A dynamic production plan management method and system based on a feedback mechanism

Abstract

The application provides a dynamic production plan management method and system based on a feedback mechanism, a dynamic early warning model for production execution data is constructed, the dynamic early warning model is composed of a plurality of resource nodes reflecting node processing capacity and production execution state warning ranges, bottleneck identification is performed on the production execution data using the dynamic early warning model, whether there is a bottleneck resource node in the production and processing process is judged, the influence degree of the bottleneck resource node on the overall scheduling of the production plan is evaluated, scheduling exception data is generated when the influence degree of the bottleneck resource node on the overall scheduling of the production plan is greater than a preset value, production plan adjustment decision data is generated based on the scheduling exception data, cross-department feedback is performed on the production plan adjustment decision data, and more comprehensive resource optimization capability can be provided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of production plan management, and particularly relates to a dynamic production plan management method and system based on a feedback mechanism. BACKGROUND

[0002] Since the implementation process of production and processing will be affected by various abnormal situations such as equipment failure, resource shortage or the execution result not meeting the expectation, in the traditional production plan management scheme, the monitoring on the implementation process of the factory production and processing is performed to dynamically adjust the production plan according to the feedback of the production execution situation, which is a common means to enhance production control, improve abnormal response capability and optimize production resource allocation. The production plan adjustment relying on the production execution situation feedback involves multi-department cooperation, and is limited to the solidification of the cross-department linkage mechanism. The traditional production plan management scheme cannot monitor and respond to various indicators in the implementation process of production and processing in an all-round way, but can only configure one or more levels of warning thresholds of specific monitoring dimensions within a limited range, implement the established cross-department information transmission and production plan adjustment decision according to the warning indicators, and only realize local scheduling, which has limited resource optimization capability, and it is difficult to achieve the expectation in terms of abnormal response speed under the existing system framework of execution feedback and decision optimization. SUMMARY

[0003] The present application is based on the above problems, and proposes a dynamic production plan management method and system based on a feedback mechanism, which can provide more comprehensive resource optimization capability.

[0004] Therefore, the first aspect of the present application proposes a dynamic production plan management method based on a feedback mechanism, comprising:

[0005] A dynamic early warning model oriented to production execution data is constructed, and the dynamic early warning model is composed of a plurality of resource nodes reflecting node processing capability and production execution state warning range;

[0006] The dynamic early warning model is used to identify bottlenecks of the production execution data, so as to determine whether there is a bottleneck resource node in the production and processing process;

[0007] The influence degree of the bottleneck resource node on the overall scheduling of the production plan is evaluated;

[0008] When the influence degree of the bottleneck resource node on the overall scheduling of the production plan is greater than a preset value, scheduling abnormal data is generated;

[0009] Production plan adjustment decision data is generated based on the scheduling abnormal data;

[0010] The production plan adjustment decision data is executed for cross-department feedback.

[0011] Furthermore, the specific steps for constructing a dynamic early warning model based on production execution data include:

[0012] Construct a resource sequence to represent resource dimensions, where each data element in the resource sequence is a resource;

[0013] Construct a process sequence to represent the time dimension, where each data element in the process sequence is a production and processing step;

[0014] The resource dimension represented by the resource sequence is configured as the first orthogonal dimension, and the time dimension represented by the process sequence is configured as the second orthogonal dimension to generate an orthogonal data matrix framework.

[0015] A standard data matrix is ​​generated based on the orthogonal data matrix framework, where each data element of the standard data matrix is ​​the standard resource data of the corresponding resource node;

[0016] Obtain several pre-configured deviation ratio warning values;

[0017] Configure the standard data matrix and the deviation ratio warning value as the basic data elements of the dynamic early warning model.

[0018] Furthermore, after configuring the standard resource data and the deviation ratio warning value as the basic data elements of the dynamic early warning model, the method further includes:

[0019] Construct a lateral matching function for identifying lateral bottleneck nodes, a longitudinal matching function for identifying longitudinal bottleneck nodes, and a bottleneck node identification function for outputting bottleneck node identification results based on the lateral matching values ​​and longitudinal matching values ​​output by the lateral matching function and the longitudinal matching function;

[0020] The horizontal matching function, the vertical matching function, and the bottleneck node identification function are configured as the dynamic warning function of the dynamic early warning model.

[0021] Furthermore, the step of using the dynamic early warning model to identify bottlenecks in the production execution data specifically includes:

[0022] At the start of the production and processing process, a real-time data matrix of the current production and processing process is generated based on the orthogonal data matrix framework. The real-time data matrix is ​​a dynamic matrix whose size changes as the production and processing process progresses.

[0023] Real-time monitoring of production execution data during the production and processing process;

[0024] After each production and processing step is completed, the real-time data matrix is ​​updated based on the production execution data;

[0025] inputting the real-time resource data matrix into the dynamic early-warning model to perform bottleneck identification.

[0026] Further, the step of inputting the real-time resource data matrix into the dynamic early-warning model to perform bottleneck identification specifically comprises:

[0027] traversing the real-time data matrix to read real-time data of each resource node;

[0028] determining the traversed resource node as a first target resource node to perform the following steps:

[0029] generating a first horizontal sequence and a first vertical sequence of the first target resource node based on data of the real-time data matrix, and generating a second horizontal sequence and a second vertical sequence of the first target resource node based on data of the standard data matrix;

[0030] calculating a horizontal matching value of the first horizontal sequence and the second horizontal sequence, and calculating a vertical matching value of the first vertical sequence and the second vertical sequence;

[0031] judging whether the horizontal matching value and the vertical matching value of the first target resource node satisfy a preset condition;

[0032] when the horizontal matching value and the vertical matching value of the first target resource node satisfy the preset condition, determining the first target resource node as a candidate bottleneck resource node.

[0033] Further, after traversing each resource node in the real-time data matrix, further comprising:

[0034] judging whether the number of the candidate bottleneck resource nodes is zero;

[0035] when the number of the candidate bottleneck resource nodes is not zero, traversing the candidate bottleneck resource nodes;

[0036] determining the traversed candidate bottleneck resource node as a second target resource node to perform the following steps:

[0037] generating an alert level sequence of the second target resource node;

[0038] judging whether a proportion of the number of alert levels greater than zero in the alert level sequence of the second target resource node is greater than a preset bottleneck threshold;

[0039] when the proportion of the number of alert levels greater than zero in the alert level sequence of the second target resource node is greater than the preset bottleneck threshold, determining the second target resource node as a bottleneck resource node.

[0040] Further, the step of generating the alert level sequence of the second target resource node specifically comprises:

[0041] obtaining a first longitudinal sequence of the bottleneck resource node, the first longitudinal sequence being a sequence of resource values of completed processes of a resource type corresponding to the bottleneck resource node;

[0042] predicting a third longitudinal sequence of the bottleneck resource node based on the first longitudinal sequence, the third longitudinal sequence being a sequence of resource values in uncompleted processes of the resource type corresponding to the bottleneck resource node;

[0043] reading a fourth longitudinal sequence corresponding to the third longitudinal sequence from the standard data matrix;

[0044] calculating a deviation proportion sequence of the third longitudinal sequence relative to the fourth longitudinal sequence;

[0045] mapping the deviation proportion sequence into the alert level sequence of the second target resource node.

[0046] Further, the step of mapping the deviation proportion sequence into the alert level sequence of the second target resource node specifically comprises:

[0047] constructing a null value sequence of the alert level sequence of the second target resource node, the null value sequence having the same length as the deviation proportion sequence, and each data element in the null value sequence corresponding to each data element in the deviation proportion sequence;

[0048] traversing each deviation proportion value in the deviation proportion sequence to determine whether the deviation proportion value is less than a minimum deviation proportion alert value, the minimum deviation proportion alert value being an alert line of a minimum alert level of the second target resource node;

[0049] when any deviation proportion value in the deviation proportion sequence is less than the minimum deviation proportion alert value, filling a data element in the null value sequence corresponding to the deviation proportion value with 0;

[0050] when any deviation proportion value in the deviation proportion sequence is greater than or equal to the minimum deviation proportion alert value, determining an alert level corresponding to the deviation proportion value;

[0051] filling a data element in the null value sequence corresponding to the deviation proportion value with the alert level.

[0052] Further, the step of evaluating the influence degree of the bottleneck resource node on the overall scheduling of the production plan specifically comprises:

[0053] Each data element in the alert level sequence of the bottleneck resource node is represented as c l , wherein l∈[1, m-k+1], m-k+1 is the length of the alert level sequence of the bottleneck resource node;

[0054] The influence degree of the bottleneck resource node on the overall scheduling of the production plan is calculated based on the alert level sequence of the bottleneck resource node:

[0055]

[0056] A second aspect of the present application provides a dynamic production plan management system based on a feedback mechanism, comprising a warning model construction module for constructing a dynamic warning model oriented to production execution data, a bottleneck node identification module for identifying bottlenecks using the dynamic warning model to determine whether there are bottleneck resource nodes in the production process, and a production plan management module, the dynamic warning model is composed of a plurality of resource nodes reflecting node processing capacity and production execution state warning range, the production plan management module is configured to evaluate the influence degree of the bottleneck resource node on the overall scheduling of the production plan, generate scheduling exception data when the influence degree of the bottleneck resource node on the overall scheduling of the production plan is greater than a preset value, generate production plan adjustment decision data based on the scheduling exception data, and perform cross-department feedback on the production plan adjustment decision data.

[0057] The present application provides a dynamic production plan management method and system based on a feedback mechanism, by constructing a dynamic warning model oriented to production execution data, the dynamic warning model is composed of a plurality of resource nodes reflecting node processing capacity and production execution state warning range, using the dynamic warning model to identify bottlenecks to determine whether there are bottleneck resource nodes in the production process, evaluating the influence degree of the bottleneck resource node on the overall scheduling of the production plan, generating scheduling exception data when the influence degree of the bottleneck resource node on the overall scheduling of the production plan is greater than a preset value, generating production plan adjustment decision data based on the scheduling exception data, and performing cross-department feedback on the production plan adjustment decision data, which can provide more comprehensive resource optimization capability. BRIEF DESCRIPTION OF DRAWINGS

[0058] Figure 1 is a flowchart of a dynamic production plan management method based on a feedback mechanism provided by an embodiment of the present application;

[0059] Figure 2 is a schematic diagram of a dynamic production plan management system based on a feedback mechanism provided by an embodiment of the present application. DETAILED DESCRIPTION

[0060] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0061] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0062] In the description of this invention, the term "multiple" refers to two or more. Unless otherwise explicitly defined, the terms "upper," "lower," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. The terms "connect," "install," "fix," etc., should be interpreted broadly. For example, "connect" can be a fixed connection, a detachable connection, or an integral connection; it can be a direct connection or an indirect connection through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "multiple" means two or more.

[0063] In the description of this specification, the terms "one embodiment," "some implementations," "specific embodiment," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0064] The following description, with reference to the accompanying drawings, describes a dynamic production planning management method and system based on a feedback mechanism, according to some embodiments of the present invention.

[0065] like Figure 1 As shown, the first aspect of the present invention proposes a dynamic production planning management method based on a feedback mechanism, comprising:

[0066] A dynamic early warning model oriented to production execution data is constructed, and the dynamic early warning model is composed of a plurality of resource nodes reflecting node processing capacity and production execution state warning ranges;

[0067] A bottleneck identification is performed on the production execution data using the dynamic early warning model to determine whether there is a bottleneck resource node in the production process;

[0068] An influence degree of the bottleneck resource node on overall scheduling of the production plan is evaluated;

[0069] When the influence degree of the bottleneck resource node on overall scheduling of the production plan is greater than a preset value, scheduling exception data is generated;

[0070] Production plan adjustment decision data is generated based on the scheduling exception data;

[0071] Cross-department feedback is performed on the production plan adjustment decision data.

[0072] Specifically, the production execution data is real-time data generated in a production process of executing a production plan, and includes but is not limited to process time consumption, device state, and resource consumption data. In the technical solution of the present application, an MES (Manufacturing Execution System) collects the state of a production device in real time through an IoT device, records time consumption data and resource consumption data of each process, and integrates the data to obtain the production execution data.

[0073] The dynamic early warning model is a dynamic data model composed of a plurality of resource nodes arranged in the order of process execution, and is a comprehensive embodiment of real-time node processing capacity and production execution state warning ranges. In the technical solution of the present application, the consumption or use of a resource in a production process is personified as an entity, and this entity is referred to as a resource node. The resource referred to herein includes time resource, device resource, mold resource, material resource, and human resource.

[0074] The bottleneck resource node is a node whose processing capacity is significantly lower than that of other execution links. The processing capacity of a node refers to the deviation of the resource consumption or usage of the node from the standard data. The other execution links include other resource nodes of the same resource type in different production processes and resource nodes of different resource types in the same production process. Considering the normal fluctuation of production parameters caused by various objective factors in the production environment, the deviation of the processing capacity of each resource node, i.e., the resource consumption or usage of the resource node, from the standard data within a certain range is within expectations, and therefore only the resource node whose processing capacity is significantly lower than that of the other execution links is regarded as a bottleneck resource node. In actual implementation, the threshold value of the processing capacity of each resource node being regarded as "significantly lower" can be configured according to specific requirements, for example, a specific deviation value or a deviation ratio. In some embodiments, in order to simplify the configuration process, the threshold value can be configured as a unified deviation ratio.

[0075] In the technical solutions of some embodiments of the present application, the influence degree of the bottleneck resource node on the overall scheduling of the production plan can be represented by the influence degree of the bottleneck resource node on the alert level of the same resource type in the subsequent production process. In the technical solutions of some other embodiments of the present application, the influence degree of the bottleneck resource node on the overall scheduling of the production plan can be represented by the degree of change in production capacity caused by the bottleneck resource node within a specific time range.

[0076] In the technical solutions of the present application, the generated scheduling exception data includes but is not limited to at least one of the node number, resource type, deviation ratio, and alert level of the bottleneck resource node.

[0077] In the step of generating production plan adjustment decision data based on the scheduling exception data, the dynamic production plan management system formulates a corresponding resource allocation scheme, production plan delay or scheduling scheme, or equipment troubleshooting and maintenance scheme, etc.

[0078] Further, the step of constructing a dynamic early warning model for production execution data specifically includes:

[0079] A resource sequence for representing the resource dimension is constructed, and each data element in the resource sequence is a resource;

[0080] A process sequence for representing the time dimension is constructed, and each data element in the process sequence is a production process;

[0081] configuring a resource dimension represented by the resource sequence as a first orthogonal dimension, and configuring a time dimension represented by the process sequence as a second orthogonal dimension, to generate an orthogonal data matrix framework;

[0082] generating a standard data matrix on the basis of the orthogonal data matrix framework, each data element in the standard data matrix being standard resource data of a corresponding resource node;

[0083] obtaining a plurality of deviation proportion alarm values pre-configured;

[0084] configuring the standard data matrix and the deviation proportion alarm values as basic data elements of the dynamic early warning model.

[0085] Specifically, various types of resources are used in the production process, and in the step of constructing a resource sequence representing a resource dimension, a fixed number is configured for each type of resource to map it to a data element at a specific position in the resource sequence. Each type of resource corresponds to a data element with a specific number in the resource sequence, and the data elements are sequentially arranged in the resource sequence according to the number, i.e., the position of each type of resource in the resource sequence is a fixed position.

[0086] The orthogonal data matrix framework is a two-dimensional matrix, and each data element in the orthogonal data matrix framework represents a resource node, i.e., each resource node has a two-dimensional coordinate, the time coordinate of which corresponds to a data element in the process sequence, and the resource coordinate of which corresponds to a data element in the resource sequence.

[0087] The standard resource data is pre-configured standard data, for example, the standard time consumption of a production process is 35 minutes, and the standard data of the resource node corresponding to the resource type "process time consumption" of this production process is 35 minutes.

[0088] The deviation proportion alarm value is a plurality of alarm lines pre-configured, which are used to identify the bottleneck resource node according to the size of the deviation proportion of the real-time monitoring value of any monitoring parameter of the production execution data with respect to the standard data with respect to these alarm lines. In the technical solution of the present application, each alarm line is called an alarm level. For example, five alarm levels from 1 to 5 can be configured, each alarm level corresponding to a deviation proportion range determined by adjacent two deviation proportion alarm values, and the higher the deviation proportion, the higher the alarm level.

[0089] Further, after the step of configuring the standard resource data and the deviation proportion alarm values as basic data elements of the dynamic early warning model, it further includes:

[0090] constructing a lateral matching function for identifying lateral bottleneck nodes, a longitudinal matching function for identifying longitudinal bottleneck nodes, and a bottleneck node identification function for outputting a bottleneck node identification result according to lateral matching values and longitudinal matching values output by the lateral matching function and the longitudinal matching function;

[0091] configuring the lateral matching function, the longitudinal matching function, and the bottleneck node identification function as a dynamic warning function of the dynamic early warning model.

[0092] Specifically, the lateral bottleneck node is a resource node whose processing capacity is significantly lower than other resource nodes in the same production and processing procedure. Similarly, the longitudinal bottleneck node is a resource node whose processing capacity is significantly lower than other resource nodes of the same resource type in other production and processing procedures of the production and processing process.

[0093] Further, the step of using the dynamic early warning model to identify bottlenecks in the production execution data specifically includes:

[0094] At the beginning of the production and processing process, a real-time data matrix of the current production and processing process is generated based on the orthogonal data matrix framework, and the real-time data matrix is a dynamic matrix whose size changes as the production and processing process progresses;

[0095] Real-time monitoring of production execution data during the production and processing process;

[0096] After each production and processing procedure is completed, the real-time data matrix is updated according to the production execution data;

[0097] The real-time resource data matrix is input into the dynamic early warning model to perform bottleneck identification.

[0098] Specifically, the orthogonal data matrix framework corresponds to the complete production and processing process of one product, that is, the orthogonal data matrix framework corresponds to all production and processing procedures of one product in the time dimension. The standard data matrix contains standard data of various resource types in each procedure in the complete production and processing process of one product.

[0099] In the technical solution of the above embodiment, the real-time data matrix is a dynamic matrix, and the matrix size changes with the progress of the production and processing process of the product. More specifically, when the production and processing process of a product is in progress, i.e., the production and processing process is not completed, in the step of updating the real-time data matrix according to the production execution data, the real-time data matrix only contains the completed production and processing procedures. For example, when the production of a product involves n resources, and the production and processing process includes m production and processing procedures, both the orthogonal data matrix framework and the standard data matrix are n x m matrices, and when the production and processing process of the product is in the kth production and processing procedure, the real-time data matrix is an n x (k-1) matrix.

[0100] Further, the step of inputting the real-time resource data matrix into the dynamic early warning model to perform bottleneck identification specifically includes:

[0101] traversing the real-time data matrix to read the real-time data of each resource node;

[0102] determining the traversed resource node as a first target resource node to perform the following steps:

[0103] generating a first horizontal sequence and a first vertical sequence of the first target resource node based on the data of the real-time data matrix, and generating a second horizontal sequence and a second vertical sequence of the first target resource node based on the data of the standard data matrix;

[0104] calculating a horizontal matching value of the first horizontal sequence and the second horizontal sequence, and calculating a vertical matching value of the first vertical sequence and the second vertical sequence;

[0105] judging whether the horizontal matching value and the vertical matching value of the first target resource node satisfy a preset condition;

[0106] when the horizontal matching value and the vertical matching value of the first target resource node satisfy the preset condition, determining the first target resource node as a candidate bottleneck resource node.

[0107] Specifically, the first transverse sequence and the first longitudinal sequence are respectively a data sequence of transverse resource nodes and longitudinal resource nodes of the first target resource node in the real-time data matrix and the first target resource node itself. Similarly, the second transverse sequence and the second longitudinal sequence are respectively a data sequence of transverse resource nodes and longitudinal resource nodes of the first target resource node in the standard data matrix and the first target resource node itself. The transverse resource node is another resource node having the same time dimension coordinate as the first target resource node in the data matrix, and the longitudinal resource node is another resource node having the same resource dimension coordinate as the first target resource node in the data matrix.

[0108] Preferably, the second longitudinal sequence has the same length as the first longitudinal sequence, and the data elements in the second longitudinal sequence correspond to the data elements in the first longitudinal sequence one by one. More specifically, the second longitudinal sequence only includes the standard data of the resource nodes of the same completed production processing procedure in the first longitudinal sequence, and does not include the standard data of the resource nodes of the uncompleted production processing procedure.

[0109] Further, the step of calculating the transverse matching value of the first transverse sequence and the second transverse sequence specifically comprises:

[0110] The data elements in the first transverse sequence are represented as hi i The data elements in the second transverse sequence are represented as h2 i wherein i∈[1, n], n is the length of the first transverse sequence and the second transverse sequence;

[0111] The first deviation ratio sequence of the first transverse sequence and the second transverse sequence is calculated and generated, and each data element in the first deviation ratio sequence is represented as rh i

[0112]

[0113] It is judged whether the deviation ratio of the first target resource node is less than a minimum deviation ratio warning value, wherein the minimum deviation ratio warning value is a warning line of a minimum warning level of the first target resource node, and i0 is the serial number of the first target resource node in the first transverse sequence and the second transverse sequence. When the deviation ratio of the first target resource node is less than the minimum deviation ratio warning value, the transverse matching value of the first target resource node is configured as 0.

[0114]

[0115] ​​​When the deviation ratio of the first target resource node When the deviation ratio is greater than or equal to the minimum deviation ratio warning value, the deviation ratio of the first target resource node is determined. Does it meet the following requirements:

[0116]

[0117] When the condition is met, the horizontal matching value of the first target resource node is configured to 1;

[0118] If the condition is not met, the horizontal matching value of the first target resource node is configured to 0.

[0119] Furthermore, the step of calculating the vertical matching value of the first vertical sequence and the second vertical sequence specifically includes:

[0120] Let the data element in the first vertical sequence be represented as v1. j The data elements in the second vertical sequence are represented as v2. j , where j∈[1,k-1], and k-1 is the length of the first vertical sequence and the second vertical sequence;

[0121] Calculate and generate a second deviation ratio sequence for the first longitudinal sequence and the second longitudinal sequence, and represent each data element in the second deviation ratio sequence as rv. j ,but

[0122]

[0123] Determine the deviation ratio of the first target resource node Whether it is less than the minimum deviation ratio warning value, where j0 is the sequence number of the first target resource node in the first vertical sequence and the second vertical sequence, and

[0124] When the deviation ratio of the first target resource node When the value is less than the minimum deviation ratio warning value, the vertical matching value of the first target resource node is configured to 0;

[0125] When the deviation ratio of the first target resource node When the deviation ratio is greater than or equal to the minimum deviation ratio warning value, the deviation ratio of the first target resource node is determined. Does it meet the following requirements:

[0126]

[0127] When the condition is met, the vertical matching value of the first target resource node is configured to 1;

[0128] If not, the vertical matching value of the first target resource node is configured as 0.

[0129] In the technical scheme of some embodiments of the present application, the step of judging whether the horizontal matching value and the vertical matching value of the first target resource node satisfy the preset condition specifically comprises:

[0130] judging whether at least one of the horizontal matching value and the vertical matching value of the first target resource node is 1;

[0131] When at least one of the horizontal matching value and the vertical matching value of the first target resource node is 1, the first target resource node is determined as a candidate bottleneck resource node.

[0132] Further, after traversing each resource node in the real-time data matrix, the method further comprises:

[0133] judging whether the number of the candidate bottleneck resource nodes is zero;

[0134] When the number of the candidate bottleneck resource nodes is not zero, traversing the candidate bottleneck resource nodes;

[0135] determining the traversed candidate bottleneck resource node as a second target resource node to execute the following steps:

[0136] generating an alert level sequence of the second target resource node;

[0137] judging whether the proportion of the number of alert levels greater than zero in the alert level sequence of the second target resource node is greater than a preset bottleneck threshold;

[0138] When the proportion of the number of alert levels greater than zero in the alert level sequence of the second target resource node is greater than the preset bottleneck threshold, the second target resource node is determined as a bottleneck resource node.

[0139] Specifically, when the values of each resource node in the production and processing process in its corresponding production and processing process are all within the expected range, i.e., none of the deviation proportions of the resource nodes is greater than the minimum deviation proportion alert value, there is no candidate bottleneck resource node. In this case, the number of the candidate bottleneck resource nodes is zero.

[0140] Further, the step of judging whether the proportion of the number of alert levels greater than zero in the alert level sequence of the second target resource node is greater than the preset bottleneck threshold specifically comprises:

[0141] obtaining a preset bottleneck threshold r0;

[0142] determining the length m-k+1 of the alert level sequence;

[0143] determining a number w of alert levels greater than zero in the alert level sequence of the second target resource node;

[0144] calculating a proportion of the number of alert levels greater than zero in the alert level sequence of the second target resource node:

[0145]

[0146] determining whether the proportion r of the number of alert levels greater than zero in the alert level sequence of the second target resource node is greater than a preset bottleneck threshold r0.

[0147] Further, the step of generating the alert level sequence of the second target resource node specifically comprises:

[0148] obtaining a first longitudinal sequence of the bottleneck resource node, the first longitudinal sequence being a resource value sequence of completed working procedures of a resource type corresponding to the bottleneck resource node;

[0149] predicting a third longitudinal sequence of the bottleneck resource node based on the first longitudinal sequence, the third longitudinal sequence being a resource value sequence in an uncompleted working procedure of the resource type corresponding to the bottleneck resource node;

[0150] reading a fourth longitudinal sequence corresponding to the third longitudinal sequence from the standard data matrix;

[0151] calculating a deviation proportion sequence of the third longitudinal sequence relative to the fourth longitudinal sequence;

[0152] mapping the deviation proportion sequence as the alert level sequence of the second target resource node.

[0153] Similarly, the fourth longitudinal sequence has the same length as the third longitudinal sequence, and data elements in the fourth longitudinal sequence one-to-one correspond to data elements in the third longitudinal sequence. More specifically, the fourth longitudinal sequence only includes standard data of resource nodes of the same uncompleted production and processing working procedure as the third longitudinal sequence, and does not include standard data of resource nodes of completed production and processing working procedures.

[0154] In the technical solution of some embodiments of the present application, after the production and processing process of each product ends, a complete real-time resource data matrix, i.e., a matrix data of size n x m, is recorded as historical production data. After a sufficient number of historical production data are collected, the real-time resource data matrix in the historical production data is used to train a prediction model of resource value sequences in uncompleted working procedures, and the step of training the prediction model specifically comprises:

[0155] configuring a minimum input data length minL in and a minimum output data length minLout ;

[0156] using the historical production data to construct a plurality of effective input data sets within a range from minL in to m-minL in ; out ;

[0157] using the historical production data to construct a plurality of effective output data sets with an effective output data length of m-L in ;

[0158] padding at the end of each input data sequence and output data sequence in the effective input data sets and the effective output data sets to supplement the data length of the input data sequence and the output data sequence in the effective input data sets and the effective output data sets to m, to obtain an input sample data set and a corresponding output sample data set of training sample data;

[0159] configuring a convolutional neural network framework with a plurality of convolutional layers and training parameters thereof;

[0160] training the convolutional neural network framework based on the training parameters using the input sample data set and the corresponding output sample data set to obtain the prediction model.

[0161] Further, the step of mapping the deviation ratio sequence to the alert level sequence of the second target resource node specifically comprises:

[0162] constructing a null sequence of the alert level sequence of the second target resource node, the length of the null sequence being the same as the length of the deviation ratio sequence, and each data element in the null sequence corresponding to each data element in the deviation ratio sequence one by one;

[0163] traversing each deviation ratio value in the deviation ratio sequence to determine whether the deviation ratio value is less than a minimum deviation ratio alert value, the minimum deviation ratio alert value being an alert line of a minimum alert level of the second target resource node;

[0164] when any deviation ratio value in the deviation ratio sequence is less than the minimum deviation ratio alert value, filling the data element corresponding to the deviation ratio value in the null sequence with 0;

[0165] when any deviation ratio value in the deviation ratio sequence is greater than or equal to the minimum deviation ratio alert value, determining the alert level corresponding to the deviation ratio value;

[0166] filling the data element corresponding to the deviation ratio value in the null sequence with the alert level.

[0167] In the technical solution of the above embodiment, the process of filling the null sequence is the process of generating the alert level sequence of the second target resource node. After traversing the deviation ratio sequence, each null value in the null sequence is replaced by 0 or an alert level, thereby obtaining the alert level sequence of the second target resource node.

[0168] Further, the step of evaluating the influence degree of the bottleneck resource node on the overall scheduling of the production plan specifically comprises:

[0169] each data element in the alert level sequence of the bottleneck resource node is represented as c l , where l∈[1, m-k+1], and m-k+1 is the length of the alert level sequence of the bottleneck resource node.

[0170] The influence degree of the bottleneck resource node on the overall scheduling of the production plan is calculated based on the alert level sequence of the bottleneck resource node:

[0171]

[0172] Specifically, in the alert level sequence of the bottleneck resource node, each data element is the alert level corresponding to the deviation ratio value of a data element in the third longitudinal sequence.

[0173] As Figure 2 shown, the second aspect of the present application proposes a dynamic production plan management system based on a feedback mechanism, which includes a warning model construction module for constructing a dynamic warning model oriented to production execution data, a bottleneck node identification module for identifying a bottleneck resource node in a production process using the dynamic warning model to determine whether there is a bottleneck resource node in the production process, and a production plan management module. The dynamic warning model is composed of a plurality of resource nodes reflecting node processing capacity and production execution state warning range. The production plan management module is configured to evaluate the influence degree of the bottleneck resource node on the overall scheduling of the production plan, generate scheduling exception data when the influence degree of the bottleneck resource node on the overall scheduling of the production plan is greater than a preset value, generate production plan adjustment decision data based on the scheduling exception data, and perform cross-department feedback on the production plan adjustment decision data.

[0174] In the technical scheme of some embodiments of the present application, the dynamic production plan management system based on the feedback mechanism is an ASP (Advanced Planning and Scheduling) system, which acquires production execution data from the MES, generates scheduling exception data after identifying bottleneck resource nodes, and generates production plan adjustment decision data to feed back to the MES and the ERP (Enterprise Resource Planning) system.

[0175] Specifically, the production execution data are real-time data generated in the production and processing process of the executed production plan, including but not limited to process time consumption, equipment state, and resource consumption data. In the technical scheme of the present application, the MES collects the state of the production and processing equipment in real time through IoT devices, records the time consumption data and resource consumption data of each process, and integrates them to obtain the production execution data.

[0176] The dynamic early warning model is a dynamic data model composed of resource nodes arranged in the order of process execution, which is a comprehensive embodiment of real-time node processing capacity and production execution state warning range. In the technical scheme of the present application, the consumption or use of a resource in a production and processing process is personified as an entity, and this entity is referred to as a resource node. The resources referred to herein include time resources, equipment resources, mold resources, material resources, and human resources.

[0177] The bottleneck resource node is a resource node whose real-time node processing capacity is significantly lower than that of other execution links. The node processing capacity refers to the deviation of the resource consumption or use of the resource node in the production execution data from the standard data. The other execution links include other resource nodes of the same resource type in different production and processing processes, and resource nodes of different resource types in the same production and processing process. Considering the normal fluctuation of production and processing parameters caused by various objective factors in the production and processing environment, the deviation of the processing capacity of each resource node, i.e., the resource consumption or use of the resource node, from the standard data within a certain range is within expectations, so only the resource node whose node processing capacity is significantly lower than that of other execution links is regarded as a bottleneck resource node. In actual implementation, the threshold value of each resource node being regarded as "significantly lower" can be configured according to specific needs, such as a specific deviation value or a deviation ratio. In some embodiments, in order to simplify the configuration process, the threshold value can be configured as a unified deviation ratio.

[0178] In the technical solution of some embodiments of the present application, the influence degree of the bottleneck resource node on the overall scheduling of the production plan can be represented by the influence degree of the bottleneck resource node on the alert level of the same resource type in the subsequent production and processing procedures. In the technical solution of some other embodiments of the present application, the influence degree of the bottleneck resource node on the overall scheduling of the production plan can be represented by the degree of change in production capacity caused by the bottleneck resource node within a specific time range.

[0179] In the technical solution of the present application, the generated scheduling exception data includes but is not limited to at least one of the node number, resource type, deviation ratio, and alert level of the bottleneck resource node.

[0180] In the step of generating production plan adjustment decision data based on the scheduling exception data, the dynamic production plan management system formulates a corresponding resource allocation scheme, production plan delay or scheduling scheme, or equipment fault diagnosis, maintenance scheme, etc. based on the scheduling exception data.

[0181] Further, the early warning model construction module is configured to:

[0182] construct a resource sequence for representing the resource dimension, each data element in the resource sequence being a kind of resource;

[0183] construct a procedure sequence for representing the time dimension, each data element in the procedure sequence being a production and processing procedure;

[0184] configure the resource dimension represented by the resource sequence as a first orthogonal dimension, and configure the time dimension represented by the procedure sequence as a second orthogonal dimension, to generate an orthogonal data matrix framework;

[0185] generate a standard data matrix based on the orthogonal data matrix framework, each data element in the standard data matrix being standard resource data of a corresponding resource node;

[0186] obtain a plurality of deviation ratio alert values pre-configured;

[0187] configure the standard data matrix and the deviation ratio alert values as the basic data elements of the dynamic early warning model.

[0188] Specifically, various types of resources are used in the production and processing process. In the step of constructing a resource sequence to represent the resource dimension, a fixed number is assigned to each resource to map it to a data element at a specific position in the resource sequence. Each resource corresponds to a data element with a specific number in the resource sequence. The data elements are arranged in order in the resource sequence according to the number, that is, the position of each resource in the resource sequence is a fixed position.

[0189] The orthogonal data matrix framework is a two-dimensional matrix. Each data element in the orthogonal data matrix framework represents a resource node. That is, each resource node has a two-dimensional coordinate. The time coordinate in its two-dimensional coordinate corresponds to a data element in the process sequence, and the resource coordinate in its two-dimensional coordinate corresponds to a data element in the resource sequence.

[0190] The standard resource data is pre-configured standard data. For example, if the standard time for a production and processing step is 35 minutes, then the standard data for the resource node with the resource type "process time" corresponding to that production and processing step is 35 minutes.

[0191] The deviation ratio warning values ​​are a set of pre-configured warning lines used to identify bottleneck resource nodes based on the magnitude of the deviation ratio of the real-time monitoring value of a certain monitoring parameter corresponding to any resource node in the production execution data relative to the standard data and these warning lines. In the technical solution of this invention, each warning line is called a warning level. For example, five warning levels from 1 to 5 can be configured, each warning level corresponding to the deviation ratio range determined by two adjacent deviation ratio warning values; the larger the deviation ratio, the higher the warning level.

[0192] Furthermore, after configuring the standard resource data and the deviation ratio warning value as the basic data elements of the dynamic early warning model, the early warning model construction module is configured as follows:

[0193] Construct a lateral matching function for identifying lateral bottleneck nodes, a longitudinal matching function for identifying longitudinal bottleneck nodes, and a bottleneck node identification function for outputting bottleneck node identification results based on the lateral matching values ​​and longitudinal matching values ​​output by the lateral matching function and the longitudinal matching function;

[0194] The horizontal matching function, the vertical matching function, and the bottleneck node identification function are configured as the dynamic warning function of the dynamic early warning model.

[0195] Specifically, the transverse bottleneck node is a resource node whose processing capacity is significantly lower than other resource nodes in the same production and processing procedure. Similarly, the longitudinal bottleneck node is a resource node whose processing capacity is significantly lower than other resource nodes of the same resource type in other production and processing procedures of the production and processing process.

[0196] Further, the bottleneck node identification module is configured to:

[0197] At the beginning of the production and processing process, a real-time data matrix of the current production and processing process is generated based on the orthogonal data matrix framework, and the real-time data matrix is a dynamic matrix whose size changes with the progress of the production and processing process;

[0198] The production execution data is monitored in real time during the production and processing process;

[0199] After each production and processing procedure is completed, the real-time data matrix is updated according to the production execution data;

[0200] The real-time resource data matrix is input into the dynamic early warning model to perform bottleneck identification.

[0201] Specifically, the orthogonal data matrix framework corresponds to the complete production and processing process of one product, that is, the orthogonal data matrix framework corresponds to all production and processing procedures of one product in the time dimension. The standard data matrix contains standard data of various resource types in each procedure of the complete production and processing process of one product.

[0202] In the technical solution of the above embodiment, the real-time data matrix is a dynamic matrix, and its matrix size changes with the progress of the production and processing process of the product. More specifically, when the production and processing process of one product is in progress, that is, the production and processing process is not completed, in the step of updating the real-time data matrix according to the production execution data, the real-time data matrix only contains the completed production and processing procedures. For example, when the production of one product involves n resources, and the production and processing process includes m production and processing procedures, the orthogonal data matrix framework and the standard data matrix are both n×m matrices, and when the production and processing process of the product is in the kth production and processing procedure, the real-time data matrix is an n×(k-1) matrix.

[0203] Further, in the step of inputting the real-time resource data matrix into the dynamic early warning model to perform bottleneck identification, the bottleneck node identification module is configured to:

[0204] Traverse the real-time data matrix to read the real-time data of each resource node;

[0205] Determine the traversed resource node as a first target resource node to perform the following steps:

[0206] generate a first horizontal sequence and a first vertical sequence of the first target resource node based on data of the real-time data matrix, and generate a second horizontal sequence and a second vertical sequence of the first target resource node based on data of the standard data matrix;

[0207] calculate a horizontal matching value of the first horizontal sequence and the second horizontal sequence, and calculate a vertical matching value of the first vertical sequence and the second vertical sequence;

[0208] determine whether the horizontal matching value and the vertical matching value of the first target resource node satisfy a preset condition;

[0209] determine the first target resource node as a candidate bottleneck resource node when the horizontal matching value and the vertical matching value of the first target resource node satisfy the preset condition.

[0210] Specifically, the first horizontal sequence and the first vertical sequence are data sequences respectively composed of horizontal resource nodes and vertical resource nodes of the first target resource node in the real-time data matrix and the first target resource node itself. Similarly, the second horizontal sequence and the second vertical sequence are data sequences respectively composed of horizontal resource nodes and vertical resource nodes of the first target resource node in the standard data matrix and the first target resource node itself. The horizontal resource nodes are other resource nodes in the data matrix having the same time dimension coordinate as the first target resource node, and the vertical resource nodes are other resource nodes in the data matrix having the same resource dimension coordinate as the first target resource node.

[0211] Preferably, the second vertical sequence has the same length as the first vertical sequence, and data elements in the second vertical sequence one-to-one correspond to data elements in the first vertical sequence. More specifically, the second vertical sequence only includes standard data of resource nodes of the same completed production and processing procedures as the first vertical sequence, and does not include standard data of resource nodes of uncompleted production and processing procedures.

[0212] Further, in the step of calculating the horizontal matching value of the first horizontal sequence and the second horizontal sequence, the bottleneck node identification module is configured to:

[0213] express data elements in the first horizontal sequence as h1 i express data elements in the second horizontal sequence as h2 i wherein i∈[1, n], and n is the length of the first horizontal sequence and the second horizontal sequence;

[0214] a first deviation ratio sequence of the first lateral sequence and the second lateral sequence is calculated and generated, each data element in the first deviation ratio sequence is represented as rh i , then

[0215]

[0216] a deviation ratio of the first target resource node is determined whether the deviation ratio of the first target resource node is less than a minimum deviation ratio alarm value, wherein the minimum deviation ratio alarm value is an alarm line of a minimum alarm level of the first target resource node, i0 is a serial number of the first target resource node in the first lateral sequence and the second lateral sequence;

[0217] when the deviation ratio of the first target resource node is less than the minimum deviation ratio alarm value, a lateral matching value of the first target resource node is configured as 0;

[0218] when the deviation ratio of the first target resource node is greater than or equal to the minimum deviation ratio alarm value, whether the deviation ratio of the first target resource node satisfies:

[0219] when the deviation ratio of the first target resource node satisfies, the lateral matching value of the first target resource node is configured as 1;

[0220] when the deviation ratio of the first target resource node does not satisfy, the lateral matching value of the first target resource node is configured as 0.

[0221] Further, in the step of calculating the longitudinal matching value of the first longitudinal sequence and the second longitudinal sequence, the bottleneck node identification module is configured to:

[0222] a data element in the first longitudinal sequence is represented as v1 j , and a data element in the second longitudinal sequence is represented as v2 j , wherein j ∈ [1, k-1], and k-1 is a length of the first longitudinal sequence and the second longitudinal sequence;

[0223] a second deviation ratio sequence of the first longitudinal sequence and the second longitudinal sequence is calculated and generated, each data element in the second deviation ratio sequence is represented as rv j , then

[0224]

[0225] a deviation ratio of the first target resource node is determined

[0226] whether the deviation ratio of the first target resource node is less than a minimum deviation ratio alarm value, wherein the minimum deviation ratio alarm value is an alarm line of a minimum alarm level of the first target resource node, i0 is a serial number of the first target resource node in the first lateral sequence and the second lateral sequence; ​​is less than a minimum deviation proportion alarm value, where j0 is the serial number of the first target resource node in the first longitudinal sequence and the second longitudinal sequence, and

[0227] when the deviation proportion of the first target resource node is less than the minimum deviation proportion alarm value, the longitudinal matching value of the first target resource node is configured as 0;

[0228] when the deviation proportion of the first target resource node is greater than or equal to the minimum deviation proportion alarm value, it is determined whether the deviation proportion of the first target resource node satisfies:

[0229]

[0230] when the deviation proportion of the first target resource node satisfies, the longitudinal matching value of the first target resource node is configured as 1;

[0231] when the deviation proportion of the first target resource node does not satisfy, the longitudinal matching value of the first target resource node is configured as 0.

[0232] In the technical solution of some embodiments of the application, in the step of determining whether the horizontal matching value and the longitudinal matching value of the first target resource node satisfy the preset condition, the bottleneck node identification module is configured to:

[0233] determine whether at least one of the horizontal matching value and the longitudinal matching value of the first target resource node is 1;

[0234] when at least one of the horizontal matching value and the longitudinal matching value of the first target resource node is 1, the first target resource node is determined as a candidate bottleneck resource node.

[0235] Further, after traversing each resource node in the real-time data matrix, the bottleneck node identification module is configured to:

[0236] determine whether the number of candidate bottleneck resource nodes is zero;

[0237] when the number of candidate bottleneck resource nodes is not zero, traverse the candidate bottleneck resource nodes;

[0238] determine the traversed candidate bottleneck resource node as a second target resource node to perform the following steps:

[0239] generate an alarm level sequence of the second target resource node;

[0240] determine whether the proportion of the number of alarm levels greater than zero in the alarm level sequence of the second target resource node is greater than a preset bottleneck threshold.​​​

[0241] When the proportion of alert levels greater than zero in the alert level sequence of the second target resource node is greater than the preset bottleneck threshold, the second target resource node is determined as a bottleneck resource node.

[0242] Specifically, when the values ​​of each resource node in its corresponding production and processing step are all within the expected range, that is, when the deviation ratio of no resource node exceeds the minimum deviation ratio warning value, then there are no candidate bottleneck resource nodes. In this case, the number of candidate bottleneck resource nodes is zero.

[0243] Furthermore, in the step of determining whether the proportion of alert levels greater than zero in the alert level sequence of the second target resource node is greater than a preset bottleneck threshold, the bottleneck node identification module is configured as follows:

[0244] Obtain the preset bottleneck threshold r0;

[0245] Determine the length m-k+1 of the alert level sequence;

[0246] Determine the number w of alert levels greater than zero in the alert level sequence;

[0247] Calculate the percentage of alert levels greater than zero in the alert level sequence of the second target resource node:

[0248]

[0249] Determine whether the percentage r of the number of warning levels greater than zero in the warning level sequence of the second target resource node is greater than the preset bottleneck threshold r0.

[0250] Furthermore, in the step of generating the alert level sequence of the second target resource node, the bottleneck node identification module is configured as follows:

[0251] Obtain the first vertical sequence of the bottleneck resource node, wherein the first vertical sequence is the resource value sequence of the completed process of the resource type corresponding to the bottleneck resource node;

[0252] Based on the first vertical sequence, a third vertical sequence of the bottleneck resource node is predicted, wherein the third vertical sequence is a sequence of resource values ​​in the unfinished process of the resource type corresponding to the bottleneck resource node;

[0253] Read the fourth vertical sequence corresponding to the third vertical sequence from the standard data matrix;

[0254] The deviation ratio sequence of the third longitudinal sequence relative to the fourth longitudinal sequence is calculated;

[0255] The deviation ratio sequence is mapped to the warning level sequence of the second target resource node.

[0256] Similarly, the fourth vertical sequence has the same length as the third vertical sequence, and the data elements in the fourth vertical sequence correspond one-to-one with the data elements in the third vertical sequence. More specifically, the fourth vertical sequence only includes standard data of resource nodes for incomplete production and processing steps that are the same as those in the third vertical sequence, and does not include standard data of resource nodes for completed production and processing steps.

[0257] In some embodiments of the present invention, after the production and processing of each product is completed, a complete real-time resource data matrix, i.e., a matrix of size n×m, is recorded as historical production data. After collecting a sufficient amount of historical production data, the real-time resource data matrix in the historical production data is used to train a predictive model for the resource value sequence in incomplete processes. The dynamic production planning management system based on the feedback mechanism further includes a predictive model training module, which is configured as follows:

[0258] Configure minimum input data length minL in and minimum output data length minL out ;

[0259] Using the historical production data within the effective input data length L in From minL in to m-minL out Construct several valid input datasets within the specified interval;

[0260] The historical production data is used to construct an effective output data length of mL. in Several valid output datasets;

[0261] Padding numbers are added to the end of each input data sequence and output data sequence in the valid input dataset and the valid output dataset to make the data length of the input data sequence and output data sequence in the valid input dataset and the valid output dataset m, so as to obtain the input sample dataset and the corresponding output sample dataset of the training sample data;

[0262] Configure a convolutional neural network framework with multiple convolutional layers and its training parameters;

[0263] Based on the training parameters, the prediction model is obtained by training the input sample dataset and the corresponding output sample dataset within the convolutional neural network framework.

[0264] Further, in the step of mapping the deviation ratio sequence to the alert level sequence of the second target resource node, the bottleneck node identification module is configured to:

[0265] constructing a null sequence of the alert level sequence of the second target resource node, the null sequence having the same length as the deviation ratio sequence, and each data element in the null sequence corresponding to each data element in the deviation ratio sequence;

[0266] traversing each deviation ratio value in the deviation ratio sequence to determine whether the deviation ratio value is less than a minimum deviation ratio alert value, the minimum deviation ratio alert value being an alert line of a minimum alert level of the second target resource node;

[0267] when any deviation ratio value in the deviation ratio sequence is less than the minimum deviation ratio alert value, filling the data element in the null sequence corresponding to the deviation ratio value with 0;

[0268] when any deviation ratio value in the deviation ratio sequence is greater than or equal to the minimum deviation ratio alert value, determining the alert level corresponding to the deviation ratio value;

[0269] filling the data element in the null sequence corresponding to the deviation ratio value with the alert level.

[0270] In the technical solution of the above embodiment, the process of filling the null sequence is the process of generating the alert level sequence of the second target resource node. After traversing the deviation ratio sequence, each null value in the null sequence is replaced by 0 or an alert level, thereby obtaining the alert level sequence of the second target resource node.

[0271] Further, the production plan management module is configured to:

[0272] representing each data element in the alert level sequence of the bottleneck resource node as c l , where l∈[1, m-k+1], m-k+1 being the length of the alert level sequence of the bottleneck resource node;

[0273] calculating the influence degree of the bottleneck resource node on the overall scheduling of the production plan based on the alert level sequence of the bottleneck resource node:

[0274]

[0275] Specifically, in the alert level sequence of the bottleneck resource node, each data element is the alert level corresponding to the deviation ratio value of a data element in the third longitudinal sequence.

[0276] It should be noted that, as used in this place, the terms "first", "second", etc. are employed merely to identify various elements or acts, without necessarily requiring or implying any actual relationship or order between such elements or acts. Moreover, the terms "comprises", "comprising", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a" does not, without more constraints, exclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the recited element.

[0277] In accordance with the practices of the present application, these embodiments have been described in relation to the above-described embodiments, which are intended to be illustrative only and not restrictive of the application. There many modifications and variations to the embodiments described herein that will be apparent to those of ordinary skill in the art. It is therefore contemplated that the application shall not be limited to the particular embodiments described herein, but rather the scope of the present application is to be given the broadest interpretation of the appended claims to use all equivalents thereof.

Claims

1. A dynamic production planning management method based on a feedback mechanism, characterized in that, include: A dynamic early warning model for production execution data is constructed, which consists of several resource nodes that reflect the processing capacity of nodes and the warning range of production execution status. The dynamic early warning model is used to identify bottlenecks in the production execution data in order to determine whether there are bottleneck resource nodes in the production process. Assess the impact of the bottleneck resource nodes on the overall production schedule; When the impact of the bottleneck resource node on the overall production schedule exceeds a preset value, scheduling anomaly data is generated. Based on the scheduling anomaly data, production plan adjustment decision data is generated; Cross-departmental feedback is performed on the production plan adjustment decision data; The specific steps for building a dynamic early warning model based on production execution data include: Construct a resource sequence to represent resource dimensions, where each data element in the resource sequence is a resource; Construct a process sequence to represent the time dimension, where each data element in the process sequence is a production and processing step; The resource dimension represented by the resource sequence is configured as the first orthogonal dimension, and the time dimension represented by the process sequence is configured as the second orthogonal dimension to generate an orthogonal data matrix framework. A standard data matrix is ​​generated based on the orthogonal data matrix framework, where each data element of the standard data matrix is ​​the standard resource data of the corresponding resource node; Obtain several pre-configured deviation ratio warning values; Configure the standard data matrix and the deviation ratio warning value as the basic data elements of the dynamic early warning model; After configuring the standard resource data and the deviation ratio warning value as the basic data elements of the dynamic early warning model, the method further includes: Construct a lateral matching function for identifying lateral bottleneck nodes, a longitudinal matching function for identifying longitudinal bottleneck nodes, and a bottleneck node identification function for outputting bottleneck node identification results based on the lateral matching values ​​and longitudinal matching values ​​output by the lateral matching function and the longitudinal matching function; The horizontal matching function, the vertical matching function, and the bottleneck node identification function are configured as the dynamic warning function of the dynamic early warning model; The steps for using the dynamic early warning model to identify bottlenecks in the production execution data specifically include: At the start of the production and processing process, a real-time data matrix of the current production and processing process is generated based on the orthogonal data matrix framework. The real-time data matrix is ​​a dynamic matrix whose size changes as the production and processing process progresses. Real-time monitoring of production execution data during the production and processing process; After each production and processing step is completed, the real-time data matrix is ​​updated based on the production execution data; The real-time data matrix is ​​input into the dynamic early warning model to perform bottleneck identification.

2. The dynamic production planning management method based on a feedback mechanism according to claim 1, characterized in that, The steps of inputting the real-time data matrix into the dynamic early warning model to perform bottleneck identification specifically include: The real-time data matrix is ​​traversed to read the real-time data of each resource node; The resource node that has been traversed is identified as the first target resource node to perform the following steps: Based on the data from the real-time data matrix, a first horizontal sequence and a first vertical sequence of the first target resource node are generated, and based on the data from the standard data matrix, a second horizontal sequence and a second vertical sequence of the first target resource node are generated. Calculate the horizontal matching value between the first horizontal sequence and the second horizontal sequence, and calculate the vertical matching value between the first vertical sequence and the second vertical sequence; Determine whether the horizontal and vertical matching values ​​of the first target resource node meet the preset conditions; When the horizontal and vertical matching values ​​of the first target resource node meet the preset conditions, the first target resource node is determined as a candidate bottleneck resource node.

3. The dynamic production planning management method based on a feedback mechanism according to claim 2, characterized in that, After traversing each resource node in the real-time data matrix, the process also includes: Determine whether the number of candidate bottleneck resource nodes is zero; When the number of candidate bottleneck resource nodes is not zero, traverse the candidate bottleneck resource nodes. The candidate bottleneck resource nodes encountered during the traversal are identified as the second target resource nodes to perform the following steps: Generate a warning level sequence for the second target resource node; Determine whether the proportion of alert levels greater than zero in the alert level sequence of the second target resource node is greater than a preset bottleneck threshold. When the proportion of alert levels greater than zero in the alert level sequence of the second target resource node is greater than the preset bottleneck threshold, the second target resource node is determined as a bottleneck resource node.

4. The dynamic production planning management method based on a feedback mechanism according to claim 3, characterized in that, The steps for generating the alert level sequence for the second target resource node specifically include: Obtain the first vertical sequence of the bottleneck resource node, wherein the first vertical sequence is the resource value sequence of the completed process of the resource type corresponding to the bottleneck resource node; Based on the first vertical sequence, a third vertical sequence of the bottleneck resource node is predicted, wherein the third vertical sequence is a sequence of resource values ​​in the unfinished process of the resource type corresponding to the bottleneck resource node; Read the fourth vertical sequence corresponding to the third vertical sequence from the standard data matrix; The deviation ratio sequence of the third longitudinal sequence relative to the fourth longitudinal sequence is calculated; The deviation ratio sequence is mapped to the warning level sequence of the second target resource node.

5. The dynamic production planning management method based on a feedback mechanism according to claim 4, characterized in that, The step of mapping the deviation ratio sequence to the alert level sequence of the second target resource node specifically includes: Construct a null value sequence of the alert level sequence of the second target resource node, wherein the length of the null value sequence is the same as the length of the deviation ratio sequence, and each data element in the null value sequence corresponds one-to-one with each data element in the deviation ratio sequence. Traverse each deviation ratio value in the deviation ratio sequence to determine whether the deviation ratio value is less than the minimum deviation ratio warning value, where the minimum deviation ratio warning value is the warning line for the minimum warning level of the second target resource node; When any deviation ratio value in the deviation ratio sequence is less than the minimum deviation ratio warning value, the data element in the null value sequence corresponding to the deviation ratio value is filled with 0. When any deviation ratio value in the deviation ratio sequence is greater than or equal to the minimum deviation ratio warning value, the warning level corresponding to the deviation ratio value is determined. The data elements in the null value sequence corresponding to the deviation ratio value are filled with the warning level.

6. The dynamic production planning management method based on a feedback mechanism according to claim 4, characterized in that, The specific steps for assessing the impact of the bottleneck resource node on the overall production schedule include: Each data element in the alert level sequence of the bottleneck resource node is represented as: ,in , The length of the alert level sequence for the bottleneck resource node; Calculate the impact of the bottleneck resource node on the overall production schedule based on the warning level sequence of the bottleneck resource node: 。 7. A dynamic production planning and management system based on a feedback mechanism, characterized in that, The system includes an early warning model building module for constructing a dynamic early warning model for production execution data, a bottleneck node identification module for using the dynamic early warning model to identify bottlenecks in the production execution data to determine whether there are bottleneck resource nodes in the production process, and a production planning management module. The dynamic early warning model consists of several resource nodes that reflect the node processing capacity and the warning range of the production execution status. The production planning management module is configured to evaluate the impact of the bottleneck resource nodes on the overall production plan scheduling. When the impact of the bottleneck resource nodes on the overall production plan scheduling is greater than a preset value, it generates scheduling anomaly data, generates production plan adjustment decision data based on the scheduling anomaly data, and performs cross-departmental feedback on the production plan adjustment decision data. The early warning model construction module is configured as follows: Construct a resource sequence to represent resource dimensions, where each data element in the resource sequence is a resource; Construct a process sequence to represent the time dimension, where each data element in the process sequence is a production and processing step; The resource dimension represented by the resource sequence is configured as the first orthogonal dimension, and the time dimension represented by the process sequence is configured as the second orthogonal dimension to generate an orthogonal data matrix framework. A standard data matrix is ​​generated based on the orthogonal data matrix framework, where each data element of the standard data matrix is ​​the standard resource data of the corresponding resource node; Obtain several pre-configured deviation ratio warning values; Configure the standard data matrix and the deviation ratio warning value as the basic data elements of the dynamic early warning model; After configuring the standard resource data and the deviation ratio warning value as the basic data elements of the dynamic early warning model, the early warning model construction module is configured as follows: Construct a lateral matching function for identifying lateral bottleneck nodes, a longitudinal matching function for identifying longitudinal bottleneck nodes, and a bottleneck node identification function for outputting bottleneck node identification results based on the lateral matching values ​​and longitudinal matching values ​​output by the lateral matching function and the longitudinal matching function; The horizontal matching function, the vertical matching function, and the bottleneck node identification function are configured as the dynamic warning function of the dynamic early warning model; The bottleneck node identification module is configured as follows: At the start of the production and processing process, a real-time data matrix of the current production and processing process is generated based on the orthogonal data matrix framework. The real-time data matrix is ​​a dynamic matrix whose size changes as the production and processing process progresses. Real-time monitoring of production execution data during the production and processing process; After each production and processing step is completed, the real-time data matrix is ​​updated based on the production execution data; The real-time data matrix is ​​input into the dynamic early warning model to perform bottleneck identification.

Images