Dynamic capacity subcontracting optimization method and system based on machine learning

Abstract

The present invention relates to the field of machine learning technology, and in particular to a dynamic capacity subcontracting optimization method and system based on machine learning, including collecting production line data, subcontractor data and project data, and performing data preprocessing, thereby improving the accuracy and reliability of data, and facilitating the construction of a more accurate machine learning model; by constructing a machine learning model, dynamic capacity prediction, intelligent grouping of subcontractors and project optimization are realized; by formulating capacity adjustment strategies, subcontracting optimization strategies and resource allocation optimization strategies in a targeted manner, and performing dynamic adjustments, production efficiency and project quality are significantly improved; by establishing comprehensive feedback channels and quantitative analysis, it is conducive to continuously optimizing capacity and resource allocation, while improving the flexibility and adaptability of decision-making. The present invention is used to solve the technical problems of low production efficiency, unstable project quality and unreasonable resource allocation in current industrial production management.

Description

Technical Field

[0001] The present invention relates to the technical field of machine learning, and in particular to a dynamic capacity subcontracting optimization method and system based on machine learning. Background Art

[0002] Machine learning is a branch of artificial intelligence that aims to enable computers to acquire knowledge and experience from data through learning and automated reasoning of computer systems, and use this knowledge and experience for pattern recognition, prediction, and decision-making. It is a data-based method that uses computer algorithms to automatically analyze data, learn patterns, and make predictions or decisions.

[0003] In the current industrial production field, there are multiple challenges: low production efficiency. The current production line may be inefficient due to inaccurate data, inaccurate capacity forecasting or unreasonable resource allocation. Project quality problems caused by poor management of subcontractors. The selection and management of subcontractors are crucial to project quality. Under the traditional subcontract management model, it may be difficult to accurately assess the qualifications and capabilities of subcontractors, resulting in unstable project quality and difficulty in obtaining effective guarantees. Resource waste caused by unreasonable resource allocation. Irrational resource allocation may lead to idle or overused resources, resulting in resource waste. Decision-making lacks flexibility and adaptability. Traditional production management methods often rely on experience or fixed rules, lack of decision-making flexibility and adaptability, and are difficult to cope with rapidly changing market demands and production line conditions. To this end, the present invention proposes a dynamic capacity subcontracting optimization method and system based on machine learning. Summary of the invention

[0004] The purpose of the present invention is to solve the problems in the background technology and to propose a dynamic capacity subcontracting optimization method and system based on machine learning.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] The dynamic capacity subcontracting optimization method based on machine learning includes:

[0007] Step 1: Determine the sources of production line data, subcontractor data and project data, collect production line data, subcontractor data and project data, and pre-process the collected production line data, subcontractor data and project data; data sources include sensors, control systems, project management systems, subcontractor management systems, etc.;

[0008] Step 2: Based on the pre-processed production line data, subcontractor data and project data, machine learning models suitable for dynamic capacity forecasting, subcontractor grouping evaluation and project optimization are constructed, namely, a dynamic capacity forecasting model, a subcontractor grouping evaluation model and a project optimization model;

[0009] The construction process of the dynamic capacity forecasting model is as follows: select a time series analysis algorithm, analyze the preprocessed production line data, identify the law of capacity changes over time, and on the basis of identifying the law of capacity changes over time, use the supervised learning algorithm to further analyze the relationship between capacity and market demand, so as to predict future capacity trends;

[0010] The construction process of the subcontractor grouping evaluation model is as follows: using clustering algorithms to group subcontractors and obtain subcontractor grouping evaluation results to facilitate subsequent intelligent matching;

[0011] The construction process of the project optimization model is as follows: considering multiple goals such as project cost, project quality, and project duration, and setting corresponding proportional coefficients for each goal, while defining the constraints of the project optimization model to obtain the output results of the project optimization model;

[0012] Step 3: According to the output results of each model, formulate intelligent strategies in a targeted manner and make dynamic adjustments according to the actual situation; the formulation of intelligent strategies includes capacity adjustment strategy, subcontracting optimization strategy and resource allocation optimization strategy;

[0013] Step 4: Establish feedback channels, collect feedback from production lines, subcontractors, and project management, use databases to record feedback data, calculate and analyze quantitative indicator values ​​in feedback data, and generate feedback signals.

[0014] Furthermore, the process of determining the sources of the production line data, subcontractor data and project data, collecting the production line data, subcontractor data and project data, and preprocessing the collected production line data, subcontractor data and project data includes:

[0015] Collect real-time production line data from production line sensors and control systems; production line data includes equipment status, production speed, and process parameters;

[0016] Mark the collected equipment status, production speed and process parameters respectively to obtain the equipment status , production speed And process parameters ;

[0017] The collected production line data is organized into a production line data matrix ,Right now

[0018] ,

[0019] In the formula, is the number of production line data points collected;

[0020] Preprocessing the collected production line data; the preprocessing specifically includes removing outliers;

[0021] Collect subcontractor data from the project management system and subcontractor management system; subcontractor data includes historical performance, qualifications, reputation, and project completion status;

[0022] Mark historical performance, qualifications, reputation and project completion status respectively to get historical performance Qualifications , Reputation and project completion ;

[0023] The collected subcontractor data is organized into a subcontractor data matrix ,Right now

[0024] ,

[0025] In the formula, The number of subcontractor data points collected;

[0026] Use data standardization methods to pre-process subcontractor data;

[0027] Collect project data from the project management system; the project data includes project cost, project quality and project duration;

[0028] Set labels for the collected project cost, project quality and project duration respectively: Project cost , Project Quality and project duration , and form the project data vector ;

[0029] The project data are preprocessed using the weighted moving average method.

[0030] Furthermore, a time series analysis algorithm is selected to analyze the preprocessed production line data to identify the law of capacity changes over time. On the basis of identifying the law of capacity changes over time, a supervised learning algorithm is used to further analyze the relationship between capacity and market demand. The process of predicting future capacity trends includes:

[0031] Get pre-processed production line data , and integrate to form time series data reflecting the production line capacity ;in, Indicates time;

[0032] Based on time series data reflecting production line capacity , constructed using time series analysis algorithms , identify the law of production capacity changes over time; among them, The formula is:

[0033] ,

[0034] In the formula, is the backshift operator, is an autoregressive polynomial, is the moving average polynomial, represents the self-noise sequence, represents the autoregressive order, represents the difference order, represents the moving average order;

[0035] Preset production line data Key factors related to market demand There is a linear relationship. The supervised learning algorithm is used to establish a correlation model between production capacity and market demand, namely the linear regression model:

[0036] ,

[0037] Among them, the key factors of market demand are is the independent variable, is the number of key factors of market demand, represents the regression coefficient, represents the error term;

[0038] According to the analysis results of production capacity and market demand, predict future production capacity trends and obtain dynamic production capacity forecast results .

[0039] Furthermore, the subcontractors are grouped using a clustering algorithm, and the process of obtaining the subcontractor grouping evaluation results includes:

[0040] Get the preprocessed subcontractor data and extract the corresponding subcontractor data points ;in, represents the index of the extracted subcontractor data point, and , Indicates the number of subcontractor data points extracted;

[0041] Divide the subcontractor data points into Clusters, randomly initialized Cluster Centers ;in, Represents the cluster center index;

[0042] Calculate the distance from each subcontractor data point to the cluster center:

[0043] ,

[0044] In the formula, Represents the subcontractor data feature index, Represents the characteristic dimension of subcontractor data;

[0045] Assign the subcontractor data points to the cluster to which the nearest cluster center belongs and update the cluster center:

[0046] ,

[0047] In the formula, Indicates The set of subcontractor data points in the cluster;

[0048] Repeat the clustering algorithm steps until the cluster center no longer changes, and obtain the subcontractor grouping evaluation results.

[0049] Furthermore, considering multiple goals such as project cost, project quality, and project duration, and setting corresponding proportional coefficients for each goal, while defining the constraints of the project optimization model, the process of obtaining the output results of the project optimization model includes:

[0050] Get preprocessed project data and extract the corresponding project cost , Project Quality and project duration ;

[0051] The extracted project cost, project quality and project duration are integrated into a multi-objective function, namely

[0052] ,

[0053] In the formula, represents the output value of multiple objective functions, represents the proportionality factor of the project cost, A proportionality factor representing the quality of the project, Represents the proportionality coefficient of the project duration, satisfying ;

[0054] Define the constraints of the project optimization model; the constraints specifically include capacity constraints and resource constraints.

[0055] Furthermore, according to the output results of each model, intelligent strategies are formulated in a targeted manner, and the process of dynamic adjustment according to actual conditions includes:

[0056] Step H1: Obtain current market demand trends ;

[0057] Compare and analyze the prediction results of the dynamic capacity prediction model with the current market demand trend. When When the production capacity is reduced, the early warning signal of capacity reduction is immediately triggered;

[0058] Step H2: According to the subcontractor grouping assessment results, match the subcontractor that best suits the current subcontract project and capacity requirements, and establish a subcontractor monitoring mechanism;

[0059] Step H3: Combine the output results of the project optimization model and define the resource allocation vector as ;in, Represents the dimension of resource allocation vector;

[0060] Computing resource utilization , where represents the index of the resource allocation vector, Indicates the maximum supply of resources;

[0061] If the resource utilization is lower than or higher than a preset resource threshold, a resource adjustment signal is triggered.

[0062] Furthermore, the process of collecting feedback from production lines, subcontractors, and project management and recording feedback data using a database includes:

[0063] Step S1: Determine feedback categories and quantitative indicators based on feedback from production lines, subcontractors, and project management;

[0064] Step S2: design a feedback collection form according to the feedback category and quantitative index; wherein the feedback collection form contains the following key information: feedback category, quantitative index and feedback time;

[0065] Step S3: Create a feedback database for storing all collected feedback data; wherein the database includes a feedback record table, which includes key fields such as feedback category, quantitative index value, and corresponding timestamp.

[0066] Dynamic capacity subcontracting optimization system based on machine learning, including:

[0067] The data collection and preprocessing module is used to determine the sources of production line data, subcontractor data and project data, collect production line data, subcontractor data and project data, and preprocess the collected production line data, subcontractor data and project data;

[0068] The comprehensive multi-model building module is used to build machine learning models suitable for dynamic capacity forecasting, subcontractor grouping evaluation and project optimization based on pre-processed production line data, subcontractor data and project data, namely, dynamic capacity forecasting model, subcontractor grouping evaluation model and project optimization model;

[0069] The intelligent strategy formulation and adjustment module is used to formulate intelligent strategies based on the output results of each model and make dynamic adjustments based on actual conditions;

[0070] The feedback collection and quantitative analysis module is used to establish feedback channels, collect feedback from production lines, subcontractors, and project management, use the database to record feedback data, calculate and analyze the quantitative indicator values ​​in the feedback data, and generate feedback signals.

[0071] Compared with the existing technology, the advantages of the dynamic capacity subcontracting optimization method and system based on machine learning provided by the present invention are:

[0072] 1. The present invention determines the sources of production line data, subcontractor data and project data, collects production line data, subcontractor data and project data, and pre-processes the collected production line data, subcontractor data and project data, thereby ensuring the comprehensiveness and diversity of the data, facilitating the analysis and processing of subsequent algorithms, and improving data processing efficiency;

[0073] 2. The present invention constructs a dynamic capacity prediction model through time series analysis algorithm and supervised learning algorithm, which can accurately predict future capacity trends and provide a scientific basis for capacity adjustment. It uses clustering algorithm to group and evaluate subcontractors, realizes intelligent matching of subcontractors, improves subcontracting efficiency and project quality, and comprehensively considers multiple goals such as project cost, quality, and construction period to construct a project optimization model. By setting proportional coefficients and constraints, the optimal project implementation plan is obtained, and the project benefits are maximized. According to the output results of each model, the capacity adjustment strategy, subcontracting optimization strategy and resource allocation optimization strategy are formulated in a targeted manner, which improves the pertinence and effectiveness of decision-making;

[0074] 3. The present invention establishes feedback channels, collects feedback from production lines, subcontractors, and project management, uses a database to record feedback data, calculates and analyzes quantitative indicator values ​​in the feedback data, and generates feedback signals, thereby providing an objective and accurate basis for decision-making and facilitating subsequent implementation and optimization.

[0075] In summary, the present invention performs dynamic capacity subcontracting optimization through machine learning, which provides strong support for achieving efficient, intelligent and sustainable production management, thereby ensuring the normal operation of subsequent dynamic capacity subcontracting optimization methods and systems based on machine learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 This is a flow chart of the dynamic capacity subcontracting optimization method based on machine learning proposed in the present invention.

[0077] Figure 2This is a module diagram of the dynamic capacity subcontracting optimization system based on machine learning proposed in the present invention. DETAILED DESCRIPTION

[0078] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the implementation regulations described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0079] See also Figure 1 The present invention provides a dynamic capacity subcontracting optimization method based on machine learning, comprising:

[0080] Step 1: Determine the sources of production line data, subcontractor data and project data, collect production line data, subcontractor data and project data, and pre-process the collected production line data, subcontractor data and project data; data sources include sensors, control systems, project management systems, subcontractor management systems, etc.;

[0081] Step 2: Based on the pre-processed production line data, subcontractor data and project data, machine learning models suitable for dynamic capacity forecasting, subcontractor grouping evaluation and project optimization are constructed, namely, a dynamic capacity forecasting model, a subcontractor grouping evaluation model and a project optimization model;

[0082] The construction process of the dynamic capacity forecasting model is as follows: select a time series analysis algorithm, analyze the preprocessed production line data, identify the law of capacity changes over time, and on the basis of identifying the law of capacity changes over time, use the supervised learning algorithm to further analyze the relationship between capacity and market demand, so as to predict future capacity trends;

[0083] The construction process of the subcontractor grouping evaluation model is as follows: using clustering algorithms to group subcontractors and obtain subcontractor grouping evaluation results to facilitate subsequent intelligent matching;

[0084] The construction process of the project optimization model is as follows: considering multiple goals such as project cost, project quality, and project duration, and setting corresponding proportional coefficients for each goal, while defining the constraints of the project optimization model to obtain the output results of the project optimization model;

[0085] Step 3: According to the output results of each model, formulate intelligent strategies in a targeted manner and make dynamic adjustments according to the actual situation; the formulation of intelligent strategies includes capacity adjustment strategy, subcontracting optimization strategy and resource allocation optimization strategy;

[0086] Step 4: Establish feedback channels, collect feedback from production lines, subcontractors, and project management, use databases to record feedback data, calculate and analyze quantitative indicator values ​​in feedback data, and generate feedback signals.

[0087] See also Figure 1 The present invention provides a dynamic capacity subcontracting optimization method based on machine learning. In the step 1, the sources of production line data, subcontractor data and project data are determined, the production line data, subcontractor data and project data are collected, and the steps of pre-processing the collected production line data, subcontractor data and project data include:

[0088] Step 101: Collect real-time production line data from sensors and control systems of the production line; wherein the production line data includes equipment status, production speed, and process parameters;

[0089] Step 102: Mark the collected equipment status, production speed and process parameters respectively to obtain the equipment status , production speed And process parameters ;

[0090] Step 103: The collected production line data is combined into a production line data matrix. ,Right now

[0091] ,

[0092] In the formula, is the number of production line data points collected;

[0093] Step 104: pre-process the collected production line data; the pre-processing specifically includes removing outliers. satisfy , then the production line data point is an outlier and is removed, where represents the mean of the production line data points, Represents the production line data point index, represents the standard deviation of the production line data points, Indicates the set outlier detection threshold, such as ;

[0094] Step 105: Collect subcontractor data from the project management system and the subcontractor management system; wherein the subcontractor data includes historical performance, qualifications, reputation, and project completion status;

[0095] Step 106: Mark the historical performance, qualifications, reputation and project completion status respectively to obtain the historical performance Qualifications , Reputation and project completion ;

[0096] Step 107: The collected subcontractor data is formed into a subcontractor data matrix ,Right now

[0097] ,

[0098] In the formula, The number of subcontractor data points collected;

[0099] Step 108: Pre-process the subcontractor data using a data standardization method:

[0100] ,

[0101] In the formula, is the original subcontractor data, and , is the standardized subcontractor data, represents a collection of subcontractor data;

[0102] Step 109: Collect project data from the project management system; wherein the project data includes project cost, project quality and project duration;

[0103] Step 110, set labels for the collected project cost, project quality and project duration respectively: project cost , Project Quality and project duration , and form the project data vector ;

[0104] Step 111: Use the weighted moving average method to pre-process the project data: If there are data points with large cost fluctuations, use the weighted moving average method to smooth them. For the The project cost after weighted moving average is:

[0105] ,

[0106] In the formula, An index representing the cost of an item, Indicates the index of different durations used in the weighted moving average calculation. Indicates the period of the moving average (such as ), Represents the weight factor corresponding to the project cost, satisfying .

[0107] See also Figure 1 The present invention provides a dynamic capacity subcontracting optimization method based on machine learning. In the step 2, the steps of constructing a dynamic capacity prediction model, a subcontractor grouping evaluation model and a project optimization model include:

[0108] Step 201: Obtain pre-processed production line data , and integrate to form time series data reflecting the production line capacity ;in, Indicates time;

[0109] Step 202: Based on the time series data reflecting the production line capacity , constructed using time series analysis algorithms , identify the law of production capacity changes over time; among them, The formula is:

[0110] ,

[0111] In the formula, is the backshift operator ( ), is an autoregressive polynomial, is the moving average polynomial, represents the self-noise sequence, represents the autoregressive order, represents the difference order, represents the moving average order;

[0112] In step 202, the model parameters can be determined by observing the autocorrelation function and partial autocorrelation function graphs, or by using the information criterion, such as -Akaike Information Guidelines: , where represents the number of model parameters, represents the likelihood function value;

[0113] Step 203: Preset production line data Key factors related to market demand There is a linear relationship. The supervised learning algorithm is used to establish a correlation model between production capacity and market demand, namely the linear regression model:

[0114] ,

[0115] Among them, the key factors of market demand are As an independent variable, it can represent market size, market growth rate, etc. is the number of key factors of market demand, represents the regression coefficient, represents the error term, where the regression coefficient It can be solved by the least squares method, the specific process is:

[0116] Production line data As the response variable vector, for the independent variable Constructing the data matrix , through the formula Solving for regression coefficients , where is a OK The data matrix of columns (the first column is a vector of all 1s, representing the intercept term), Indicates production line data the number of

[0117] Step 204: Based on the analysis results of production capacity and market demand, predict the future production capacity trend and obtain dynamic production capacity prediction results. ;

[0118] Step 205: Obtain the pre-processed subcontractor data and extract the corresponding subcontractor data points ;in, represents the index of the extracted subcontractor data point, and , Indicates the number of subcontractor data points extracted;

[0119] Step 206: Divide the subcontractor data points into Clusters, randomly initialized Cluster Centers ;in, Represents the cluster center index;

[0120] Step 207: Calculate the distance from each subcontractor data point to the cluster center:

[0121] ,

[0122] In the formula, Represents the subcontractor data feature index, Represents the characteristic dimension of subcontractor data;

[0123] Step 208: Assign the subcontractor data point to the cluster to which the nearest cluster center belongs, and update the cluster center:

[0124] ,

[0125] In the formula, Indicates The set of subcontractor data points in the cluster;

[0126] Step 209, repeat the clustering algorithm steps until the cluster center no longer changes, and obtain the subcontractor grouping evaluation result;

[0127] Step 210: Obtain pre-processed project data and extract corresponding project costs , Project Quality and project duration ;

[0128] Step 211: Integrate the extracted project cost, project quality and project duration into a multi-objective function, namely

[0129] ,

[0130] In the formula, represents the output value of multiple objective functions, represents the proportionality factor of the project cost, A proportionality factor representing the quality of the project, Represents the proportionality coefficient of the project duration, satisfying For example, if cost is the most important objective, different scaling factors can be set as scaling factors for project costs. , the proportionality coefficient of project quality , the proportional coefficient of project duration ;

[0131] Step 212: define constraints of the project optimization model; wherein the constraints specifically include capacity constraints and resource constraints:

[0132] As for capacity constraints, the capacity of the production line is affected by many factors, such as the number of equipment , the average operating efficiency of the equipment , Equipment operational time ,Product production cycle and the loss rate during the production process , set the theoretical maximum capacity , and the theoretical minimum capacity May be subject to some basic production maintenance requirements, such as the minimum number of equipment startups , minimum average operating efficiency , shortest run time And consider a certain safety margin ( ) and minimum loss rate during production Influenced by factors such as ;

[0133] Resource constraints include raw material resources, labor resources, and equipment resources;

[0134] In the Raw Materials resource, set the total amount of raw materials to The amount of raw materials required per unit product is , then the raw material resource limit is:

[0135] ,

[0136] In the formula, Indicates the product category index. Indicates the number of product types. Indicates The production quantity of the product, Indicates the minimum amount of raw materials required to maintain production (taking into account factors such as inventory safety). Indicates the maximum available amount of raw materials (subject to purchase, inventory, etc. restrictions);

[0137] In the labor resource, set the total number of workers to , the labor hours required to produce a unit of product is , the work efficiency of the labor force is , then the labor resource limit is:

[0138] ,

[0139] In the formula, Indicates the minimum number of workers required to maintain production operations (considering some basic positions, etc.), Indicates the maximum amount of labor available (subject to limitations on the number of employees, working hours, etc.);

[0140] In the device resources, set the total number of devices to , the time that a unit of production product occupies the equipment is The average availability of equipment is , the device resource limit is:

[0141] ,

[0142] In the formula, Indicates the minimum equipment resources required to maintain production (considering key equipment, etc.), Indicates the maximum device resources that can be provided (subject to limitations such as device quantity and device maintenance).

[0143] See also Figure 1 The present invention provides a dynamic capacity subcontracting optimization method based on machine learning. In step three, according to the output results of each model, an intelligent strategy is formulated in a targeted manner, and the steps of dynamically adjusting according to the actual situation include:

[0144] Step 301: Obtain current market demand trends ;

[0145] Compare and analyze the prediction results of the dynamic capacity prediction model with the current market demand trend. When the production capacity is increased, the early warning signal of capacity increase is immediately triggered, indicating that the production scale needs to be expanded to meet the current market demand. When the production capacity is reduced, the early warning signal of capacity reduction is immediately triggered, and it is recommended to adjust the production plan to avoid overcapacity and waste of resources;

[0146] For example, define the capacity adjustment as ,like ,but , where is the adjustment factor, and the adjustment factor Greater than 0 and less than 1;

[0147] Step 302: According to the subcontractor grouping evaluation results, match the subcontractor that best suits the current subcontract project and capacity requirements, and establish a subcontractor monitoring mechanism: according to the current subcontractor's project completion status, calculate the subcontractor's work completion rate (work completion rate = completed workload / total workload × 100%), and compare the calculated work completion rate with the preset target completion rate range: if the subcontractor's work completion rate falls within the preset target completion rate range, it means that the subcontractor's work progress is in line with expectations, and further subcontract cost control will be carried out to ensure the economic benefits of the project; if it is not within the preset target completion rate range, it means that the subcontractor's work may be delayed, and it is necessary to adjust the subcontracting strategy in a timely manner, including re-evaluating the subcontractor's qualifications, adjusting the work plan, or finding alternative subcontractors, etc., to ensure that the project can be completed on time and with quality;

[0148] Step 303: Based on the output results of the project optimization model, define the resource allocation vector as ;in, Represents the dimension of resource allocation vector;

[0149] Computing resource utilization , where represents the index of the resource allocation vector, Indicates the maximum supply of resources;

[0150] If the resource utilization is lower than or higher than the preset resource threshold, a resource adjustment signal is triggered; for example, the preset resource threshold is marked as , the expected interval of resource utilization is ;when or , a resource adjustment signal is triggered.

[0151] See also Figure 1 The present invention provides a dynamic capacity subcontracting optimization method based on machine learning. In step 4, a feedback channel is established to collect feedback from production lines, subcontractors, and project management, and a database is used to record feedback data, and quantitative index values ​​in the feedback data are calculated and analyzed, and the steps of generating feedback signals include:

[0152] Step 401: Determine feedback categories and quantitative indicators based on feedback from production lines, subcontractors, and project management;

[0153] In step 401, the feedback categories correspond to the feedback opinions of the production line, subcontractor, and project management, respectively, wherein the quantitative indicators may be production line efficiency, subcontractor quality, and project progress. Specifically, the output volume, product quality, etc. are used as the standards for measuring production line efficiency, the subcontractor quality is evaluated by the project completion rate and customer satisfaction, and the project progress is quantified by the number of delay days. It is understandable that a special feedback channel (such as a production line employee suggestion box, an online feedback platform, etc.) can be set up to collect feedback on production line efficiency, and feedback on subcontractor quality can be collected through questionnaires, project review meetings, etc., and the project progress can be obtained based on real-time recording of project progress.

[0154] Step 402: design a feedback collection form based on the feedback category and quantitative indicators to ensure that the collected information is comprehensive and accurate; wherein the feedback collection form includes the following key information: feedback category, quantitative indicators and feedback time;

[0155] Step 403: Create a feedback database for storing all collected feedback data; wherein the database includes a feedback record table, which includes key fields such as feedback category, quantitative index value, and corresponding timestamp;

[0156] Step 404, taking production line efficiency as an example, the specific steps of calculating and analyzing the quantitative index value in the feedback data and generating a feedback signal include:

[0157] Step W1: Based on product quality, production cost and production time, use weighted efficiency index For quantitative analysis, the calculation formula is:

[0158] ,

[0159] In the formula, The weight coefficients corresponding to product quality, production cost and production time respectively ( ), Indicates the number of qualified products, represents the total production quantity, represents the actual production cost, represents the budgeted production cost, represents the theoretical production time, Indicates the actual production time;

[0160] Step W2: Use time series analysis to predict future production efficiency. Product quality, production cost, and production time are multiple dimensions as input feature matrices and are marked as ;definition is the corresponding efficiency value vector, through the regression model Predict new efficiency values ;

[0161] Step W3: Perform abnormal diagnosis on the predicted efficiency value and set the standard error of the prediction to , calculated using the formula Confidence Interval:

[0162] ,

[0163] In the formula, is a standard normal distribution Quantiles are used to determine the width of confidence intervals (e.g., for a 95% confidence interval, ,therefore ); Indicates the significance level, which is used to determine the confidence level of the confidence interval. A higher value means a higher confidence level, but may also result in a wider confidence interval; They represent the lower and upper limits of the confidence interval respectively;

[0164] Step W4: Based on the results of the confidence interval, different feedback signals are generated: if the new efficiency value is significantly lower than the lower limit of the confidence interval, a strong negative feedback signal is generated; if the new efficiency value is within the confidence interval but lower, a mild negative feedback signal is generated; if the new efficiency value is significantly higher than the upper limit of the confidence interval, a strong positive feedback signal is generated;

[0165] Step W5: Dynamically adjust the parameters and strategies of the production line according to the feedback signal. For example, if a strong negative feedback signal is received, increase resource input, optimize the production process or adjust the production plan. At the same time, use optimization algorithms (such as genetic algorithms, particle swarm optimization, etc.) to find the optimal combination of production parameters.

[0166] In the embodiment of the present invention, by collecting data from multiple sources such as sensors, control systems, project management systems, and subcontractor management systems, the comprehensiveness and diversity of the data are ensured, providing a solid foundation for subsequent analysis and modeling. By preprocessing the collected data, removing outliers, standardizing data, etc., the accuracy and reliability of the data are improved, and the influence of data errors on subsequent analysis is avoided. Through the time series analysis algorithm and the supervised learning algorithm, the future capacity trend can be accurately predicted, providing a scientific basis for enterprises to formulate production plans and adjust capacity. By using the clustering algorithm to group and evaluate subcontractors, it is convenient to intelligently match the subcontractors that are most suitable for the current subcontracting projects and capacity requirements, thereby improving the subcontracting efficiency and project quality. By constructing a multi-objective optimization model based on multiple objectives such as project cost, project quality, and project duration, the comprehensive optimization of the project is achieved, and the overall benefits of the enterprise are improved. By formulating the capacity adjustment strategy, subcontracting optimization strategy, and resource allocation optimization strategy in a targeted manner according to the output results of each model, the intelligent and automated decision-making is achieved. Through feedback and quantitative analysis, possible problems can be discovered and solved in a timely manner, providing strong support for enterprises to achieve intelligent decision-making, improve production efficiency, reduce costs, and enhance competitiveness. In summary, the examples of the present invention involve decision-making on data collection and analysis, result generation and performance evaluation, and solve the technical problems of low production efficiency, unstable project quality and unreasonable resource allocation in current industrial production management. In actual situations, more data and contextual information may be needed to make specific decisions and optimization plans.

[0167] See also Figure 2 The present invention provides a dynamic capacity subcontracting optimization system based on machine learning, comprising:

[0168] The data collection and preprocessing module is used to determine the sources of production line data, subcontractor data and project data, collect production line data, subcontractor data and project data, and preprocess the collected production line data, subcontractor data and project data;

[0169] The comprehensive multi-model building module is used to build machine learning models suitable for dynamic capacity forecasting, subcontractor grouping evaluation and project optimization based on pre-processed production line data, subcontractor data and project data, namely, dynamic capacity forecasting model, subcontractor grouping evaluation model and project optimization model;

[0170] The intelligent strategy formulation and adjustment module is used to formulate intelligent strategies based on the output results of each model and make dynamic adjustments based on actual conditions;

[0171] The feedback collection and quantitative analysis module is used to establish feedback channels, collect feedback from production lines, subcontractors, and project management, use the database to record feedback data, calculate and analyze the quantitative indicator values ​​in the feedback data, and generate feedback signals.

[0172] In addition, the formulas involved in the above are all calculated by removing dimensions and taking their numerical values. They are a formula that is closest to the actual situation obtained by collecting a large amount of data and performing software simulation. The proportional coefficient in the formula and the various preset thresholds in the analysis process are set by technical personnel in this field according to actual conditions or obtained by simulating a large amount of data; the size of the proportional coefficient is to quantify each parameter to obtain a specific value for subsequent comparison. The size of the proportional coefficient depends on the amount of sample data and the preliminary setting of the corresponding processing coefficient for each group of sample data by technical personnel in this field; as long as it does not affect the proportional relationship between the parameter and the quantized value.

[0173] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically based on the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0174] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing the present application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0175] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0176] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0177] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0178] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0179] Secondly: In the drawings of the embodiments disclosed in the present invention, only the structures related to the embodiments disclosed in the present invention are involved, and other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of the present invention can be combined with each other;

[0180] Finally: The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A dynamic capacity subcontracting optimization method based on machine learning, characterized by: Step 1: Determine the sources of production line data, subcontractor data and project data, collect production line data, subcontractor data and project data, and pre-process the collected production line data, subcontractor data and project data; Step 2: Based on the pre-processed production line data, subcontractor data and project data, machine learning models suitable for dynamic capacity forecasting, subcontractor grouping evaluation and project optimization are constructed, namely, a dynamic capacity forecasting model, a subcontractor grouping evaluation model and a project optimization model; The construction process of the dynamic capacity forecasting model is as follows: select a time series analysis algorithm, analyze the preprocessed production line data, identify the law of capacity changes over time, and on the basis of identifying the law of capacity changes over time, use the supervised learning algorithm to further analyze the relationship between capacity and market demand, so as to predict future capacity trends; The construction process of the subcontractor grouping evaluation model is as follows: using clustering algorithm to group subcontractors and obtain subcontractor grouping evaluation results; The construction process of the project optimization model is as follows: considering multiple goals such as project cost, project quality, and project duration, and setting corresponding proportional coefficients for each goal, while defining the constraints of the project optimization model to obtain the output results of the project optimization model; Step 3: According to the output results of each model, formulate intelligent strategies in a targeted manner and make dynamic adjustments according to the actual situation; the formulation of intelligent strategies includes capacity adjustment strategy, subcontracting optimization strategy and resource allocation optimization strategy; Step 4: Establish feedback channels, collect feedback from production lines, subcontractors, and project management, use databases to record feedback data, calculate and analyze the quantitative indicator values ​​in the feedback data, and generate feedback signals; The process of step three, formulating intelligent strategies in a targeted manner according to the output results of each model and making dynamic adjustments according to the actual situation, includes: H1. Obtain current market demand trends ; Compare and analyze the prediction results of the dynamic capacity prediction model with the current market demand trend. When When the production capacity is reduced, the early warning signal of capacity reduction is immediately triggered; H2. According to the results of the subcontractor grouping assessment, match the subcontractor that best suits the current subcontracting project and capacity requirements, and establish a subcontractor monitoring mechanism; H3. Combined with the output results of the project optimization model, the resource allocation vector is defined as ;in, Represents the dimension of resource allocation vector; Computing resource utilization , where represents the index of the resource allocation vector, Indicates the maximum supply of resources; If the resource utilization is lower than or higher than the preset resource threshold, a resource adjustment signal is triggered; The process of collecting feedback from production lines, subcontractors, and project management and using a database to record feedback data includes: S1. Determine feedback categories and quantitative indicators based on feedback from production lines, subcontractors, and project management; S2. Design a feedback collection form based on feedback categories and quantitative indicators; S3. Create a feedback database to store all collected feedback data.

2. The dynamic capacity subcontracting optimization method based on machine learning according to claim 1 is characterized in that: In the step 1, the process of determining the sources of the production line data, subcontractor data and project data, collecting the production line data, subcontractor data and project data, and preprocessing the collected production line data, subcontractor data and project data includes: Collect real-time production line data from production line sensors and control systems; production line data includes equipment status, production speed, and process parameters; Mark the collected equipment status, production speed and process parameters respectively to obtain the equipment status , production speed And process parameters ; The collected production line data is organized into a production line data matrix ,Right now ， In the formula, is the number of production line data points collected; Preprocessing the collected production line data; the preprocessing specifically includes removing outliers; Collect subcontractor data from the project management system and subcontractor management system; subcontractor data includes historical performance, qualifications, reputation, and project completion status; Mark historical performance, qualifications, reputation and project completion status respectively to get historical performance Qualifications , Reputation and project completion ; The collected subcontractor data is organized into a subcontractor data matrix ,Right now ， In the formula, The number of subcontractor data points collected; Use data standardization methods to pre-process subcontractor data; Collect project data from the project management system; the project data includes project cost, project quality and project duration; Set labels for the collected project cost, project quality and project duration respectively: Project cost , Project Quality and project duration , and form the project data vector ; The project data are preprocessed using the weighted moving average method.

3. The dynamic capacity subcontracting optimization method based on machine learning according to claim 1 is characterized in that: In the step 2, a time series analysis algorithm is selected to analyze the preprocessed production line data to identify the law of capacity changes over time. On the basis of identifying the law of capacity changes over time, a supervised learning algorithm is used to further analyze the relationship between capacity and market demand. The process of predicting future capacity trends includes: Get pre-processed production line data , and integrate to form time series data reflecting the production line capacity ;in, Indicates time; Based on time series data reflecting production line capacity , constructed using time series analysis algorithms , identify the law of production capacity changes over time; among them, The formula is: ， In the formula, is the backshift operator, is an autoregressive polynomial, is the moving average polynomial, represents the self-noise sequence, represents the autoregressive order, represents the difference order, represents the moving average order; Preset production line data Key factors related to market demand There is a linear relationship. The supervised learning algorithm is used to establish a correlation model between production capacity and market demand, namely the linear regression model: ， Among them, the key factors of market demand are is the independent variable, is the number of key factors of market demand, represents the regression coefficient, represents the error term; According to the analysis results of production capacity and market demand, predict future production capacity trends and obtain dynamic production capacity forecast results .

4. The dynamic capacity subcontracting optimization method based on machine learning according to claim 1 is characterized in that: In the step 2, the process of grouping the subcontractors using a clustering algorithm to obtain the subcontractor grouping evaluation results includes: Get the preprocessed subcontractor data and extract the corresponding subcontractor data points ;in, represents the index of the extracted subcontractor data point, and , Indicates the number of subcontractor data points extracted; Divide the subcontractor data points into Clusters, randomly initialized Cluster Centers ;in, Represents the cluster center index; Calculate the distance from each subcontractor data point to the cluster center: ， In the formula, Represents the subcontractor data feature index, Represents the characteristic dimension of subcontractor data; Assign the subcontractor data points to the cluster to which the nearest cluster center belongs and update the cluster center: ， In the formula, Indicates The set of subcontractor data points in the cluster; Repeat the clustering algorithm steps until the cluster center no longer changes, and obtain the subcontractor grouping evaluation results.

5. The dynamic capacity subcontracting optimization method based on machine learning according to claim 1 is characterized in that: In the step 2, multiple goals including project cost, project quality and project duration are considered, and a corresponding proportional coefficient is set for each goal. At the same time, the constraint conditions of the project optimization model are defined. The process of obtaining the output result of the project optimization model includes: Get preprocessed project data and extract the corresponding project cost , Project Quality and project duration ; The extracted project cost, project quality and project duration are integrated into a multi-objective function, namely ， In the formula, represents the output value of multiple objective functions, represents the proportionality factor of the project cost, A proportionality factor representing the quality of the project, Represents the proportionality coefficient of the project duration, satisfying ; Define the constraints of the project optimization model; the constraints specifically include capacity constraints and resource constraints.

6. The dynamic capacity subcontracting optimization method based on machine learning according to claim 1 is characterized in that: The designed feedback collection table contains the following key information: feedback category, quantitative index and feedback time; the created feedback database contains a feedback record table, and the table contains key fields such as feedback category, quantitative index value and corresponding timestamp.

7. Dynamic capacity subcontracting optimization system based on machine learning, characterized by: The method for optimizing dynamic capacity subcontracting based on machine learning as claimed in any one of claims 1 to 6 comprises: The data collection and preprocessing module is used to determine the sources of production line data, subcontractor data and project data, collect production line data, subcontractor data and project data, and preprocess the collected production line data, subcontractor data and project data; The comprehensive multi-model building module is used to build machine learning models suitable for dynamic capacity forecasting, subcontractor grouping evaluation and project optimization based on pre-processed production line data, subcontractor data and project data, namely, dynamic capacity forecasting model, subcontractor grouping evaluation model and project optimization model; The intelligent strategy formulation and adjustment module is used to formulate intelligent strategies based on the output results of each model and make dynamic adjustments based on actual conditions; The feedback collection and quantitative analysis module is used to establish feedback channels, collect feedback from production lines, subcontractors, and project management, use the database to record feedback data, calculate and analyze the quantitative indicator values ​​in the feedback data, and generate feedback signals.

Images