Production plan scheduling method and system

Abstract

The invention belongs to the technical field of production planning and scheduling systems, and discloses a production planning and scheduling method and system. The method comprises the following steps: acquiring process capability data of each machine tool and process demand data of each order; based on the process capability data and the process demand data, calculating to obtain a process matching degree matrix of each machine tool to each order; based on the process matching degree matrix, the orders are distributed to all machine tools; generating a remodeling time matrix of remodeling time between any two orders of each machine tool based on a pre-collected basic remodeling time database, the process capability data and the process demand data; on the basis of the remodeling time matrix of each machine tool, optimizing order sorting on each machine tool, and minimizing total remodeling time to obtain a final order allocation table; and the overall remodeling total time length is accurately evaluated and minimized, so that the utilization rate and the productivity of the machine tool are remarkably improved in a multi-variety and small-batch production mode, the delivery cycle is shortened, and the plan and execution deviation is reduced.

Description

Technical Field

[0001] This invention relates to the field of production planning and scheduling systems, and more specifically, to a production planning and scheduling method and system. Background Technology

[0002] In the manufacturing sector, the rational allocation of orders to machine tools and optimized scheduling are crucial for improving production efficiency. Traditional production scheduling methods typically only consider matching the process requirements of orders with the basic capabilities of machine tools, neglecting the changeover time costs incurred when switching machine tools to different orders. Especially in multi-variety, small-batch production models, the proportion of changeover time increases significantly, severely impacting machine tool utilization and overall capacity.

[0003] Existing patent application CN118095572A discloses a battery production scheduling method, apparatus, electronic device, and storage medium, comprising: performing gradient splitting and demand merging on battery product demand data to obtain target demand data, wherein the target demand data includes the production demand of each level of battery products; performing production scheduling calculations based on the target demand data to obtain a first set of production scheduling schemes; optimizing the first set of production scheduling schemes based on multiple preset optimization objectives to obtain a second set of production scheduling schemes; determining the production process corresponding to each second production scheduling scheme in the second set of production scheduling schemes based on preset sorting reference indicators to obtain a target set of production scheduling schemes; and outputting the target set of production scheduling schemes.

[0004] The existing patent application CN113627759B proposes a dynamic scheduling method for manufacturing resources in a mixed-line manufacturing system. The method includes: constructing a process model representation using data structures to complete data modeling of the actual product process model; completing data modeling of the correspondence between actual processes and manufacturing resources, and the corresponding theoretical process durations; constructing and maintaining information such as the number of various manufacturing resources, manufacturing resource preparation time, product changeover time, and other necessary information related to manufacturing resource utilization intervals; constructing a dynamic scheduling program encoder and decoder; designing and constructing a complete optimization program and setting program parameters; constructing a scheduling solution solver and solving the solution; publishing the scheduling manufacturing plan; constructing a detector; constructing a core data monitor for the manufacturing system operation; and triggering a rescheduling mechanism.

[0005] A patent application with publication number CN118569555A discloses an automatic production scheduling method, apparatus, equipment, and storage medium, comprising: acquiring target sales orders created for discrete manufacturing products, including the delivery priority of discrete manufacturing products, combined production mode identifiers, and combined production identifiers; determining the sub-product types of each sub-product in the product, and determining the production lines that can produce each sub-product based on the sub-product types to obtain target production lines; acquiring the production scheduling orders already created on each target production line; determining the production time intervals that can produce sub-products on each target production line according to the delivery priority based on the already scheduled production orders; and using digital twin technology combined with heuristic algorithms and combined production mode identifiers to perform forward scheduling and rearrangement of the production time intervals on the target production lines to obtain the production scheduling result.

[0006] However, the above-mentioned patent applications have the following drawbacks: they separate order allocation and sorting, and adopt static, single-factor changeover time assumptions, ignoring dynamic factors such as real-time equipment status. Ultimately, they can only rely on local heuristic algorithm iterations, which can easily fall into local optima, resulting in low machine tool utilization, high changeover costs, and frequent delivery delays.

[0007] In view of this, the present invention proposes a production planning and scheduling method and system to solve the above-mentioned defects. Summary of the Invention

[0008] To overcome the aforementioned deficiencies of the prior art and achieve the above objectives, the present invention provides the following technical solution: a production planning and scheduling method, comprising:

[0009] Obtain the process capability data of each machine tool and the process requirement data of each order;

[0010] Based on process capability data and process requirement data, the process matching degree matrix of each machine tool to each order is calculated;

[0011] Based on the process matching degree matrix, orders are allocated to each machine tool;

[0012] Based on the pre-collected basic changeover time database, process capability data, and process requirement data, a changeover time matrix is ​​generated for any two orders of each machine tool.

[0013] Based on the changeover time matrix of each machine tool, the order sorting on each machine tool is optimized to minimize the total changeover time, and the final order allocation table is obtained.

[0014] Furthermore, the process capability data includes the machine tool supported process types, process parameter ranges, and real-time status data; the process requirement data includes the order process type, process parameter requirements, and delivery weight.

[0015] The process parameters include temperature, pressure, rotational speed, and feed rate; the real-time status data includes mold wear and load rate.

[0016] Furthermore, methods for obtaining the process matching degree matrix of each machine tool for each order include:

[0017] Determine whether the machine tool supports process types that include order process types, and generate a process type matching matrix;

[0018] The order's process parameter requirements are compared with the machine tool's process parameter range to generate a parameter compatibility matrix;

[0019] The process type matching matrix and the parameter compatibility matrix are multiplied element by element to generate the process matching degree matrix of each machine tool for each order.

[0020] Furthermore, methods for generating the process type matching matrix include:

[0021] If the machine tool supports process types that include the order's process types, then the corresponding element in the process type matrix is ​​1;

[0022] If the machine tool's supported process types do not include the order's process types, then the corresponding element in the process type matrix is ​​0;

[0023] In the process type matrix, rows represent machine tools and columns represent orders.

[0024] Furthermore, methods for generating the parameter compatibility matrix include:

[0025] Obtain the midpoint value of the range of various process parameters of the machine tool;

[0026] Calculate the absolute value of the difference between each process parameter of the order and the midpoint value of each process parameter range of the machine tool;

[0027] The elements in the parameter compatibility matrix are calculated based on the absolute value and the width of the range of corresponding process parameters of the machine tool.

[0028] In the parameter compatibility matrix, rows represent machine tools and columns represent orders.

[0029] Furthermore, methods for allocating orders to individual machine tools include:

[0030] The matching degree threshold is obtained based on historical data analysis;

[0031] For each order, a set of machine tools with a matching degree greater than the matching degree threshold is selected based on the process matching degree matrix to form an order-machine tool allocation pair;

[0032] For each order-machine tool allocation pair, the allocation priority coefficient is calculated by combining the matching degree of the corresponding allocation pair, the load rate of the machine tool, and the delivery weight of the order.

[0033] Orders are assigned to the corresponding machine tools in descending order of priority coefficient.

[0034] Furthermore, the methods for generating the changeover time matrix for each machine tool include:

[0035] Extract the process parameter requirements of the assigned orders for each machine tool to form a process parameter group for each machine tool;

[0036] For any two orders in the assigned orders of each machine tool, calculate the Euclidean distance to generate the parameter difference matrix of each machine tool;

[0037] Collect real-time status data of machine tools, and sum the mold wear degree and load rate in the real-time status data by weighting, to obtain the status correction value matrix of each machine tool with the same matrix dimension as the parameter difference matrix;

[0038] From the pre-collected basic changeover time database, the corresponding basic changeover time is retrieved according to the machine tool model and order process type, forming a diagonal matrix with the same matrix dimension as the number of each machine tool and the number of allocated orders;

[0039] By using preset parameter difference weighting coefficients and state correction weighting coefficients, the diagonal matrix of the basic changeover time of each machine tool, combined with the parameter difference matrix and the state correction value matrix, is linearly combined to generate the changeover time matrix of each machine tool.

[0040] Furthermore, methods for obtaining the final order allocation table include:

[0041] For each machine tool with assigned orders, a sequence of machine tool orders is generated using random permutation or heuristic rules;

[0042] Based on the transformation time matrix, construct the objective function;

[0043] A neighborhood search is performed on each machine tool order sequence using a search algorithm to retain the order sequence with the minimum total changeover time.

[0044] After a preset number of iterations or when the convergence condition is met, the optimal order sequence for each machine tool is output, and the final order allocation table is obtained.

[0045] Furthermore, the objective function is to minimize the conversion time of all adjacent orders in the order sequence.

[0046] Furthermore, when generating the changeover time matrix for each machine tool, the basic changeover time is adjusted based on the operator team's skill level, including:

[0047] The skill disturbance coefficient is calculated based on the preset operator team skill level. The skill disturbance coefficient is then multiplied by the basic changeover time to obtain the operator skill correction matrix.

[0048] By using preset parameter difference weighting coefficients and state correction weighting coefficients, the diagonal matrix of the basic changeover time, the parameter difference matrix, the state correction value matrix, and the operator skill correction matrix are linearly combined to generate the changeover time matrix for each machine tool.

[0049] Compared with existing technologies, the technical effects and advantages of the production planning and scheduling method and system of the present invention are as follows:

[0050] This invention obtains machine tool process capability data and order process requirement data, calculates a process matching degree matrix to accurately match orders and machine tools, allocates orders based on the matching degree, generates a changeover time matrix by combining real-time machine tool status data and process parameter differences, and then optimizes order sorting through 2-opt and simulated annealing search algorithms to minimize the total changeover time.

[0051] This solution introduces priority coefficients and continuously generates matching degrees, order allocation, changeover time matrices, and sorting optimizations to ensure optimal synergy between allocation and sorting results. Instead of using static constants, changeover time incorporates multi-dimensional data such as process parameter differences, real-time equipment status, and operator team skill levels to generate a dynamic changeover time matrix in real time. At the solution level, a hierarchical heuristic strategy of 2-opt local improvement and simulated annealing global jump is introduced, retaining the advantage of rapid convergence while increasing the ability to escape local optima. It can accurately assess and minimize the overall changeover time, significantly improving machine tool utilization and capacity in multi-variety, small-batch production models, thereby shortening delivery cycles and reducing planning and execution deviations. Attached Figure Description

[0052] Figure 1 This is a schematic diagram of a production planning and scheduling system according to an embodiment of the present invention;

[0053] Figure 2 This is a flowchart of a production planning and scheduling method according to an embodiment of the present invention;

[0054] Figure 3 This is a flowchart of the method for obtaining the process matching degree matrix according to an embodiment of the present invention;

[0055] Figure 4 This is a schematic diagram of the 2-opt neighborhood search algorithm according to an embodiment of the present invention;

[0056] Figure 5 This is a schematic diagram of the simulated annealing algorithm according to an embodiment of the present invention. Detailed Implementation

[0057] To further clarify the technical problem to be solved by this application and its background, before proceeding with specific implementation methods, the relevant reaction mechanism, the limitations of the prior art, and the practical difficulties faced by those skilled in the art in solving the problem will be explained in detail.

[0058] Specifically, in discrete manufacturing, changeover time typically refers to the downtime required for a production line to switch from processing one product model to another. This includes equipment adjustments after downtime, process parameter settings, and necessary mold and tooling replacements. During changeover, the production line cannot produce anything, and this period is considered a non-value-adding production downtime, representing a significant form of production waste that is currently difficult to completely avoid. The greater the process differences between different product batches and the more complex the molds and fixtures that need to be replaced, the longer the changeover time tends to be. Conversely, if the products being produced continuously are similar and require fewer adjustments, the changeover time will be relatively shorter. This means that changeover time directly impacts production cycle time and production line capacity utilization: excessively long changeover times will reduce effective production time, delay production cycles, and even affect order delivery.

[0059] Although changeover time has a significant impact on production scheduling efficiency and capacity, existing scheduling technologies still have many limitations in handling it. Some existing solutions often pre-set the changeover time required for different products as a fixed constant and store it in a basic database for use as a scheduling parameter. This static approach fails to reflect the dynamic fluctuations of changeover time in actual production. For example, in patent CN118095572A, although factors such as tooling changeover time and waiting time during the changeover of different battery products are considered in calculating the changeover time, these time parameters themselves are fixed values ​​based on experience and lack a mechanism for adjustment according to real-time production conditions. In addition, existing scheduling models do not fully consider the impact of differences in the current state of machine tools and the skill level of operators on changeover efficiency. For example, traditional changeover processes mainly rely on manual experience, resulting in low efficiency and high error rates; the time required for different operators to perform changeovers may vary significantly, but this factor is often ignored in scheduling models. More importantly, existing technologies generally lack dynamic prediction mechanisms for changeover time: production scheduling typically uses historical averages or fixed estimates, making it impossible to predict the duration of upcoming changeovers and adjust plans in a timely manner based on real-time data. These limitations result in production scheduling schemes lacking flexibility and accuracy when it comes to changeovers, making it difficult to adapt to complex and ever-changing production environments.

[0060] In summary, due to the significant uncertainty and context-dependent nature of changeover time, production planners in the industry currently face numerous practical difficulties when scheduling production. First, the time required for changeover is difficult to accurately predict. Influenced by factors such as equipment status, product differences, the complexity of process adjustments, and the skill level of operators, changeover time fluctuates considerably in practice, and using fixed values ​​for evaluation often results in significant errors. Existing evaluation methods are crude and cannot precisely quantify the impact of each changeover in advance. Second, it is difficult to optimize the production sequence in a timely manner. Due to the lack of dynamic prediction of future changeover time, schedulers cannot adjust the production sequence in advance to avoid lengthy changeovers or to concentrate the production of similar products, missing opportunities to optimize scheduling and reduce changeover losses. As a result, discrepancies between plans and actual production frequently occur during the production process, either due to underestimating changeover time leading to production delays or due to overly conservative approaches resulting in wasted capacity. Furthermore, the uncertainty of changeover time makes it difficult to guarantee delivery dates, increasing the risk of contract fulfillment for manufacturing companies. These difficulties highlight the inadequacies of existing technologies in handling changeover time, urgently requiring an improved technical solution that can dynamically predict changeover time and optimize the production sequence. Therefore, this application proposes a production planning and scheduling method and system to address the above pain points, providing more accurate changeover time prediction and intelligent scheduling optimization to reduce changeover losses, improve capacity utilization and on-time delivery capability.

[0061] The technical solutions of the embodiments of the present invention will be described in detail, clearly, and completely below with reference to the accompanying drawings. It should be particularly noted that the specific embodiments described below are only for better illustrating and explaining the technical solutions of the present invention, and are intended to enable those skilled in the art to better understand and implement the present invention, and should not be construed as limiting the scope of protection of the present invention. Without departing from the spirit and substance of the present invention, those skilled in the art can modify, adjust, or make equivalent substitutions based on the content disclosed in the present invention, and these should all be considered within the scope of protection of the present invention.

[0062] Example 1

[0063] Please see Figure 1 As shown in the figure, this embodiment discloses a production planning and scheduling system, including a data acquisition module, an order matching module, an order allocation module, an order calculation module, and an order optimization module. Each module is connected by wired and / or wireless means to realize data transmission.

[0064] The data acquisition module is used to obtain the process capability data of each machine tool and the process requirement data of each order.

[0065] The process capability data includes the machine tool supported process types, process parameter ranges, and real-time status data; the process requirement data includes the order process type, process parameter requirements, and delivery weight.

[0066] The process parameters include temperature, pressure, rotational speed, and feed rate; the real-time status data includes mold wear and load rate.

[0067] For example, the data acquisition module acquires data by integrating an Enterprise Resource Planning (ERP) system, a Manufacturing Execution System (MES), and Industrial Internet of Things (IIoT) devices. For static process capability data such as machine tool supported process types and process parameter ranges, data is directly retrieved from the equipment management database via a standardized API interface. This database pre-stores machine tool factory parameters and process capability data. Real-time status data, such as mold wear, is collected in real-time by pressure sensors installed on the machine tool mold components, collecting the number of mold uses and converting the data into a wear coefficient within the 0-1 range. The load rate is calculated by real-time acquisition of spindle running time, feed rate, and other data from the machine tool's own control system, combined with the machine tool's rated capacity. In the order's process requirements data, the order process type and process parameter requirements are entered into the MES by the process design department during the product development stage and then synchronized to the data acquisition module via a data integration interface. Delivery weights are manually set by planning personnel in the MES based on factors such as order delivery urgency and customer priority, or automatically generated through preset rules. All data is standardized after acquisition and stored in a unified production data center for the data acquisition module to access.

[0068] The order matching module calculates the process matching degree matrix of each machine tool for each order based on process capability data and process requirement data.

[0069] Further, please refer to Figure 3 As shown, the methods for obtaining the process matching degree matrix of each machine tool for each order include:

[0070] Determine whether the machine tool supports process types that include order process types, and generate a process type matching matrix;

[0071] The order's process parameter requirements are compared with the machine tool's process parameter range to generate a parameter compatibility matrix;

[0072] Based on the process type matching matrix and parameter compatibility matrix, a process matching degree matrix for each machine tool to each order is generated.

[0073] For example, in the manufacturing execution system of an automotive parts manufacturing company, three machine tools M1, M2, and M3 need to process eight orders O1-O8. Machine tool M1 supports injection molding, with a temperature range of 180-220℃ and a pressure range of 5-10MPa; machine tool M2 supports milling, with a speed range of 1000-3000rpm and a feed rate of 0.05-0.2mm / r; machine tool M3 supports both injection molding and welding, with a temperature range of 150-250℃ and a pressure range of 4-12MPa. The process requirements for orders O1-O8 are as follows:

[0074] Order 01: Process type is injection molding, temperature requirement is 200℃, pressure requirement is 8MPa;

[0075] Order 02: The process type is milling, the required rotation speed is 2000 rpm, and the required feed rate is 0.1 mm / r;

[0076] Order 03: Process type is injection molding, temperature requirement is 210℃, pressure requirement is 9MPa;

[0077] Order 04: Process type is welding, temperature requirement is 180℃, pressure requirement is 6MPa;

[0078] Order 05: The process type is milling, the required rotation speed is 1500 rpm, and the required feed rate is 0.15 mm / r;

[0079] Order 06: Process type is injection molding, temperature requirement is 190℃, pressure requirement is 6MPa;

[0080] Order 07: Process type is welding, temperature requirement is 200℃, pressure requirement is 8MPa;

[0081] Order 08: The process type is milling, the required speed is 2500 rpm, and the required feed rate is 0.08 mm / r.

[0082] Furthermore, methods for generating the process type matching matrix include:

[0083] If the machine tool supports process types that include the order's process types, then the corresponding element in the process type matrix is ​​1;

[0084] If the machine tool's supported process types do not include the order's process types, then the corresponding element in the process type matrix is ​​0;

[0085] In the process type matrix, rows represent machine tools and columns represent orders.

[0086] For example, for the process type of order O1, machine tools M1 and M3 support this process, and the corresponding element in the process type matching matrix is ​​1; machine tool M2 does not support this process, and the corresponding element in the process type matching matrix is ​​0. Similarly, orders O2, O5, and O8 are only supported by machine tool M2; orders O4 and O7 are only supported by machine tool M3; and orders O3 and O6 are supported by both machine tools M1 and M3. This results in a 3×8 process type matching matrix.

[0087] .

[0088] Furthermore, methods for generating the parameter compatibility matrix include:

[0089] Obtain the midpoint value of the range of various process parameters of the machine tool;

[0090] Calculate the absolute value of the difference between each process parameter of the order and the midpoint value of each process parameter range of the machine tool;

[0091] The corresponding element in the parameter compatibility matrix = 1 - absolute value / width of the range of corresponding process parameters for the machine tool;

[0092] In the parameter compatibility matrix, rows represent machine tools and columns represent orders.

[0093] For example, order O1 requires a temperature of 200℃, while machine tool M1 has a temperature range of 180-220℃. The parameter compatibility score is 1 - |200-200| / (220-180) = 1; the score for a pressure of 8MPa is 1 - |8-7.5| / (10-5) = 0.9. The average of these two scores is 0.95. Similarly, the parameter compatibility scores for other orders and machine tools are calculated, generating a parameter compatibility matrix.

[0094] .

[0095] Multiply the process type matching matrix and the parameter compatibility matrix element by element to obtain the final process matching degree matrix:

[0096] ;

[0097] In this context, rows represent machine tools, and columns represent orders.

[0098] The purpose of calculating the process matching matrix is ​​to provide a quantitative basis for order allocation, ensuring that orders are assigned to the most suitable machine tools and reducing additional changeover costs or production delays caused by process mismatch. The reason for choosing to support process types and process parameters in calculating the process matching matrix is ​​as follows: Supported process types are a fundamental condition for whether an order can be processed on a machine tool; if the machine tool does not support the basic process type of the order, production cannot proceed. Process parameters determine the ease and efficiency with which the machine tool executes the order. Higher process parameter compatibility results in better machine tool stability when processing orders and shorter changeover adjustment time. Therefore, the combination of both comprehensively reflects the machine tool's ability to match orders.

[0099] The order allocation module assigns orders to each machine tool based on the process matching degree matrix.

[0100] Furthermore, methods for allocating orders to individual machine tools include:

[0101] The matching degree threshold is obtained based on historical data analysis;

[0102] For each order, a set of machine tools with a matching degree greater than the matching degree threshold is selected based on the process matching degree matrix to form an order-machine tool allocation pair;

[0103] For each order-machine tool allocation pair, the allocation priority coefficient is calculated by combining the matching degree of the corresponding allocation pair, the load rate of the machine tool, and the delivery weight of the order.

[0104] Orders are assigned to the corresponding machine tools in descending order of priority coefficient.

[0105] Furthermore, the matching degree threshold is based on statistical analysis of historical production data. When the defect rate of machine tool processing orders exceeds a predetermined threshold, and the average changeover time exceeds a predetermined threshold, the corresponding matching degree is used as the matching degree threshold. For example, when the matching degree is below 0.9, the defect rate of machine tool processing orders increases significantly, and the average changeover time increases by more than 20%. Therefore, 0.9 is set as the threshold to ensure production quality and efficiency.

[0106] For example, for order O1, the machine tool matching degree corresponding to it in the process matching degree matrix is ​​M1 (0.95) and M3 (0.93), and the matching degree threshold is 0.9. Therefore, machine tools M1 and M3 are selected to form order-machine tool allocation pairs (O1,M1) and (O1,M3). Similarly, order O2 only forms an allocation pair with machine tool M2, order O3 forms an allocation pair with machine tool M1 (0.92) and machine tool M3 (0.88), but M3 is excluded because its matching degree is lower than the matching degree threshold. Order O4 only forms an allocation pair with machine tool M3 (0.91), order O5 only forms an allocation pair with machine tool M2 (0.96), order O6 forms an allocation pair with machine tool M1 (0.97) and M3 (0.95), order O7 only forms an allocation pair with machine tool M3 (0.90), and order O8 only forms an allocation pair with machine tool M2 (0.94).

[0107] Furthermore, the method for calculating the allocation priority coefficient for each order-machine tool allocation pair is as follows: multiply the matching degree between the order and the machine tool, the current load of the machine tool, and the delivery weight of the order by three factors, specifically: allocation priority coefficient = matching degree × (1 − current load rate of the machine tool) × order delivery weight.

[0108] For example, machine tool M1 currently has a load rate of 70%, M2 has a load rate of 65%, and M3 has a load rate of 80%; the delivery weights of orders O1-O8 are 0.8, 0.6, 0.9, 0.7, 0.5, 0.85, 0.65, and 0.75, respectively. For the allocation pair (O1, M1), the allocation priority coefficient is the matching degree 0.95 multiplied by the load factor (1-0.7=0.3) and then multiplied by the delivery weight 0.8, resulting in 0.228; the coefficient for the allocation pair (O1, M3) is 0.93×(1-0.8)×0.8=0.1488. Similarly, calculate the priority coefficients for other allocation pairs: (O2,M2) is 0.98×(1-0.65)×0.6=0.2058, (O3,M1) is 0.92×0.3×0.9=0.2484, (O4,M3) is 0.91×0.2×0.7=0.1274, (O5,M2) is 0.96×0.35×0.5=0.168, (O6,M1) is 0.97×0.3×0.85=0.24735, (O6,M3) is 0.95×0.2×0.85=0.1615, (O7,M3) is 0.90×0.2×0.65=0.117, and (O8,M2) is 0.94×0.35×0.75=0.24975.

[0109] Orders are allocated in descending order of priority coefficient: First, (O8, M2) (priority coefficient 0.24975), order O8 is assigned to M2; next, (O3, M1) (priority coefficient 0.2484), order O3 is assigned to M1; then (O6, M1), with a priority coefficient of 0.24735, is assigned to M1 because it is higher than (O6, M3)'s 0.1615; then (O1, M1) (priority coefficient 0.228), which is higher than (O1, M3)'s 0.1488, so order O1 is assigned to M1; followed by (O2, M2), (O5, M2), (O4, M3), and (O7, M3). The final initial order allocation table is as follows: machine tool M1 is assigned to orders O3, O6, and O1; machine tool M2 is assigned to orders O8, O2, and O5; and machine tool M3 is assigned to orders O4 and O7.

[0110] The purpose of calculating the allocation priority coefficient is to comprehensively consider the process matching degree, machine tool load status and order delivery urgency, so as to avoid the problem of some machine tools being overloaded and others being idle due to allocation based solely on matching degree. This ensures that order allocation can meet process requirements, balance machine tool load, and prioritize urgent orders, thereby optimizing the overall production plan.

[0111] Furthermore, the methods for resetting the delivery option of an order include:

[0112] Get the time difference between the order's required delivery date and the current date. Then calculate the total time the order is expected to occupy the machine tool. The weight of the delivery order is calculated as follows: (This refers to the sum of processing time and changeover time.)

[0113] ;

[0114] If a certain order =10 days =6 days, then the weight of the option is 0.4; if < If the delivery date is urgent, the delivery weight is automatically set to 1.0 to trigger expedited processing. This method avoids the subjectivity of human intervention and ensures the repeatability of the algorithm.

[0115] The order calculation module generates a changeover time matrix for any two orders for each machine tool, based on process capability data and process requirement data.

[0116] Methods for generating the changeover time matrix for each machine tool include:

[0117] Extract the process parameter requirements of the assigned orders for each machine tool to form a process parameter group for each machine tool;

[0118] For any two orders in the assigned orders of each machine tool, calculate the Euclidean distance to generate the parameter difference matrix for each machine tool.

[0119] Collect real-time status data of machine tools, and sum the mold wear degree and load rate in the real-time status data by weighting them to obtain a status correction value matrix with the same matrix dimension as the parameter difference matrix for each machine tool.

[0120] From the pre-collected basic changeover time database, the corresponding basic changeover time is retrieved according to the machine tool model and order process type, forming a diagonal matrix with dimensions consistent with the number of each machine tool and the number of allocated orders.

[0121] By using preset parameter difference weighting coefficients and state correction weighting coefficients, the diagonal matrix of the basic changeover time of each machine tool, combined with the parameter difference matrix and the state correction value matrix, is linearly combined to generate the changeover time matrix of each machine tool.

[0122] For example, starting with the initial order allocation table, the changeover time matrix for each machine tool is generated.

[0123] For machine tool M2, first extract the process parameter groups of the assigned orders: order O8 (2500 rpm, 0.08 mm / r), order O2 (2000 rpm, 0.1 mm / r), and order O5 (1500 rpm, 0.15 mm / r), forming a three-dimensional process parameter group matrix. Calculate the Euclidean distance between any two orders. For example, if the speed difference between order O8 and order O2 is 500 rpm and the feed rate difference is 0.02 mm / r, the difference is... The speed difference between order O2 and order O5 is 500 rpm, and the feed rate difference is 0.05 mm / r, with a degree of difference of [missing information]. Similarly, the difference between order 08 and order 05 can be calculated to be approximately 0.45, generating a 3×3 parameter difference matrix D:

[0124] ;

[0125] in, This represents the parameter difference when switching from the i-th order to the j-th order.

[0126] Real-time status data of machine tool M2 was collected, yielding a mold wear degree (MJMS) of 0.7 and a load rate (FZL) of 75%, generating a status correction value matrix S. Each element in the status correction value matrix... The calculation method is as follows:

[0127] ;

[0128] Finally, we obtain the state correction value matrix S, which has the same dimension as the parameter difference matrix.

[0129] ;

[0130] in, This represents the status correction value when switching from the i-th order to the j-th order.

[0131] The mold wear coefficient of 0.6 and the load rate coefficient of 0.4 in the state correction value matrix were derived through a combination of historical data correlation analysis and expert experience. For example, a company conducted regression analysis on the past 100 changeover records of machine tool M2 and found that mold wear accounts for approximately 60% of the changeover time, meaning that for every 0.1 increase in wear, the changeover time increases by an average of 8 minutes. The load rate accounts for approximately 40% of the changeover time, meaning that for every 10% increase in load rate above the threshold, the setup time increases by 5 minutes. Therefore, the weights were set to 0.6 and 0.4. These proportions can be dynamically optimized in subsequent production using machine learning algorithms.

[0132] The basic changeover time for milling on machine tool M2 is retrieved from the basic changeover time database and is set to 1.2 hours. This forms a 3×3 diagonal matrix, where the main diagonal is 1.2 and the rest are 0. Using empirical values, parameter difference weighting coefficient α = 0.6 and state correction weighting coefficient β = 0.4 are set. A linear combination of the basic changeover time matrix, parameter difference matrix, and state correction matrix is ​​then performed to generate the changeover time matrix T. Each element in the changeover time matrix... The calculation method is as follows:

[0133] ;

[0134] Finally, the changeover time matrix T of machine tool M2 is obtained:

[0135] ;

[0136] in, This represents the time required to switch from the i-th order to the j-th order.

[0137] The basic changeover time database collects historical changeover data for each machine tool through the Manufacturing Execution System (MES), storing the basic changeover times categorized by "machine tool model + process type". For example, the basic changeover time of 1.2 hours for milling on machine tool M2 is derived from the average changeover time of the past 50 pure milling orders for this machine tool, and is corrected by combining the theoretical changeover time in the machine tool manual. The data collection cycle is 3 months, and the basic changeover time is automatically updated every 50 new records to ensure that the basic changeover time database reflects the latest equipment performance.

[0138] Furthermore, the method for setting the parameter difference weighting coefficient and the state correction weighting coefficient is as follows:

[0139] Based on production experience, the influence coefficients of process parameter differences and machine tool real-time status on changeover time are analyzed to determine the initial weight values, with higher influence coefficients assigned higher weights.

[0140] Through stress testing of the simulated production scheduling system, the weight combination was verified and optimized, and the weight combination that minimized the changeover time prediction error rate and maximized the actual execution consistency was selected.

[0141] The weights should be adjusted appropriately according to the specific needs of the industry scenario to meet the special requirements of actual production.

[0142] For example, through multiple production trials, it was found that differences in process parameters between orders are the primary driver of changeover time. For instance, a 10% increase in milling speed difference leads to a 12% increase in changeover time, while real-time machine tool status is a secondary factor; for example, a 10% increase in load rate leads to an 8% increase in changeover time. Therefore, process parameter differences are given higher weight. After initial weighting, stress tests were conducted using a simulation scheduling system: when α=0.6 and β=0.4, the changeover time prediction error rate was lowest, and the scheduling plan showed the highest consistency in actual production. This weight can be adjusted according to different industry characteristics; for example, in precision machining scenarios, the status correction weight can be increased to 0.5.

[0143] Similarly, the changeover time matrix is ​​calculated for machine tools M1 and M3, and the changeover time matrix for each machine tool is finally generated.

[0144] The parameter difference matrix reflects the actual degree of difference in process parameters between orders. The greater the parameter difference, the longer the time required for machine tool parameter adjustment, mold replacement, and other operations, making it a core factor affecting changeover time. For example, differences in spindle speed in milling orders directly affect spindle adjustment time, while differences in feed rate affect tool position calibration time. Quantifying these differences using Euclidean distance allows for a scientific assessment of the basic time required for process changeover.

[0145] The state correction value matrix incorporates the dynamic influence of the machine tool's real-time state. For example, the higher the mold wear, the longer the preparation time for mold replacement; the higher the load rate, the more the debugging time is caused by the decrease in machine tool stability.

[0146] The combination of the two approaches avoids the limitations of relying solely on historical average data. It takes into account both the inherent differences in the process characteristics of orders and the real-time changes in the current status of machine tools, ensuring the accuracy and dynamic adaptability of changeover time calculations and providing a reliable time cost basis for subsequent order sorting optimization.

[0147] The order optimization module optimizes the order sorting on each machine tool based on the changeover time matrix of each machine tool, minimizes the total changeover time, and obtains the final order allocation table.

[0148] Methods for obtaining the final order allocation table include:

[0149] For each machine tool with assigned orders, a sequence of machine tool orders is generated using random permutation or heuristic rules;

[0150] Based on the transformation time matrix, construct the objective function;

[0151] The search algorithm is used to perform a neighborhood search on each machine tool order sequence, calculate the total changeover time after swapping adjacent orders, and retain the order sequence with the shorter total changeover time.

[0152] After a preset number of iterations or when the convergence condition is met, the optimal order sequence for each machine tool is output, and the final order allocation table is obtained.

[0153] For example, for the order sequence O8→O2→O5 for machine tool M2, the system uses the NEH heuristic to generate an initial sequence, i.e., sorting the orders in descending order of processing time, and constructs a function with the objective of minimizing the total changeover time. Based on the changeover time matrix of M2:

[0154] ;

[0155] The total transition time for the current sequence O8→O2→O5 is 1.692(O8→O2) + 1.734(O2→O5) = 3.426 hours.

[0156] Please see Figure 4 As shown, the system uses the 2-opt neighborhood search algorithm to exchange adjacent orders, generating a candidate sequence O2→O8→O5. Its total exchange time is 1.692(O2→O8) + 1.83(O8→O5) = 3.522 hours, which is longer than the original sequence, so the original sequence is retained. Similarly, checking the sequence O8→O5→O2, its total exchange time is 1.83(O8→O5) + 1.734(O5→O2) = 3.564 hours, still longer than the original sequence. Therefore, the optimal order sequence for M2 remains O8→O2→O5.

[0157] Please see Figure 5 As shown, for the order sequence O3→O6→O1 of machine tool M1, simulated annealing is used for neighborhood search. The total transformation time of the initial sequence is: O3→O6 (1.9 hours) + O6→O1 (1.81 hours) = 3.71 hours. Randomly swapping the positions of O1 and O6 yields the candidate sequence O3→O1→O6, with a total transformation time of: O3→O1 (1.9 hours) + O1→O6 (1.81 hours) = 3.71 hours. After 50 iterations, the final sequence O1→O3→O6 is adopted, with a total transformation time of 3.740 hours, slightly higher than the initial sequence, but accepted during the annealing process due to the probability mechanism's escape from local optima.

[0158] The strategy of accepting inferior solutions is essentially an exploration cost paid for global optimization. During the high-temperature phase, inferior solutions are widely accepted to traverse the solution space, gradually converging to the optimal solution domain as the temperature decreases. For example, after 50 iterations, the algorithm may discover a better sequence with a total changeover time of less than 3.7 hours, which is difficult for deterministic algorithms to achieve. This capability directly translates into production benefits: a global reduction of 1 hour in changeover time means that machine tools can utilize that time to process more urgent orders, improving overall capacity utilization by 3-5%. Especially in large-scale order scenarios such as more than 10 orders per machine tool, simulated annealing can reduce the total changeover time by 10-15% compared to the 2-opt algorithm, while maintaining computational efficiency—by balancing optimization quality and real-time requirements through a preset number of iterations (e.g., 50), avoiding the impracticality of exhaustive search. Therefore, this optimization mechanism of retreating to advance allows the production system to approach the global optimum even under complex constraints, ultimately achieving the dual goals of minimizing changeover costs and improving on-time delivery.

[0159] The order sequence for machine tool M3 is O4→O7, which consists of only two orders. The changeover time for both is 2.8 hours, so no sorting optimization is required.

[0160] Finally, the optimal order sequence for each machine tool is output, forming the final order allocation table: M1(O1→O3→O6), M2(O8→O2→O5), M3(O4→O7). The total changeover time for the final order allocation table is 3.740 (M1) + 3.426 (M2) + 2.8 (M3) = 9.966 hours.

[0161] In this embodiment, the 2-opt algorithm and simulated annealing algorithm leverage their respective advantages to minimize changeover time for order sequences of different sizes. The 2-opt algorithm generates candidate sequences by swapping adjacent orders and calculates the total changeover time, strictly preserving better solutions. It quickly finds locally optimal order sequences within a smaller solution space, avoiding complex calculations and achieving efficient convergence, ensuring rapid optimization of small-scale order sorting. The simulated annealing algorithm utilizes its probabilistic acceptance mechanism, allowing for the acceptance of inferior solutions with a certain probability, thus escaping the local optimum trap and exploring better solutions in a larger solution space. By gradually approaching the global optimum through a preset number of iterations, it effectively addresses the problem of complex solution space caused by an increase in the number of orders. The combination of these two algorithms forms a hierarchical optimization strategy: 2-opt handles simple, small-scale sequences, while simulated annealing tackles complex, large-scale sequences. Both calculations are based on accurate cost data provided by the changeover time matrix, ensuring efficiency in sorting small-scale orders while improving the solution quality for large-scale orders. Together, they help the production scheduling system output high-quality order allocation schemes that balance process matching and changeover efficiency within a reasonable timeframe, significantly reducing total changeover time and improving production efficiency.

[0162] Example 2

[0163] This embodiment provides a production planning and scheduling system, which also includes:

[0164] When generating the changeover time matrix for each machine tool, the base changeover time is adjusted based on the operator team's skill level, specifically including:

[0165] The skill disturbance coefficient is calculated based on the preset operator team skill level. The skill disturbance coefficient is then multiplied by the basic changeover time to obtain the operator skill correction matrix.

[0166] By using preset parameter difference weighting coefficients and state correction weighting coefficients, the diagonal matrix of the basic changeover time, the parameter difference matrix, the state correction value matrix, and the operator skill correction matrix are linearly combined to generate the changeover time matrix for each machine tool.

[0167] Furthermore, the preset methods for operator team skill levels include:

[0168] Based on historical production data, the changeover efficiency of operators is tracked over a long period of time. Through statistical analysis, the actual skill level of each shift is determined, and the skill performance of the shifts is divided into multiple discrete levels. The level division is based on the deviation of the shift's actual changeover time from the basic changeover time in the historical production process.

[0169] Specifically, teams with historical changeover efficiency higher than the basic changeover time are defined as high-level skill teams, teams with efficiency close to the basic changeover time are defined as medium-level skill teams, and teams with efficiency significantly lower than the basic changeover time are defined as low-level skill teams. Each level corresponds to a preset skill disturbance coefficient, so that the basic changeover time can be corrected accordingly by applying the above skill disturbance coefficient during the calculation of the changeover time matrix.

[0170] Furthermore, the preset methods for the skill perturbation coefficient include:

[0171] The skill disturbance coefficient for high-skilled work teams is set to a value less than 1, for medium-skilled work teams it is set to 1, and for low-skilled work teams it is set to a value greater than 1. For example, high-skilled operator teams, due to their proficiency and high efficiency, typically have shorter actual changeover times than the basic changeover time; therefore, a skill disturbance coefficient of less than 1 is set to reflect this improved changeover efficiency. Low-skilled teams, due to insufficient experience or poor proficiency, have significantly longer actual changeover times than the basic changeover time; therefore, a skill disturbance coefficient of greater than 1 is set to reflect this reduced changeover efficiency. Medium-skilled teams, whose actual performance is close to the standard basic changeover time, have a coefficient of 1 without further adjustment. The main purpose of this setting is to quantify the actual impact of operator skills on changeover efficiency through the skill disturbance coefficient, thereby more accurately predicting changeover times in the actual production environment and improving the accuracy and practicality of production scheduling plans.

[0172] Furthermore, the method for generating the changeover time matrix for each machine tool by linearly combining the diagonal matrix, parameter difference matrix, state correction value matrix, and operator skill correction matrix of the basic changeover time using preset parameter difference weighting coefficients and state correction weighting coefficients includes:

[0173] First, the parameter difference weight coefficient is multiplied by the parameter difference matrix, and the state correction weight coefficient is multiplied by the state correction value matrix. Then, the diagonal matrix of the basic changeover time and the operator skill correction matrix are combined together. Through matrix addition, a comprehensive changeover time matrix considering parameter differences, real-time machine tool status, and operator team skill level is obtained. Finally, it is used to optimize the production scheduling plan to minimize the total changeover time.

[0174] Adjusting basic changeover times based on operator team skill levels can more accurately reflect changeover efficiency in actual production environments. Since operator skill levels directly impact the efficiency of process adjustments and equipment changeovers during production, incorporating team skill levels allows for a more comprehensive consideration of the influence of human factors on changeover times, thereby improving the accuracy of changeover time predictions. This method makes production scheduling plans closer to actual production, helping to reduce deviations between production plans and execution, and improving the stability and effectiveness of plan execution. Furthermore, this method clarifies the contribution of operator skill differences to production efficiency, providing crucial information for companies to specifically improve employee skills and optimize human resource allocation.

[0175] Furthermore, the operator team skill level can be dynamically adjusted based on the order delivery weight during the changeover time correction. Specifically, during the changeover process for orders with high delivery weight, the system prioritizes scheduling operator teams with higher skill levels to perform changeover tasks, thereby improving changeover efficiency and ensuring on-time delivery. When high-skilled teams are unavailable, the system adjusts the changeover time based on the skill level of currently available teams and assigns them higher scheduling priority in conjunction with the order delivery weight, thus reflecting the potential delay risk caused by insufficient skills in the changeover time matrix.

[0176] Linking operator team skill levels with order delivery options can effectively enhance the system's sensitivity to critical orders, dynamically allocate highly skilled human resources during production scheduling, and reduce delivery deviations caused by human error in changing order settings.

[0177] Example 3

[0178] Please see Figure 2 As shown, this embodiment provides a production planning and scheduling method, including:

[0179] Obtain the process capability data of each machine tool and the process requirement data of each order;

[0180] Based on process capability data and process requirement data, the process matching degree matrix of each machine tool to each order is calculated;

[0181] Based on the process matching degree matrix, orders are allocated to each machine tool;

[0182] Based on process capability data and process requirement data, generate a changeover time matrix for any two orders for each machine tool;

[0183] Based on the changeover time matrix of each machine tool, the order sorting on each machine tool is optimized to minimize the total changeover time, and the final order allocation table is obtained.

[0184] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

[0185] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A production planning and scheduling method, characterized in that, include: Obtain the process capability data of each machine tool and the process requirement data of each order; Based on process capability data and process requirement data, the process matching degree matrix of each machine tool to each order is calculated; Based on the process matching degree matrix, orders are allocated to each machine tool; Based on the pre-collected basic changeover time database, process capability data, and process requirement data, a changeover time matrix is ​​generated for any two orders of each machine tool. Based on the changeover time matrix of each machine tool, the order sorting on each machine tool is optimized to minimize the total changeover time, and the final order allocation table is obtained.

2. The production planning and scheduling method according to claim 1, characterized in that, The process capability data includes the machine tool supported process types, process parameter ranges, and real-time status data; the process requirement data includes the order process type, process parameter requirements, and delivery weight. The process parameters include temperature, pressure, rotational speed, and feed rate; the real-time status data includes mold wear and load rate.

3. The production planning and scheduling method according to claim 1, characterized in that, Methods for obtaining the process matching degree matrix of each machine tool for each order include: Determine whether the machine tool supports process types that include order process types, and generate a process type matching matrix; The order's process parameter requirements are compared with the machine tool's process parameter range to generate a parameter compatibility matrix; The process type matching matrix and the parameter compatibility matrix are multiplied element by element to generate the process matching degree matrix of each machine tool for each order.

4. The production planning and scheduling method according to claim 3, characterized in that, Methods for generating process type matching matrices include: If the machine tool supports process types that include the order's process types, then the corresponding element in the process type matrix is ​​1; If the machine tool's supported process types do not include the order's process types, then the corresponding element in the process type matrix is ​​0; In the process type matrix, rows represent machine tools and columns represent orders.

5. The production planning and scheduling method according to claim 3, characterized in that, Methods for generating parameter compatibility matrices include: Obtain the midpoint value of the range of various process parameters of the machine tool; Calculate the absolute value of the difference between each process parameter of the order and the midpoint value of each process parameter range of the machine tool; The elements in the parameter compatibility matrix are calculated based on the absolute value and the width of the range of corresponding process parameters of the machine tool. In the parameter compatibility matrix, rows represent machine tools and columns represent orders.

6. The production planning and scheduling method according to claim 1, characterized in that, Methods for allocating orders to machine tools include: The matching degree threshold is obtained based on historical data analysis; For each order, a set of machine tools with a matching degree greater than the matching degree threshold is selected based on the process matching degree matrix to form an order-machine tool allocation pair; For each order-machine tool allocation pair, the allocation priority coefficient is calculated by combining the matching degree of the corresponding allocation pair, the load rate of the machine tool, and the delivery weight of the order. Orders are assigned to the corresponding machine tools in descending order of priority coefficient.

7. The production planning and scheduling method according to claim 1, characterized in that, Methods for generating the changeover time matrix for each machine tool include: Extract the process parameter requirements of the assigned orders for each machine tool to form a process parameter group for each machine tool; For any two orders in the assigned orders of each machine tool, calculate the Euclidean distance to generate the parameter difference matrix of each machine tool; Collect real-time status data of machine tools, and sum the mold wear degree and load rate in the real-time status data by weighting, to obtain the status correction value matrix of each machine tool with the same matrix dimension as the parameter difference matrix; From the pre-collected basic changeover time database, the corresponding basic changeover time is retrieved according to the machine tool model and order process type, forming a diagonal matrix with the same matrix dimension as the number of each machine tool and the number of allocated orders; By using preset parameter difference weighting coefficients and state correction weighting coefficients, the diagonal matrix of the basic changeover time of each machine tool, combined with the parameter difference matrix and the state correction value matrix, is linearly combined to generate the changeover time matrix of each machine tool.

8. The production planning and scheduling method according to claim 1, characterized in that, Methods for obtaining the final order allocation table include: For each machine tool with assigned orders, a sequence of machine tool orders is generated using random permutation or heuristic rules; Based on the transformation time matrix, construct the objective function; A neighborhood search is performed on each machine tool order sequence using a search algorithm to retain the order sequence with the minimum total changeover time. After a preset number of iterations or when the convergence condition is met, the optimal order sequence for each machine tool is output, and the final order allocation table is obtained.

9. A production planning and scheduling method according to claim 7, characterized in that, The objective function is to minimize the changeover time of all adjacent orders in the order sequence.

10. A production planning and scheduling method according to claim 7, characterized in that, When generating the changeover time matrix for each machine tool, the base changeover time is adjusted based on the operator team's skill level, including: The skill disturbance coefficient is calculated based on the preset operator team skill level. The skill disturbance coefficient is then multiplied by the basic changeover time to obtain the operator skill correction matrix. By using preset parameter difference weighting coefficients and state correction weighting coefficients, the diagonal matrix of the basic changeover time, the parameter difference matrix, the state correction value matrix, and the operator skill correction matrix are linearly combined to generate the changeover time matrix for each machine tool.

11. A production planning and scheduling system for implementing the production planning and scheduling method according to any one of claims 1-10, characterized in that, include: The data acquisition module is used to acquire the process capability data of each machine tool and the process requirement data of each order; The order matching module calculates the process matching degree matrix of each machine tool for each order based on process capability data and process requirement data. The order allocation module assigns orders to each machine tool based on the process matching degree matrix. The order calculation module generates a changeover time matrix for any two orders for each machine tool, based on process capability data and process requirement data. The order optimization module optimizes the order sorting on each machine tool based on the changeover time matrix of each machine tool, minimizes the total changeover time, and obtains the final order allocation table.

Images