An intelligent ranking method based on a multi-scale evaluation system

Abstract

The application discloses an intelligent sequencing method based on a multi-scale evaluation system, relates to the technical field of big data analysis, and constructs an evaluation system by extracting key data of a storage continuous production line scheduling problem, judges whether adjacent workpieces can be continuously produced through a multi-scale connection matrix, calculates production cost, calculates the priority on each dimension attribute through a multi-scale material priority value, constructs an MIP model, and calculates the scheduling sequence of each product.

Description

Technical Field

[0001] This invention relates to the field of big data analytics, specifically to an intelligent ranking method based on a multi-scale evaluation system. Background Technology

[0002] Currently, in the evaluation of many complex systems, traditional evaluation methods often only consider the impact of a single scale or dimension, leading to one-sided evaluation results that fail to fully reflect the performance or behavior of the target. This is especially true in the field of intelligent manufacturing, where production scheduling problems with multidimensional characteristics require comprehensive analysis and evaluation from different levels and scales due to the diversity and dynamism of the system.

[0003] A typical continuous production scheduling problem with multidimensional characteristics is described as follows: A manufacturing company needs to formulate a short-term production plan for a flexible production line. This production line can produce several different types of products, but the products have multiple dimensions of attributes. Specifically, changes in the attributes of products in various dimensions during continuous production will affect production. For example, changes in the geometric dimensions of adjacent products within a certain range have no significant impact on production, but exceeding a certain threshold requires replacing some molds on the production line to adapt, thus affecting capacity. Furthermore, situations where changes to certain product attributes require significant adjustments to the production line are generally not allowed in practice. In continuous production line scheduling problems, due to different processes, the requirements for the product processing sequence also differ, but are basically determined by the product attributes. Since each dimension of attribute has its unique switching rules and priorities, it is necessary to evaluate the impact of switching between product attributes on the quality of the scheduling results from multiple dimensions. However, the objectives of various dimensions are difficult to measure, and in practice, it relies heavily on the rich business experience of the scheduling personnel, lacking quantifiable evaluation indicators. Furthermore, in actual production, since orders arrive on a rolling basis, and these multi-dimensional product orders often feature multiple varieties and small batches, there is generally a need for batch scheduling of orders to reduce frequent production line adjustments. Therefore, a single scheduling session does not need to include all orders. The priority of order scheduling may also be affected by other order attributes, such as order delivery date and whether it is an urgent order, which greatly increases the complexity of scheduling decisions.

[0004] Currently, there is a lack of a systematic method that can effectively integrate multi-scale information to evaluate, model, and optimize complex problems that require multi-dimensional consideration. Therefore, a new evaluation system is needed that can integrate constraints and objective weights, transform constraints into model parameters or optimization objectives, comprehensively and accurately evaluate objectives at different scales, and obtain the optimal decision-making scheme based on this. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent ranking method based on a multi-scale evaluation system to solve the problems raised in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an intelligent ranking method based on a multi-scale evaluation system, wherein the intelligent management method includes the following steps:

[0007] Step S1: For the production operation plans of different products of manufacturing enterprises, the key basic data of continuous production line scheduling problems are divided into dynamic data and static data according to the data update frequency and uncertainty. The data is extracted and stored from multiple data sources.

[0008] Step S1-1: Data that is updated ≤ once a month and is determined to remain unchanged within a production cycle is classified as static data, and data that changes continuously over time and exhibits irregular changes is classified as dynamic data.

[0009] Step S1-2: Extract dynamic and static data from multiple data sources, and store the key basic data affecting production scheduling in the continuous production line scheduling problem into the database according to product-related attributes and order-related attributes. The key basic data are dynamic data and static data.

[0010] Step S2: Design a scoring framework system that includes multiple scoring mechanisms based on the production scheduling rules of the production line, uniformly process indicators of different dimensions and scales, and classify and integrate them into the product switching and product selection scoring items in continuous production scheduling.

[0011] Step S2-1: Construction of the evaluation system. Starting from the production scheduling rules, the decision points of the continuous production line scheduling problem with multi-dimensional constraints are divided into two parts: selection decision and sequence decision. The multi-dimensional constraints are multiple independent but interrelated restrictive conditions. The evaluation system includes product selection scoring items and product switching scoring items.

[0012] S2-2. The selection decision is affected by the product selection scoring items. The scoring is given according to the priority of each product attribute determined by the scoring mechanism of the scoring framework system. The subsequent scheduling plan arranges orders in descending order of the given scores. The scoring items do not limit the product attributes, but need to be adjusted in real time according to the actual scenario.

[0013] Step S2-3: Sequential decision-making is affected by product switching scoring items. Product switching is divided into continuous switching and discrete switching. Continuous switching is for product attributes with continuous values, while discrete switching is for switching enumerable attributes. Continuous switching can be converted into discrete switching by setting a value range.

[0014] Step S3: Use the multi-scale connectivity matrix to determine whether adjacent products in a product production chain can be produced continuously and calculate the cost of continuous production.

[0015] Step S3-1: The multi-scale connection matrix can quickly integrate the constraints and objectives in the continuous production line scheduling problem, reduce the complexity of subsequent modeling and calculation, and calculate both the non-connectable workpiece combination and the connectable workpiece combination.

[0016] Step S3-2: In production scheduling, changes to predefined key product attributes of two adjacent workpieces trigger production line reconfiguration or parameter adjustment. Such attribute switching is marked and prohibited from entering subsequent scheduling plans. During production scheduling, product combinations that are prohibited from switching are pre-established. The attributes of adjacent workpieces are checked item by item according to the preset attribute priority order. If a product combination that is prohibited from switching is detected, the connection relationship of the product combination is marked as unconnectable, and the connection relationship is prohibited from entering the subsequent scheduling judgment process. In production scheduling, changes to predefined key product attributes of two adjacent workpieces will not trigger production line reconfiguration or parameter adjustment. Such attribute switching is a connectable workpiece and will not enter the product combination that is prohibited from switching.

[0017] Step S3-3: For the attribute switching of multiple dimensions involved in connectable workpieces, different connection methods have different degrees of impact on production line adjustments. Attribute switching is divided into discrete switching and continuous fluctuation. The calculation of continuous fluctuation is as follows:

[0018] ;

[0019] In the formula, For continuous fluctuations, This represents the quantified value of the fluctuation of product i to j on attribute k. The base fluctuation score for attribute k;

[0020] Discrete wave calculation is specifically as follows:

[0021] ;

[0022] In the formula, It is a discrete wave. The pre-defined unique score for product i to j on attribute k;

[0023] Continuous fluctuations are applicable when the attribute value is a continuous numerical value. The scoring logic of discrete switching is that the difference between discrete switching and continuous fluctuations is that discrete switching is a finite set that can be enumerated. Switching of different enumerated values ​​of the same attribute uniquely corresponds to a score, marking the type that cannot be switched.

[0024] Step S4: Extract order attributes, calculate the priority value of each dimension of product attribute information using multi-scale material priority value, combine the calculated priority value with the multi-scale connection matrix to determine the production priority of each product order;

[0025] Step S4-1: In the order priority calculation rules, a preset benchmark score is used as the default score for order attributes with no selection preference. For product order attributes that are expected to be included in the plan, the priority is increased by 40% of the benchmark score. For product order attributes that are not expected to be included, the priority is decreased by 40% of the benchmark score. For orders that are decided not to be included in the schedule or orders that must be included, the priority calculation is skipped.

[0026] Step S4-2: The comprehensive multi-scale material priority value is the arithmetic sum of the priority values ​​of each attribute, specifically:

[0027] ;

[0028] In the formula, This represents the priority value for materials at multiple scales, where m is the total number of attributes involved in the evaluation. The fluctuation can be either continuous or discrete, depending on the attribute.

[0029] Step S5: Use the comprehensive target evaluation result as the objective function of the integer programming model to characterize the core constraints, and perform product production-level modeling and solution based on the continuous production line scheduling problem scenario;

[0030] Step S5-1: Calculate the total material connection score and total material priority value of the materials to be included, using the following formula:

[0031] ;

[0032] In the formula: I represents the set of all candidate materials, I = {0, 1, ..., N}, where 0 is the starting point, n is the ending point, j and i are the materials in the candidate material set, and c i,j To calculate the connection cost of producing material j after material i, x i,j To determine whether material j is produced immediately following material i, x i,j As an indicator variable, x i,j =1 indicates that material j is arranged after material i, x i,j =0 indicates that material i is not followed by material j, p i Let y be the material priority value of material i. i Whether material i is discharged, y i As an indicator variable, y i =1 indicates that it is included, y i=0 indicates that it will not be included, based on the sum of the total material connection score and the total material priority value;

[0033] Step S5-2: Calculate the intermediate material connection constraints excluding the starting and ending materials. Each selected material has one and only one preceding and following workpieces. The starting and ending materials can be divided into actual product orders and virtual orders according to actual needs. The formula is as follows:

[0034] ;

[0035] For any material in set I, if j is not in the preorder set of i... If j ≠ i, then x i,j The sum is 0;

[0036] ;

[0037] For any material in set I, if j is not in the subsequent set of i... If j ≠ i, then x i,j The sum is 0;

[0038] ;

[0039] For any material i other than the starting point 0, if j is not in the set of i... If j ≠ i, then x i,j The sum of these equals yi. If yi = 1, then the sum is 1, indicating that i has one and only one subsequent material.

[0040] ;

[0041] For any material i other than the endpoint n, if j is not in the subsequent set of i... If j ≠ i, then x i,j The sum equals y i If y i If the sum is 1, then i has one and only one preceding material;

[0042] Step S5-3: Ensure the last material is selected. The formula is:

[0043] ;

[0044] Step S5-4: Set the last workpiece to not discharge material. The formula is as follows:

[0045] ;

[0046] Step S5-5: Set there to be only one preceding workpiece before the last roll of material. The formula is as follows:

[0047] ;

[0048] Step S5-6, the calculation formula for eliminating sub-loop constraints is:

[0049] ;

[0050] Step S5-7: Set a formula to ensure that the material sorting number increases by 1. The formula is as follows:

[0051] ;

[0052] Step S5-8: Exclude connection methods that are not in the adjacency set. The calculation formula is as follows:

[0053] ;

[0054] Based on the above steps, the production sequence of each material can be determined.

[0055] Step S6: Adjust the parameters of the evaluation system to match the target of the influence of each factor in actual production. Use expert evaluation method to evaluate the production scheduling results. Based on the product production priority feedback from the evaluation, locate the corresponding influencing factor parameters in the evaluation system, and adjust the corresponding influencing factor parameters. If the product production priority feedback matches the actual scenario, output it; otherwise, continue to adjust the parameters and calculate.

[0056] Step S6-1: When the connection of two adjacent products in a specification attribute does not meet the production requirements, the score of this category is reduced and the priority of the connection is lowered. If the connection causes the connection quality of other attributes to decline compared with the connection quality of other product attributes, the switching rules are dynamically adjusted based on real-time feedback in production.

[0057] Step S6-2: For orders that are expected to be included in the production plan but are not, the core is to coordinate the relationship between the multi-scale connection matrix score and the material priority value. Under the premise of keeping the multi-scale material connection matrix unchanged, increase the material priority value of the orders that are expected to be included in the production plan, so that the orders that are expected to be included in the production plan are included in the production plan.

[0058] Step S6-3: When an error occurs in the connectability judgment in the multi-scale connection matrix, causing the scheduling work to be unable to proceed normally, it is necessary to check the connectability status and update the non-connectability status.

[0059] Compared with the prior art, the beneficial effects of the present invention are:

[0060] 1. This invention characterizes various types of constraints in the production scheduling problem by constructing a scoring framework system, and incorporates multidimensional constraints into the same evaluation system to facilitate subsequent model solving.

[0061] 2. This invention utilizes a multi-scale connection matrix calculation method to quickly integrate most of the constraints and objectives in the continuous production line scheduling problem, thereby reducing the complexity of subsequent modeling and calculation.

[0062] 3. This invention achieves optimal solution to this type of problem by constructing a MIP model that conforms to the process of continuous production line scheduling. Attached Figure Description

[0063] Figure 1 This is a schematic diagram illustrating the steps of an intelligent ranking method based on a multi-scale evaluation system according to the present invention.

[0064] Figure 2 This is a flowchart illustrating an intelligent ranking method based on a multi-scale evaluation system according to the present invention. Detailed Implementation

[0065] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] Step S1: For the production operation plans of different products of manufacturing enterprises, the key basic data of continuous production line scheduling problems are divided into dynamic data and static data according to the data update frequency and uncertainty. The data is extracted and stored from multiple data sources.

[0067] Step S1-1: Data that is updated ≤ once a month and is determined to remain unchanged within a production cycle is classified as static data, and data that changes continuously over time and exhibits irregular changes is classified as dynamic data.

[0068] Step S1-2: Extract dynamic and static data from multiple data sources, and store the key basic data affecting production scheduling in the continuous production line scheduling problem into the database according to product-related attributes and order-related attributes. The key basic data are dynamic data and static data.

[0069] Step S2: Design a scoring framework system that includes multiple scoring mechanisms based on the production scheduling rules of the production line, uniformly process indicators of different dimensions and scales, and classify and integrate them into the product switching and product selection scoring items in continuous production scheduling.

[0070] Step S2-1: Construction of the evaluation system. Starting from the production scheduling rules, the decision points of the continuous production line scheduling problem with multi-dimensional constraints are divided into two parts: selection decision and sequence decision. The multi-dimensional constraints are multiple independent but interrelated restrictive conditions. The evaluation system includes product selection scoring items and product switching scoring items.

[0071] S2-2. The selection decision is affected by the product selection scoring items. The scoring is given according to the priority of each product attribute determined by the scoring mechanism of the scoring framework system. The subsequent scheduling plan arranges orders in descending order of the given scores. The scoring items do not limit the product attributes, but need to be adjusted in real time according to the actual scenario.

[0072] Step S2-3: Sequential decision-making is affected by product switching scoring items. Product switching is divided into continuous switching and discrete switching. Continuous switching is for product attributes with continuous values, while discrete switching is for switching enumerable attributes. Continuous switching can be converted into discrete switching by setting a value range.

[0073] Step S3: Use the multi-scale connectivity matrix to determine whether adjacent products in a product production chain can be produced continuously and calculate the cost of continuous production.

[0074] Step S3-1: The multi-scale connection matrix can quickly integrate the constraints and objectives in the continuous production line scheduling problem, reduce the complexity of subsequent modeling and calculation, and calculate both the non-connectable workpiece combination and the connectable workpiece combination.

[0075] Step S3-2: In production scheduling, changes to predefined key product attributes of two adjacent workpieces trigger production line reconfiguration or parameter adjustment. Such attribute switching is marked and prohibited from entering subsequent scheduling plans. During production scheduling, product combinations that are prohibited from switching are pre-established. The attributes of adjacent workpieces are checked item by item according to the preset attribute priority order. If a product combination that is prohibited from switching is detected, the connection relationship of the product combination is marked as unconnectable, and the connection relationship is prohibited from entering the subsequent scheduling judgment process. In production scheduling, changes to predefined key product attributes of two adjacent workpieces will not trigger production line reconfiguration or parameter adjustment. Such attribute switching is a connectable workpiece and will not enter the product combination that is prohibited from switching.

[0076] Step S3-3: For the attribute switching of multiple dimensions involved in connectable workpieces, different connection methods have different degrees of impact on production line adjustments. Attribute switching is divided into discrete switching and continuous fluctuation. The calculation of continuous fluctuation is as follows:

[0077] ;

[0078] In the formula, For continuous fluctuations, This represents the quantified value of the fluctuation of product i to j on attribute k. The base fluctuation score for attribute k;

[0079] Discrete wave calculation is specifically as follows:

[0080] ;

[0081] In the formula, It is a discrete wave. The pre-defined unique score for product i to j on attribute k;

[0082] Continuous fluctuations are applicable when the attribute value is a continuous numerical value. The scoring logic of discrete switching is that the difference between discrete switching and continuous fluctuations is that discrete switching is a finite set that can be enumerated. Switching of different enumerated values ​​of the same attribute uniquely corresponds to a score, marking the type that cannot be switched.

[0083] Step S4: Extract order attributes, calculate the priority value of each dimension of product attribute information using multi-scale material priority value, combine the calculated priority value with the multi-scale connection matrix to determine the production priority of each product order;

[0084] Step S4-1: In the order priority calculation rules, a preset benchmark score is used as the default score for order attributes with no selection preference. For product order attributes that are expected to be included in the plan, the priority is increased by 40% of the benchmark score. For product order attributes that are not expected to be included, the priority is decreased by 40% of the benchmark score. For orders that are decided not to be included in the schedule or orders that must be included, the priority calculation is skipped.

[0085] Step S4-2: The comprehensive multi-scale material priority value is the arithmetic sum of the priority values ​​of each attribute, specifically:

[0086] ;

[0087] In the formula, This represents the priority value for materials at multiple scales, where m is the total number of attributes involved in the evaluation. The fluctuation can be either continuous or discrete, depending on the attribute.

[0088] Step S5: Use the comprehensive target evaluation result as the objective function of the integer programming model to characterize the core constraints, and perform product production-level modeling and solution based on the continuous production line scheduling problem scenario;

[0089] Step S5-1: Calculate the total material connection score and total material priority value of the materials to be included, using the following formula:

[0090] ;

[0091] In the formula: I represents the set of all candidate materials, I = {0, 1, ..., N}, where 0 is the starting point, n is the ending point, j and i are the materials in the candidate material set, and c i,j To calculate the connection cost of producing material j after material i, x i,j To determine whether material j is produced immediately following material i, x i,j As an indicator variable, x i,j =1 indicates that material j is arranged after material i, x i,j =0 indicates that material i is not followed by material j, p i Let y be the material priority value of material i. i Whether material i is discharged, y i As an indicator variable, y i =1 indicates that it is included, y i =0 indicates that it will not be included, based on the sum of the total material connection score and the total material priority value;

[0092] Step S5-2: Calculate the intermediate material connection constraints excluding the starting and ending materials. Each selected material has one and only one preceding and following workpieces. The starting and ending materials can be divided into actual product orders and virtual orders according to actual needs. The formula is as follows:

[0093] ;

[0094] For any material in set I, if j is not in the preorder set of i... If j ≠ i, then x i,j The sum is 0;

[0095] ;

[0096] For any material in set I, if j is not in the subsequent set of i... If j ≠ i, then x i,j The sum is 0;

[0097] ;

[0098] For any material i other than the starting point 0, if j is not in the set of i... If j ≠ i, then x i,j The sum of these equals yi. If yi = 1, then the sum is 1, indicating that i has one and only one subsequent material.

[0099] ;

[0100] For any material i other than the endpoint n, if j is not in the subsequent set of i... If j ≠ i, then x i,j The sum equals y i If y i If the sum is 1, then i has one and only one preceding material;

[0101] Step S5-3: Ensure the last material is selected. The formula is:

[0102] ;

[0103] Step S5-4: Set the last workpiece to not discharge material. The formula is as follows:

[0104] ;

[0105] Step S5-5: Set there to be only one preceding workpiece before the last roll of material. The formula is as follows:

[0106] ;

[0107] Step S5-6, the calculation formula for eliminating sub-loop constraints is:

[0108] ;

[0109] Step S5-7: Set a formula to ensure that the material sorting number increases by 1. The formula is as follows:

[0110] ;

[0111] Step S5-8: Exclude connection methods that are not in the adjacency set. The calculation formula is as follows:

[0112] ;

[0113] Based on the above steps, the production sequence of each material can be determined.

[0114] Step S6: Adjust the parameters of the evaluation system to match the target of the influence of each factor in actual production. Use expert evaluation method to evaluate the production scheduling results. Based on the product production priority feedback from the evaluation, locate the corresponding influencing factor parameters in the evaluation system, and adjust the corresponding influencing factor parameters. If the product production priority feedback matches the actual scenario, output it; otherwise, continue to adjust the parameters and calculate.

[0115] Step S6-1: When the connection of two adjacent products in a specification attribute does not meet the production requirements, the score of this category is reduced and the priority of the connection is lowered. If the connection causes the connection quality of other attributes to decline compared with the connection quality of other product attributes, the switching rules are dynamically adjusted based on real-time feedback in production.

[0116] Step S6-2: For orders that are expected to be included in the production plan but are not, the core is to coordinate the relationship between the multi-scale connection matrix score and the material priority value. Under the premise of keeping the multi-scale material connection matrix unchanged, increase the material priority value of the orders that are expected to be included in the production plan, so that the orders that are expected to be included in the production plan are included in the production plan.

[0117] Step S6-3: When an error occurs in the connectability judgment in the multi-scale connection matrix, causing the scheduling work to be unable to proceed normally, it is necessary to check the connectability status and update the non-connectability status.

[0118] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. An intelligent ranking method based on a multi-scale evaluation system, characterized in that: The intelligent sorting method includes the following steps: Step S1: For the production operation plans of different products of manufacturing enterprises, the key basic data of continuous production line scheduling problems are divided into dynamic data and static data according to the data update frequency and uncertainty. The data is extracted and stored from multiple data sources. Step S2: Design a scoring framework system that includes multiple scoring mechanisms based on the production line scheduling rules, uniformly handle indicators of different dimensions, and classify and integrate them into the product switching and product selection scoring items in continuous production scheduling; the specific steps of step S2 are as follows: Step S2-1: Construction of the evaluation system. Starting from the production scheduling rules, the decision points of the continuous production line scheduling problem with multi-dimensional constraints are divided into two parts: selection decision and sequence decision. The multi-dimensional constraints are multiple independent but interrelated restrictive conditions. The evaluation system includes product selection scoring items and product switching scoring items. S2-2. The selection decision is affected by the product selection scoring items. The scoring is given according to the priority of each product attribute determined by the scoring mechanism of the scoring framework system. The subsequent scheduling plan arranges orders in descending order of the given scores. The scoring items do not limit the product attributes, but need to be adjusted in real time according to the actual scenario. Step S2-3: Sequential decision-making is affected by product switching scoring items. Product switching is divided into continuous switching and discrete switching. Continuous switching is for product attributes with continuous values, while discrete switching is for switching enumerable attributes. Continuous switching can be converted into discrete switching by setting a value range. Step S3: Use a multi-scale connectivity matrix to determine whether adjacent products in a product production chain can be produced continuously and calculate the cost of continuous production; the specific steps of step S3 are as follows: Step S3-1: The multi-scale connection matrix can quickly integrate the constraints and objectives in the continuous production line scheduling problem, reduce the complexity of subsequent modeling and calculation, and calculate both the non-connectable workpiece combination and the connectable workpiece combination. Step S3-2: In production scheduling, changes to predefined key product attributes of two adjacent workpieces trigger production line reconfiguration or parameter adjustment. Such attribute switching is marked and prohibited from entering subsequent scheduling plans. During production scheduling, product combinations that are prohibited from switching are pre-established. The attributes of adjacent workpieces are checked item by item according to the preset attribute priority order. If a product combination that is prohibited from switching is detected, the connection relationship of the product combination is marked as unconnectable, and the connection relationship is prohibited from entering the subsequent scheduling judgment process. In production scheduling, changes to predefined key product attributes of two adjacent workpieces will not trigger production line reconfiguration or parameter adjustment. Such attribute switching is a connectable workpiece and will not enter the product combination that is prohibited from switching. Step S3-3: For the attribute switching of multiple dimensions involved in connectable workpieces, different connection methods have different degrees of impact on production line adjustments. Attribute switching is divided into discrete switching and continuous fluctuation. The calculation of continuous fluctuation is as follows: ； In the formula, For continuous fluctuations, This represents the quantified value of the fluctuation of product i to j on attribute k. The base fluctuation score for attribute k; Discrete wave calculation is specifically as follows: ； In the formula, It is a discrete fluctuation. The pre-defined unique score for product i to j on attribute k; Continuous fluctuations are applicable when the attribute value is a continuous numerical value. The scoring logic of discrete switching is that the difference between discrete switching and continuous fluctuations is that discrete switching is a finite set that can be enumerated. Switching different enumerated values ​​of the same attribute uniquely corresponds to a score, marking the type that cannot be switched. Step S4: Extract order attributes, calculate the priority values ​​of product attribute information in each dimension using multi-scale material priority values, and combine the calculated priority values ​​with the multi-scale linkage matrix to determine the production priority of each product order; the specific steps of step S4 are as follows: Step S4-1: In the order priority calculation rules, a preset benchmark score is used as the default score for order attributes with no selection preference. For product order attributes that are expected to be included in the plan, the priority is increased by 40% of the benchmark score. For product order attributes that are not expected to be included, the priority is decreased by 40% of the benchmark score. For orders that are decided not to be included in the schedule or orders that must be included, the priority calculation is skipped. Step S4-2: The comprehensive multi-scale material priority value is the arithmetic sum of the priority values ​​of each attribute, specifically: ； In the formula, This represents the priority value for materials at multiple scales, where m is the total number of attributes involved in the evaluation. The fluctuation can be continuous or discrete depending on the attribute. Step S5: Use the comprehensive objective evaluation result as the objective function of the integer programming model to characterize the core constraints, and perform product production-level modeling and solution based on the continuous production line scheduling problem scenario; the specific steps of step S5 are as follows: Step S5-1: Calculate the total material connection score and total material priority value of the materials to be included, using the following formula: ； In the formula: I represents the set of all candidate materials, I = {0, 1, ..., N}, where 0 is the starting point, n is the ending point, j and i are the materials in the candidate material set, and c i,j To calculate the connection cost of producing material j after material i, x i,j To determine whether material j is produced immediately following material i, x i,j As an indicator variable, x i,j =1 indicates that material j is arranged after material i, x i,j =0 indicates that material i is not followed by material j, p i Let y be the material priority value of material i. i Whether material i is discharged, y i As an indicator variable, y i =1 indicates that it is included, y i =0 indicates that it will not be included, based on the sum of the total material connection score and the total material priority value; Step S5-2: Calculate the intermediate material connection constraints excluding the starting and ending materials. Each selected material has one and only one preceding and following workpieces. The starting and ending materials can be divided into actual product orders and virtual orders according to actual needs. The formula is as follows: ； For any material in set I, if j is not in the preorder set of i... If j ≠ i, then x i,j The sum is 0; ； For any material in set I, if j is not in the subsequent set of i... If j ≠ i, then x i,j The sum is 0; ； For any material i other than the starting point 0, if j is not in the set of i... If j ≠ i, then x i,j The sum of these equals yi. If yi = 1, then the sum is 1, indicating that i has one and only one subsequent material. ； For any material i other than the endpoint n, if j is not in the subsequent set of i... If j ≠ i, then x i,j The sum equals y i If y i If the sum is 1, then i has one and only one preceding material; Step S5-3: Ensure the last material is selected. The formula is: ； Step S5-4: Set the last workpiece to not discharge material. The formula is as follows: ； Step S5-5: Set there to be only one preceding workpiece before the last roll of material. The formula is as follows: ； Step S5-6, the calculation formula for eliminating sub-loop constraints is: ； Step S5-7: Set a formula to ensure that the material sorting number increases by 1. The formula is as follows: ； Step S5-8: Exclude connection methods that are not in the adjacency set. The calculation formula is as follows: ； Based on the above steps, the production sequence of each material can be determined; Step S6: Adjust the parameters of the evaluation system to match the target of the influence of each factor in actual production. Use expert evaluation method to evaluate the production scheduling results. Based on the product production priority feedback from the evaluation, locate the corresponding influencing factor parameters in the evaluation system, and adjust the corresponding influencing factor parameters. If the product production priority feedback matches the actual scenario, output it; otherwise, continue to adjust the parameters and calculate.

2. The intelligent ranking method based on a multi-scale evaluation system according to claim 1, characterized in that: The specific steps of step S1 are as follows: Step S1-1: Data that is updated ≤ once a month and is determined to remain unchanged within a production cycle is classified as static data, and data that changes continuously over time and exhibits irregular changes is classified as dynamic data. Step S1-2: Extract dynamic and static data from multiple data sources, and store the key basic data affecting production scheduling in the continuous production line scheduling problem into the database according to product-related attributes and order-related attributes. The key basic data are dynamic data and static data.

3. The intelligent ranking method based on a multi-scale evaluation system according to claim 1, characterized in that: The specific steps for resolving the issue of poor specification alignment between adjacent workpieces in step S6 are as follows: Step S6-1: When the connection of two adjacent products in a specification attribute does not meet the production requirements, the score of this category is reduced and the priority of the connection is lowered. If the connection causes the connection quality of other attributes to decline compared with the connection quality of other product attributes, the switching rules are dynamically adjusted based on real-time feedback in production.

4. The intelligent ranking method based on a multi-scale evaluation system according to claim 3, characterized in that: The specific steps for step S6, which involves expecting orders not to be scheduled, are as follows: Step S6-2: For orders that are expected to be included in the production plan but are not, the key is to coordinate the relationship between the multi-scale connection matrix score and the material priority value. Under the premise of keeping the multi-scale material connection matrix unchanged, increase the material priority value of the orders that are expected to be included in the production plan, so that the orders that are expected to be included in the production plan are included in the production plan.

5. The intelligent ranking method based on a multi-scale evaluation system according to claim 4, specifically addressing the connection judgment error in the multi-scale connection matrix, comprises the following steps: Step S6-3: When an error occurs in the connectability judgment in the multi-scale connection matrix, causing the scheduling work to be unable to proceed normally, it is necessary to check the connectability status and update the non-connectability status.

Images