Automobile order optimization production scheduling method based on APS

Abstract

The invention belongs to the technical field of production scheduling, and particularly relates to an APS-based automobile order optimization production scheduling method, which specifically comprises the following steps of: 1, optimizing models of an AP part and an AS part in an APS according to actual demands of enterprise orders; 2, in the AP part, the model takes the minimum constraint dissatisfaction as a target, the satisfaction of the high priority constraint is further emphasized on the basis of the priority weight, and in the algorithm, the constraint combination constraint scene is satisfied through the first round of rough arrangement, and then the constraint is further satisfied through the second round of fine arrangement; 3, in the AS part, different fitness functions are used in a single-constraint scene and a multi-constraint scene for solving.

Description

Technical Field

[0001] This invention belongs to the field of production scheduling technology, specifically relating to an APS-based method for optimizing automobile order scheduling. Background Technology

[0002] Currently in the manufacturing industry, in order to meet the matching degree between market orders and enterprise capacity, reduce inventory and improve capital turnover, enterprises are gradually starting to schedule production based on order volume. Under various constraints, they can obtain shorter product manufacturing cycles and lower production costs. The production scheduling problem usually involves two types of constraints: hard constraints and soft constraints. Hard constraints must be fully satisfied in order for the generated solution to be feasible, while soft constraints can be satisfied as much as possible.

[0003] This invention aims to provide a faster and more efficient order processing and scheduling method to overcome the problems of long solution time and low constraint satisfaction rate of existing methods. Specifically, this invention significantly improves scheduling efficiency through a superior new mathematical model and a more applicable new algorithm, making it superior to traditional methods in terms of solution speed and constraint satisfaction rate. Summary of the Invention

[0004] The purpose of this invention is to provide an APS-based method for optimizing automobile order scheduling to solve the problems mentioned in the background art.

[0005] To achieve the above-mentioned technical objectives, the technical solution adopted by the present invention is as follows: An APS-based method for optimizing automobile order scheduling, characterized by the following specific method: Step 1: Based on the actual needs of the enterprise's orders, optimize the models of the AP and AS parts in APS respectively; Step 2: In the AP part, the model aims to minimize the number of unmet constraints, and further emphasizes the satisfaction of high-priority constraints in the priority weight. In terms of algorithm, most constraints are satisfied through the first round of coarse ranking, while for combined constraint scenarios, the second round of fine ranking is needed to further satisfy the constraints. Step 3: In the AS part, different fitness functions are used to solve the problem in both single-constraint and multi-constraint scenarios.

[0006] Construct the objective function of AP and set the constraints, and then perform algorithm processing.

[0007] Construct a function model for AS, then use a genetic algorithm for coarse ranking, and finally perform fine ranking.

[0008] This application has at least the following advantages compared to the prior art: (1) In the AP (order grouping) stage, feasible solutions with the same or even higher constraint satisfaction rate can be obtained more quickly. In the actual production process, the core function of the AP part is to reasonably group all production orders to ensure that each group of orders can achieve stable and efficient production under the premise of meeting the factory's production capacity limit. Compared with the existing technical means, this invention improves the solution efficiency and constraint satisfaction level, making the scheduling process more efficient and convenient, and significantly improving the response speed and practicality of the overall scheduling system.

[0009] (2) In the AS (order sorting) stage, an optimization solution with a higher constraint satisfaction rate can be obtained. The main task of the AS part is to perform sequence optimization on daily orders so that they can reach a better state in terms of production process requirements, production cost control and production stability. A higher constraint satisfaction rate means that the production arrangement can better meet the actual operation needs, which helps to achieve the goal of cost reduction and efficiency improvement, thereby enhancing the company's overall competitiveness.

[0010] (3) Overall, this invention achieves effective collaboration and integrated optimization between AP and AS modules. While ensuring a high constraint satisfaction rate, it also takes into account the solution efficiency and solution quality. It is suitable for complex and ever-changing production environments and has good engineering application prospects and promotion value. Attached Figure Description

[0011] The present invention can be further illustrated by the non-limiting embodiments given in the accompanying drawings.

[0012] Figure 1 This is a schematic diagram of the overall method steps of the present invention.

[0013] Figure 2 This is a schematic diagram of the AP scheduling process of the present invention.

[0014] Figure 3 This is a schematic diagram of the AP coarse sorting process of the present invention.

[0015] Figure 4 This is a schematic diagram of the AP fine sorting process of the present invention.

[0016] Figure 5 This is a schematic diagram of the AP frame structure of the present invention.

[0017] Figure 6 This is a schematic diagram of the coarse sorting process of the AS algorithm of the present invention.

[0018] Figure 7 This is a schematic diagram demonstrating two cross-operations of the present invention.

[0019] Figure 8 This is a schematic diagram of a coarse-row embodiment of the present invention.

[0020] Figure 9This is a schematic diagram comparing AP combination scenarios of the present invention.

[0021] Figure 10 This is a schematic diagram comparing the AS combination scenarios of the present invention. Detailed Implementation

[0022] To enable those skilled in the art to better understand the present invention, the technical solution of the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0023] like Figure 1-10 As shown, an APS-based method for optimizing automobile order scheduling is described below: Step 1: Based on the actual needs of the enterprise's orders, optimize the models of the AP and AS parts in APS respectively; Step 2: In the AP part, the model aims to minimize the number of unmet constraints, and further emphasizes the satisfaction of high-priority constraints in the priority weight. In terms of algorithm, most constraints are satisfied through the first round of coarse ranking, while for combined constraint scenarios, the second round of fine ranking is needed to further satisfy the constraints. Step 3: In the AS part, different fitness functions are used to solve the problem in both single-constraint and multi-constraint scenarios.

[0024] Specifically, in the new model and algorithm design for the AP and AS parts in this application, the AP part aims to minimize the number of unmet constraints in the model and further emphasizes the satisfaction of high-priority constraints in the priority weight. In terms of algorithm, most constraints are satisfied through the first round of coarse sorting. Especially in single-constraint scheduling, a scheduling result with a 100% constraint satisfaction rate can be obtained in a short time through coarse sorting. For combined constraint scenarios, a second round of fine sorting is required to further satisfy the constraints.

[0025] For the AS part, the corresponding objective function is directly used as the fitness function for single-constraint scenarios, and the genetic algorithm is used to solve it directly, which can obtain the optimal solution in a short time. For multi-constraint scenarios, the corresponding objective functions are linearly weighted, transforming the multi-objective optimization problem into a single-objective optimization problem. The relative magnitude of the weights reflects the priority of the corresponding constraints. During the solution process, the data orders are encoded as integers, and an improved sequential crossover operation and multiple neighborhood search strategies are used to improve the mutation operation in the traditional genetic algorithm. The improved genetic algorithm is used to perform a coarse ranking of the combined scenario orders. After obtaining an approximate optimal solution, a second round of fine ranking is required based on the satisfaction conditions.

[0026] In step two, the objective function is constructed as follows: Formula 1:

[0027] in 0-1 variables represent the first... The constraint in the first The satisfaction level of the day, Indicates the first The priority weight values ​​corresponding to each constraint are calculated as follows: Formula 2:

[0028] Formula 3:

[0029] in, Indicates the first Constraint priority of each constraint.

[0030] The constraints on the objective function in step two are as follows: Formula 4:

[0031] Formula 5:

[0032] Formula 6:

[0033] Formula 7:

[0034] Formula 8:

[0035] Formula 9:

[0036] Formula 10:

[0037] Equation 4 represents the daily production capacity constraint, the first... Daily production volume No more than daily production capacity ; Equation 5 represents the production range constraint, the first... Daily production schedule Column attribute value The quantity satisfies the interval ; Equation 6 indicates that the constraint is included, where This is an indicator function; it returns 1 if the condition is true and 0 if it is false, i.e., at the th... Group 1 Column attribute value The quantity is At that time, in this There is a line The Column attribute value ; Equation 7 represents a mutual exclusion constraint, the first... Group 1 Column attribute value If the row number is not 0, then the attribute value is The number of rows must be 0; Equation 8 represents the upper limit constraint on the number of categories. Group 1 The number of numerical categories in a column is less than the upper limit of the number of categories. ; Equation 9 represents the equilibrium constraint, the first... Group 1 Column attribute value actual quantity Compared with the average quantity The difference is less than the upper limit of the equilibrium difference. ; Equation 10 represents a specified date constraint, for the first... For rows where the column value is not 0, the group number of the group to which it belongs. It must be equal to the first Column values .

[0038] The specific algorithm steps in step two are as follows: Step a: Data preprocessing, including: Production scheduling information processing: Read the corresponding fields of production scheduling information from the production scheduling data and build a grouping matrix according to daily production capacity; Order data processing: Read order information from production scheduling data and build a two-dimensional matrix to store the order information; Constraint processing: Read production scheduling constraint information from the production scheduling data and establish a constraint information matrix; Step b: Sorting, including: Constraint Priority Ranking: Prioritize the constraints considered in production scheduling; The first sort is a coarse sort: In the first round of sorting, the algorithm will fill the grouping matrix with the orders from the initial matrix according to the constraint information matrix, and the filling order will follow three priority principles: Constraints with higher priority take precedence: among constraints of the same level, the order of priority is: production range constraint > inclusion constraint > production ramp-up constraint > category upper limit constraint > mutual exclusion constraint > equilibrium constraint. Prioritize those with fewer attributes for the same constraint: the first round of sorting can achieve a single constraint fully satisfied and a combined scenario constraint satisfaction rate of over 69%; For example Figure 8In some embodiments, matrix A represents the initial order evidence. Column 1 represents the order number, and columns 2 to 4 represent the order attributes. First, a grouping matrix B is established based on the daily capacity constraints [{Day: 1,capacity": 3},{Day: 2,capacity: 3}]. Then, the production scheduling result is obtained based on the constraint information. Taking the production range constraint [2: [2],”1”:[1,1],”2”:[3,3]] as an example, the constraint requires column index 2 (i.e., the third column of the matrix) to have 1 unit on the first day and 3 units on the second day. Then, the remaining unconstrained orders are used to fill in the daily capacity, and the result matrix C is obtained.

[0039] The second sorting is fine sorting: it further satisfies the constraints that were not satisfied in the coarse sorting. In the first round of sorting, the algorithm fills the orders in the initial matrix into the grouping matrix according to the constraint information matrix. However, in the production scheduling of combined scenarios, due to the mutual influence between constraints, especially the situation of constraint conflict, a large number of constraints are not satisfied. At this time, it is necessary to repair these unsatisfied constraints through subsequent adjustments.

[0040] 1: Quantity Coding ([column, value, daily quantity]) [1, 1, [30, 30], [30, 30], [30, 30], [50, 60]] −−− Number of constraints required [1, 1, 20, 30, 40, 55]−−−−−−−−−−−Current Quantity [1, 1, −10, 0, 10, [−5, 5]]−−−Difference between current quantity and constraint For example, in the above situation, 10 orders with a value of 1 in column 2 from group 3 will be selected and exchanged with 10 orders with a value of 1 or less from group 1. The value [-5,5] in group 4 indicates that the constraint for group 4 requires an interval value, and currently 5 constraint orders can be reduced or accepted. 2: Type Encoding ([column, number of day types (mutual exclusions), constraint effective date]) [1, 3, [1, 2, 3]]−−−−−−−−−−−−Number of constraints required [1, 4, 3, 3]−−−−−−−−−−−−−−−Current Quantity [[1,2,3,4],[1,2,3],[2,3,4]]−−−−−−Difference between current quantity and constraint For example, in the above situation, the category with the fewest number of the four attributes in the first group will be selected and swapped to the second or third group. For example, if category 1 is swapped, it can only be swapped to the second group; while if categories 2 or 3 are swapped, they can be swapped to either the second or third group.

[0041] 3: Balanced Quantity Coding ([column, value, total quantity, production date]) is similar to quantity coding. The constraints need to be converted first. The subsequent exchange method is the same as that of quantity coding, but it is easy to break other constraints.

[0042] [1, 3, 30, 3]−−−−−−−−−−−−Number of constraints required [1, 3, 10, 10, 10]−−−−−−−−−−−−−Constraint Transformation The balance constraint is the last constraint to be satisfied. After balance repair, the algorithm will first check whether the exchange result is better. If not, it will select an order to exchange again.

[0043] The function model in step three is as follows: Formula 11:

[0044] Formula 12:

[0045] Formula 13:

[0046] Formula 14:

[0047] Formula 15:

[0048] Formula 16:

[0049] Equation 17:

[0050] Formula 18:

[0051] Formula 19:

[0052] Formula 20:

[0053] Equation 21:

[0054] Equation 22:

[0055] Equation 23:

[0056] Formula 24:

[0057] Equation 11 is the fitness function, representing the case where the minimization constraint is not satisfied. For the objective function The weights; Equation 12 represents the switching of upper limit constraints, with the daily production scheduling attribute value being... The number of switching times is less than a specified value, where ; Equation 13 represents a continuity constraint, and the attribute value is... The length of the block must be within the specified range. ,in Indicates the first Each block attribute is The length of the cluster, Indicates the attribute is The optimal block length; Equation 14 represents the interval constraint, and the attribute value is There must be at least a gap between the clusters. Other attribute values ​​of the platform, among which Indicates that the attribute value is not equal to , Indicates the first The node attribute is not The length of the cluster; Equation 15 represents the adjacency constraint, and the attribute value is The following non-adjacent production scheduling attribute values ,in

[0058] Equation 16 indicates that the constraint is included, specifying Each order range must include indivual ,in, Indicates the first The attribute value within each interval is Quantity; Equation 17 indicates that the batch contains each attribute value. The block, whose corresponding interval sequence contains at least Platform attribute value ,in Indicates the first Each block attribute is The attribute value in the corresponding interval sequence is Quantity; Equation 18 represents the number of attribute batches for attribute values. The number of its blocks needs to be . There are 1, and the length difference of each block is the current minimum, where Represents attribute value The actual number of blocks, Indicates the penalty value. Indicates the first Each block attribute is The length of the cluster, This means that all attribute values ​​are Quantity; Equation 19 represents sorting by integer multiples for attribute values. Its clumping requires Production is scheduled in multiples of integers, with the remainder being... Production is scheduled separately, with the remainder When merged into any attribute value In the clumps, among which Indicates the first Each block attribute is The length of the cluster; Equation 20 represents the mixed-row constraint, which applies to attribute values. , Mixed production is not allowed, and the specified production order is that the first production run is completed first. , then arrange ,in This means that all attribute values ​​are The position of the first order in the order list. This means that all attribute values ​​are The position of the last order in the order book; Equation 21 represents the equilibrium constraint within the cycle time, assuming the workshop production cycle time is... JPH, or 1-hour production Trolley, for attribute values ,Every The number of tables must be equal, among which Indicates each The attribute value inside the trolley is Quantity, Indicates the total number of orders; Equation 22 represents the priority ordering for attribute values. Priority is given to the first vehicle on that day; this constraint only applies to the first vehicle. Equation 23 represents the specified position constraint, which refers to the first few order numbers in the initial order list. The corresponding orders need to be scheduled for production at the designated location. ; Equation 24 represents the buffer constraint, which means that the sequence difference of the same order before and after the change from the welding workshop through the buffer to the final assembly workshop must not exceed a specified value. ,in This indicates the number of units in the final assembly workshop's production scheduling results. The location of each order. This indicates the number of units in the final assembly workshop's production scheduling results. The position of each order in the production schedule of the welding workshop.

[0059] The algorithm in step three uses a genetic algorithm for coarse ranking, and the specific method is as follows: a. Population initialization: Specifically, the initial population is generated by combining random generation with heuristic rules to improve population diversity. Therefore, two methods for initializing the population are set: random generation and generation based on constraints.

[0060] Among them, constraint-based generation refers to obtaining a production schedule result that fully satisfies a certain constraint through data processing, without considering the satisfaction of other constraints. For each constraint in the combined scenario, such a production schedule result is generated and used as an individual in the initial population for subsequent operations. b. Perform selection evaluation. Specifically, use the truncated selection method to evaluate the fitness of all individuals in the current population, obtain the fitness value of each individual, sort the population according to the fitness value from high to low, and then select the individual with the highest fitness value. [Individuals] are used for subsequent crossover operations; c. Perform crossover operation. Specifically, the crossover operation further explores the solution space by exchanging the gene information of selected chromosomes. In this invention, we improve upon the traditional PMX crossover to obtain two crossover operations. Both crossover operations are essentially the same: select a gene or gene fragment from a parent individual 1, then directly copy the selected gene to offspring 1, and fill the missing gene in offspring 1 in the order it appears in parent 2 to generate offspring individual 1. d. Insertion mutation operation. Specifically, in this invention, we use insertion mutation as the mutation operation. The specific steps are to randomly select a specified number of genes and insert them into random positions in the chromosome. Since our goal is to make all genes in the chromosome change according to the specified rules, we hope that other genes can be adjusted in the expected direction without affecting the order of local gene fragments. Therefore, insertion mutation is the optimal choice. e. Introduce heuristic rules. Specifically, (1) For orders that need to be scheduled to a specific location, we prioritize precise processing and schedule these orders to the specified location before the algorithm starts iterating. In subsequent iterations, we do not make any position changes to these orders.

[0061] (2) In actual automobile production scheduling, the buffer constraint is a crucial restriction. It requires that the sequence difference of each car in different workshops must be strictly controlled within a specified range. If at least one car exceeds this range, it will cause buffer congestion, and the scheduling scheme is considered invalid. Since it is necessary to ensure that every order satisfies the buffer constraint, randomly selecting the insertion position during mutation results in extremely slow convergence in the later stages, as verified. Therefore, we optimized the insertion position of the mutated gene, as follows: After selecting the mutated gene, we located its position in the production scheduling results of the welding workshop. Then the insertion location range of this gene is: ,in The specified sequence difference value.

[0062] The algorithm in step three obtains the optimal scheduling result after coarse sorting, and then performs fine sorting based on this result, adjusting any unmet constraints. Generally speaking, only continuous constraints may be unmet; other constraints can achieve a 100% satisfaction rate after coarse sorting, so adjustments are only made for continuous constraints.

[0063] To address the continuity constraint problem in production scheduling (i.e., the length of material clumps with a specific attribute value must satisfy a given interval), the coarse-sorting algorithm may generate invalid clumps with lengths below the lower limit. Therefore, the fine-sorting stage employs a neighborhood borrowing strategy for correction, the core logic of which is as follows: a. Perform constraint repair, specifically, regarding length The cluster, starting from the nearest satisfied Borrowing from similar clusters The number of orders makes the current block length after correction... The length of the borrowed block still meets the requirements. ; b. Perform two-way verification. Specifically, the borrowing operation needs to simultaneously verify the constraint satisfaction of both blocks. Recursive checks are used to ensure that the borrowed block will not generate new violations due to the borrowing.

[0064] This approach has two advantages: first, spatial locality, which prioritizes borrowing from adjacent nodes to minimize changes in the order sequence; and second, completeness, which means that the vast majority of nodes will ultimately satisfy the requirements. .

[0065] Specifically, in some embodiments, taking actual order data provided by an automobile manufacturer as an example, three combined scenarios were verified in both AP and AS scenarios. First, the hard constraints of the orders (such as daily production capacity constraints in AP and specified location constraints in AS) were prioritized for satisfaction; Secondly, the model is constructed with the goal of minimizing constraint violations; Then, an improved method is used to solve it.

[0066] like Figure 9-10 The figure shows a comparison of the constraint satisfaction rate and solution time of the present invention with traditional methods in combined scenarios.

[0067] Therefore, compared with some existing methods, the AP and AS parts of this method are partially or completely superior to existing methods in terms of constraint satisfaction rate and solution time.

[0068] Specifically, this method can obtain solutions with similar constraint satisfaction rates more quickly in the AP part. In scenario 1, the constraint satisfaction rate is only 2.1% lower than that of the existing method, but the solution time is only 18% of that of the existing method; in scenario 2, the constraint satisfaction rate is only 2.5% lower than that of the existing method, but the solution time is only 3.1% of that of the existing method; in scenario 3, the constraint satisfaction rate is 2.67% higher than that of the existing method, and the solution time is only 1% of that of the existing method.

[0069] Higher constraint satisfaction rates were achieved in all three scenarios of the AS section. In scenario one, the constraint satisfaction rate was only 2.96% lower than the existing method, but the solution time was only 5% of the existing method; in scenario two, the constraint satisfaction rate was 2.19% higher than the existing method, but the solution time was only 14% longer; in scenario three, the constraint satisfaction rate was only 4.29% lower than the existing method, but the solution time was only 50% longer.

[0070] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A car order optimization scheduling method based on APS, characterized in that: The specific method is as follows: Step 1: Based on the actual needs of the enterprise's orders, optimize the models of the AP and AS parts in APS respectively; Step 2: In the AP part, the model aims to minimize the number of unmet constraints, and further emphasizes the satisfaction of high-priority constraints in the priority weight. In terms of algorithm, most constraints are satisfied through the first round of coarse ranking, while for combined constraint scenarios, the second round of fine ranking is needed to further satisfy the constraints. Step 3: In the AS part, different fitness functions are used to solve the problem in both single-constraint and multi-constraint scenarios.

2. The method for optimizing automobile order scheduling based on APS according to claim 1, characterized in that: In step two, the objective function is constructed as follows: Formula 1: ; in 0-1 variables represent the first... The constraint in the first The satisfaction level of the day, Indicates the first The priority weight values ​​corresponding to each constraint are calculated as follows: Formula 2: ; Formula 3: ; in, Indicates the first Constraint priority of each constraint.

3. The automotive order optimization scheduling method based on APS according to claim 2, characterized in that: The constraints on the objective function in step two are as follows: Formula 4: ; Formula 5: ; Formula 6: ; Formula 7: ; Formula 8: ; Formula 9: ; Formula 10: ; Equation 4 represents the daily production capacity constraint, the first... Daily production volume No more than daily production capacity ; Equation 5 represents the production range constraint, the first... Daily production schedule Column attribute value The quantity satisfies the interval ; Equation 6 indicates that the constraint is included, where This is an indicator function; it returns 1 if the condition is true and 0 if it is false, i.e., at the th... Group 1 Column attribute value The quantity is At that time, in this There is a line The Column attribute value ; Equation 7 represents a mutual exclusion constraint, the first... Group 1 Column attribute value If the row number is not 0, then the attribute value is The number of rows must be 0; Equation 8 represents the upper limit constraint on the number of categories. Group 1 The number of categories in a column is less than the upper limit of the number of categories. ; Equation 9 represents the equilibrium constraint, the first... Group 1 Column attribute value actual quantity Compared with the average quantity The difference is less than the upper limit of the equilibrium difference. ; Equation 10 represents a specified date constraint, for the first... For rows where the column value is not 0, the group number of the group to which it belongs. It must be equal to the first Column values .

4. The automotive order optimization scheduling method based on APS according to claim 3, characterized in that: The specific algorithm steps in step two are as follows: Step a: Data preprocessing, including: Production scheduling information processing: Read the corresponding fields of production scheduling information from the production scheduling data and build a grouping matrix according to daily production capacity; Order data processing: Read order information from production scheduling data and build a two-dimensional matrix to store the order information; Constraint processing: Read production scheduling constraint information from the production scheduling data and establish a constraint information matrix; Step b: Sorting, including: Constraint Priority Ranking: Prioritize the constraints considered in production scheduling; The first sort is a coarse sort: In the first round of sorting, the algorithm will fill the grouping matrix with the orders from the initial matrix according to the constraint information matrix, and the filling order will follow three priority principles: Constraints with higher priority take precedence: among constraints of the same level, the order of priority is: production range constraint > inclusion constraint > production ramp-up constraint > category upper limit constraint > mutual exclusion constraint > equilibrium constraint. For multiple attributes with the same constraint, those with fewer attributes are preferred; The second sorting is fine sorting: it further satisfies the constraints that were not satisfied in the coarse sorting. In the first round of sorting, the algorithm fills the orders in the initial matrix into the grouping matrix according to the constraint information matrix. However, in the production scheduling of combined scenarios, due to the mutual influence between constraints, especially the situation of constraint conflict, a large number of constraints are not satisfied. At this time, it is necessary to repair these unsatisfied constraints through subsequent adjustments.

5. The automotive order optimization scheduling method based on APS according to claim 1, characterized in that: The function model in step three is as follows: Formula 11: ; Formula 12: ; Formula 13: ; Formula 14: ; Formula 15: ; Formula 16: ; Equation 17: ; Formula 18: ; Formula 19: ; Formula 20: ; Equation 21: ; Equation 22: ; Equation 23: ; Formula 24: ; Equation 11 is the fitness function, representing the case where the minimization constraint is not satisfied. For the objective function The weights; Equation 12 represents the switching of upper limit constraints, with the daily production scheduling attribute value being... The number of switching times is less than a specified value, where ； Equation 13 represents a continuity constraint, and the attribute value is... The length of the block must be within the specified range. ,in Indicates the first Each block attribute is The length of the cluster, Indicates the attribute is The optimal block length; Equation 14 represents the interval constraint, and the attribute value is There must be at least a gap between the clusters. Other attribute values ​​of the platform, among which Indicates that the attribute value is not equal to , Indicates the first The node attribute is not The length of the cluster; Equation 15 represents the adjacency constraint, and the attribute value is The following non-adjacent production scheduling attribute values ,in ； Equation 16 indicates that the constraint is included, specifying Each order range must include indivual ,in, Indicates the first The attribute value within each interval is Quantity; Equation 17 indicates that the batch contains each attribute value The block, whose corresponding interval sequence contains at least Platform attribute value ,in Indicates the first Each block attribute is The attribute value in the corresponding interval sequence is Quantity; Equation 18 represents the number of attribute batches for attribute values. The number of its blocks needs to be . There are 1, and the length difference of each block is the current minimum, where Represents attribute value The actual number of blocks, Indicates the penalty value. Indicates the first Each block attribute is The length of the cluster, This means that all attribute values ​​are Quantity; Equation 19 represents sorting by integer multiples for attribute values. Its clumping requires Production is scheduled in multiples of integers, with the remainder being... Production is scheduled separately, with the remainder When merged into any attribute value In the clumps, among which Indicates the first Each block attribute is The length of the cluster; Equation 20 represents the mixed-row constraint, which applies to attribute values. , Mixed production is not allowed, and the specified production order is that the first production run is completed first. , then arrange ,in This means that all attribute values ​​are The position of the first order in the order list. This means that all attribute values ​​are The position of the last order in the order book; Equation 21 represents the equilibrium constraint within the cycle time, assuming the workshop production cycle time is... JPH, or 1-hour production Trolley, for attribute values ,Every The number of tables must be equal, among which Indicates each The attribute value inside the trolley is Quantity, Indicates the total number of orders; Equation 22 represents the priority ordering for attribute values. Priority is given to the first vehicle on that day; this constraint only applies to the first vehicle. Equation 23 represents the specified position constraint, which refers to the first few order numbers in the initial order list. The corresponding orders need to be scheduled for production at the designated location. ; Equation 24 represents the buffer constraint, which means that the sequence difference of the same order before and after the change from the welding workshop through the buffer to the final assembly workshop must not exceed a specified value. ,in This indicates the number of units in the final assembly workshop's production scheduling results. The location of each order. This indicates the number of units in the final assembly workshop's production scheduling results. The position of each order in the production schedule of the welding workshop.

6. The automotive order optimization scheduling method based on APS according to claim 5, characterized in that: The algorithm in step three uses a genetic algorithm for coarse ranking, and the specific method is as follows: a. Perform population initialization; b. Conduct a selection evaluation; c. Perform cross operations; d. Insertion mutation operation; e. Introduce heuristic rules.

7. The APS-based automotive order optimization scheduling method according to claim 6, characterized in that: The algorithm in step three performs fine sorting after coarse sorting, and the specific method is as follows: a. Perform constraint repair; b. Perform two-way verification.

Images