should be further explained that, in the specific implementation process, the specific calculation formula for obtaining the overall vulnerability coefficient of each process subsequence is:

[0062] in, represents the overall vulnerability coefficient of process subsequence j; represents the stability influence coefficient of process subsequence j; It represents the influence coefficient of process subsequence j on the stability of other process subsequences; Stability influence coefficients of several other process subsequences.

[0063] like Figure 2 As shown, an intelligent distribution system for a dairy product production line includes a monitoring center, wherein the monitoring center is communicatively connected to a day-ahead demand estimation module, a data processing module, a distribution plan planning module, an intraday monitoring module, a dynamic adjustment module, and a temporary alternative distribution module; The day-ahead demand prediction module is used to build a day-ahead demand prediction model and output the time series of the day's estimated order data; The data processing module is used to obtain the current dairy production process information, set the IoT points according to the process information, collect the actual production data of each point, build a production directed graph, and obtain the preferred equipment and alternative equipment for each process subsequence; The allocation plan planning module is used to improve the multi-objective genetic algorithm by introducing adaptive mutations that change with the number of iterations and dynamically adjusting weights. The optimal production allocation plan for the day is obtained through the improved multi-objective genetic algorithm based on the time series of the estimated order data for the day. The intraday monitoring module is used to set the judgment criteria for each process subsequence according to the optimal production allocation plan of the day, and perform intraday monitoring on each process subsequence according to the judgment criteria; The dynamic adjustment module is used to perform dynamic weight adjustment and reallocation operations based on the intraday monitoring results to obtain the new optimal production allocation plan for the day; The temporary alternative allocation module is used to perform overall stability estimation operations according to the intraday monitoring results, generate early warning process subsequences according to the overall stability estimation operation results, and perform temporary alternative allocation operations on the early warning process subsequences.

[0064] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An intelligent distribution method for a dairy product production line, characterized in that: The following steps are involved: Step s1: Build a day-ahead demand forecasting model and output the day-ahead estimated order data time series; Step s2: Obtain the current dairy product production process information, set IoT points according to the process information, collect actual production data at each point, build a production directed graph, and obtain the preferred equipment and alternative equipment for each process subsequence; Step s3: Introduce adaptive mutation that changes with the number of iterations and dynamically adjust weights to improve the multi-objective genetic algorithm, and obtain the best production allocation plan for the day based on the time series of the estimated order data of the day through the improved multi-objective genetic algorithm; Step s4: setting the judgment criteria for each process subsequence according to the optimal production allocation plan for the day, and performing intraday monitoring on each process subsequence according to the judgment criteria; Step s5: Perform dynamic weight adjustment and reallocation operations based on the intraday monitoring results to obtain a new optimal production allocation plan for the day; Step s6: Perform an overall stability estimation operation based on the intraday monitoring results, generate an early warning process subsequence, and perform a temporary alternative allocation operation on the early warning process subsequence.

2. The intelligent distribution method for a dairy product production line according to claim 1, characterized in that: The process of outputting the estimated order data time series for the day includes: Collect historical multi-source data, extract time features and perform STL decomposition on the historical multi-source data, obtain trend components and seasonal components, use the trend components and seasonal components as training data, use the training data to train the day-ahead demand prediction model, output the trained day-ahead demand prediction model, and output the time series sequence of the day's estimated order data according to the day-ahead demand prediction model.

3. The intelligent distribution method for a dairy product production line according to claim 2, characterized in that: The process of obtaining the current dairy production process information, setting IoT points according to the process information, collecting the actual production data of each point, and constructing a production directed graph includes: Obtain the process characteristics of the current dairy product production equipment, obtain process information based on the process characteristics, divide the dairy product production line according to the process information, divide it into several process sub-sequences, and deploy IoT nodes in each process sub-sequence. The actual production data includes the actual equipment status feature time series sequence and the standard equipment status feature time series sequence; The dairy production equipment in each process subsequence and the assembly order and assembly relationship between each dairy production equipment are extracted, each process subsequence is used as a node of the production directed graph, and the assembly order and assembly relationship between each dairy production equipment is used as the connection relationship between the nodes to construct a production directed graph.

4. The intelligent distribution method for a dairy product production line according to claim 3, characterized in that: The process of obtaining the preferred equipment and alternative equipment for each process subsequence includes: Extract the production equipment data of dairy production equipment in each process subsequence, use the production equipment data as evaluation indicators, set the indicator weights of the evaluation indicators, and obtain the membership matrix of each dairy production equipment to the preset reliability level through fuzzy comprehensive evaluation; The reliability level of each dairy product production equipment is obtained according to the membership matrix and the indicator weight, and the reliability level threshold is preset. The dairy product production equipment with a reliability level greater than or equal to the reliability level threshold is marked as the preferred equipment, and the dairy product production equipment with a reliability level less than the reliability level threshold is marked as the alternative equipment.

5. The intelligent distribution method for a dairy product production line according to claim 4, characterized in that: The process of obtaining the best production allocation plan for the day includes: Construct several objective functions, construct constraint formulas based on the time series of the estimated order data of the day and the production equipment data of the preferred equipment, preset the initial weight coefficients of several objective functions and several production allocation plans, construct fitness functions based on several objective functions and initial weight coefficients, perform chromosome encoding and population initialization on several production allocation plans, generate initial populations, introduce adaptive mutation that changes with the number of iterations and dynamically adjust weights to improve the multi-objective genetic algorithm; Based on the initial population, fitness function and constraint formula, the optimal production allocation plan for the day is obtained through an improved multi-objective genetic algorithm.

6. The intelligent distribution method for a dairy product production line according to claim 5, characterized in that: The judgment criteria for each process subsequence are set according to the optimal production allocation plan for the day. The process of intraday monitoring of each process subsequence according to the judgment criteria includes: The judgment criteria include standard order feature timing sequence, standard equipment status feature timing sequence, order feature deviation threshold interval and equipment status feature deviation threshold interval, and the actual order feature timing sequence and actual equipment status feature timing sequence of each process subsequence are obtained; Compare the actual order feature timing sequence of each process subsequence with the standard order feature timing sequence to obtain the order feature deviation, and compare the actual equipment state feature timing sequence of each process subsequence with the standard equipment state feature timing sequence to obtain the equipment state feature deviation; If there is a process subsequence whose order feature deviation is not within the order feature deviation threshold range or the equipment status feature deviation is not within the equipment status feature deviation threshold, dynamic weight adjustment and reallocation operations are performed; If the order feature deviations of each process subsequence are within the order feature deviation threshold and the equipment state feature deviations are within the equipment state feature deviation threshold, then threshold points are selected within the equipment state feature deviation threshold to divide sub-intervals of different stability levels to obtain the stability level of the process subsequence.

7. The intelligent distribution method for a dairy product production line according to claim 6, characterized in that: The process of dynamically adjusting and reallocating weights based on the intraday monitoring results to obtain the new optimal production allocation plan for the day includes: According to the actual order feature time series of the process subsequence and the actual equipment status feature time series, the demand urgency coefficient and equipment health coefficient of the process subsequence are obtained; Obtain the historical production data of each process subsequence, and obtain the sensitivity coefficients of various types of indicators in order characteristics and equipment status characteristics corresponding to several objective functions based on the historical production data; According to the sensitivity coefficients of various types of indicators corresponding to several objective functions, a weight adjustment rule is constructed, and a dynamic weight calculation model is constructed based on the weight adjustment rule. The demand urgency coefficient, equipment health coefficient, order feature deviation and equipment status feature deviation of the process subsequence are input into the dynamic weight calculation model, and the initial weight coefficients corresponding to several objective functions in the fitness function are updated according to the dynamic weight calculation model. Set a feedback cycle, input the order feature deviation and equipment status feature deviation of the process subsequence into the dynamic weight calculation model at the end timestamp of the feedback cycle, and update the weight adjustment rules of the dynamic weight calculation model; The constraint formula is updated according to the actual production data of each process subsequence. Based on the fitness function that completes the dynamic weight adjustment and the updated constraint formula, an improved multi-objective genetic algorithm is used to obtain a new optimal production allocation plan for the day.

8. The intelligent distribution method for a dairy product production line according to claim 7, characterized in that: The process of conducting an overall stability estimation operation based on the intraday monitoring results includes: Import the stability level of each process subsequence into the nodes of each process subsequence in the production directed graph, construct the adjacency matrix of the production directed graph, obtain the stability influence coefficient of each node in the production directed graph on other nodes according to the adjacency matrix, and set the weight label of each node according to the stability influence coefficient of each node on other nodes; A full-process stability prediction model is built based on a graph convolutional neural network, which learns and represents the production directed graph and outputs the overall estimated stability.

9. The intelligent distribution method for a dairy product production line according to claim 8, characterized in that: The process of generating a warning process subsequence according to the overall stability estimation operation result and performing a temporary alternative allocation operation on the warning process subsequence includes: An overall stability threshold is preset. If the overall estimated stability is less than the overall stability threshold, the overall vulnerability coefficient of each process subsequence is obtained according to the stability level of each process subsequence and the stability influence coefficient of each process subsequence on other process subsequences. An overall vulnerability coefficient threshold is preset, and the process subsequence whose overall vulnerability coefficient is greater than the overall vulnerability coefficient threshold is marked as a warning process subsequence, and a temporary alternative allocation operation is performed on the warning process subsequence.

10. An intelligent distribution system for a dairy product production line, specifically applied to an intelligent distribution method for a dairy product production line according to any one of claims 1 to 9, characterized in that: It includes a monitoring center, which is communicatively connected to a day-ahead demand estimation module, a data processing module, an allocation plan planning module, an intraday monitoring module, a dynamic adjustment module, and a temporary alternative allocation module; The day-ahead demand prediction module is used to build a day-ahead demand prediction model and output the time series of the day's estimated order data; The data processing module is used to obtain the current dairy production process information, set the IoT points according to the process information, collect the actual production data of each point, build a production directed graph, and obtain the preferred equipment and alternative equipment for each process subsequence; The allocation plan planning module is used to improve the multi-objective genetic algorithm by introducing adaptive mutations that change with the number of iterations and dynamically adjusting weights. The optimal production allocation plan for the day is obtained through the improved multi-objective genetic algorithm based on the time series of the estimated order data for the day. The intraday monitoring module is used to set the judgment criteria for each process subsequence according to the optimal production allocation plan of the day, and perform intraday monitoring on each process subsequence according to the judgment criteria; The dynamic adjustment module is used to perform dynamic weight adjustment and reallocation operations based on the intraday monitoring results to obtain the new optimal production allocation plan for the day; The temporary alternative allocation module is used to perform overall stability estimation operations according to the intraday monitoring results, generate early warning process subsequences according to the overall stability estimation operation results, and perform temporary alternative allocation operations on the early warning process subsequences.

Images