Multi-target production scheduling method and device, equipment and storage medium
Through the multi-objective production scheduling method, the optimization of production scheduling schemes is solved by using algorithms such as constraint programming and heuristic search, and the problem of production scheduling in manufacturing enterprises depends on manual judgment, improve the accuracy and rationality of production scheduling schemes, and reduce the Production scheduling is time-consuming and resource-consuming.
Patent Information
- Application Number
- CN202411413768.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-11
- Publication Date
- 2024-11-08
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
In manufacturing enterprises, production scheduling relies on manual judgment, resulting in long-term task allocation, unreasonable resource allocation, unreasonable waiting for upstream and downstream work orders, and frequent changes in production plans, making it difficult to meet the multi-target scheduling needs, especially during delivery time , cost, operating time, and resource load balance, it is difficult to optimize labor.
采用多目标排产方法,通过获取需求排产目标及其权重信息,生成多目标函数,结合约束编程、启发式搜索、分支限算法和回溯搜索策略,构建排产模型,优化排产方案,确保方案的准确性和合理性。
It effectively improves the accuracy and rationality of the production scheduling plan, reduces the time-consuming production scheduling, improves resource utilization, and reduces the frequent changes in production plans and the difficulty of management.
Smart Images

Figure CN118917644A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of data processing technology, and in particular to a multi-objective production scheduling method, device, equipment and storage medium. Background Art
[0002] In the daily operations of manufacturing companies, production scheduling often relies on human subjective judgment. Faced with a large amount of order data, production planners often encounter the following challenges: 1. The task allocation process is time-consuming: Due to the large number of production work orders, manual task allocation takes a lot of time, affecting the overall production efficiency; 2. Unreasonable resource allocation: The allocation of production resources is often unscientific, resulting in waste of resources and reduced production efficiency; 3. Unreasonable waiting of upstream and downstream work orders: Improper allocation of upstream work orders will cause unreasonable waiting of downstream work orders, affecting the smoothness of the entire production process; 4. Frequent changes in production plans: Temporary adjustments to production orders, such as inserting orders, canceling orders, and expediting orders, will lead to frequent changes in the original production plan, increasing management difficulty; 5. Multi-objective scheduling requirements are difficult to meet: When considering multiple objectives such as delivery date, cost, operation duration, resource load balancing, etc., manual scheduling is difficult to achieve optimality and it is difficult to meet multiple objectives at the same time.
[0003] It can be seen that the existing technology still needs to be improved and enhanced. Summary of the invention
[0004] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a multi-objective production scheduling method, which combines multiple algorithms to achieve the generation of multi-objective production scheduling plans, and can effectively improve the accuracy and rationality of the generated production scheduling plans.
[0005] The first aspect of the present invention provides a multi-objective production scheduling method, including: obtaining a demand production scheduling target, and obtaining weight information and auxiliary information corresponding to the demand production scheduling target, generating a multi-objective function based on the demand production scheduling target and its weight information; obtaining basic production information, and constructing a production scheduling model based on the basic production information, auxiliary information and multi-objective function; setting hard constraints of constraint programming technology, and using constraint programming technology to solve and combinatorially optimize the production scheduling model to obtain multiple initial solutions that meet the constraints; searching for an approximate solution corresponding to the initial solution based on a heuristic search strategy; optimizing the approximate solution in combination with a branch and limit algorithm and a backtracking search strategy to obtain an optimal solution.
[0006] Optionally, in a first implementation method of the first aspect of the present invention, the demand production scheduling target is obtained, and the weight information and auxiliary information corresponding to the demand production scheduling target are obtained, and a multi-objective function is generated based on the demand production scheduling target and its weight information, including: obtaining the demand production scheduling target, the demand production scheduling target includes one or more production scheduling targets; according to the production scheduling target, using a keyword matching algorithm to match the auxiliary information corresponding to the production scheduling target from a pre-built auxiliary information library; when the production scheduling target includes multiple, obtaining the weight value corresponding to each production scheduling target, and combining multiple weight values to obtain weight information; using the production scheduling target as a variable, and using the weight value corresponding to the production scheduling target as a coefficient to construct a multi-objective function.
[0007] Optionally, in a second implementation method of the first aspect of the present invention, the obtaining of basic production information and constructing a production scheduling model based on the basic production information, auxiliary information and a multi-objective function include: obtaining basic production information, generating basic constraints based on the basic production information, and generating auxiliary constraints based on the auxiliary information; constructing a production scheduling model based on operations research ideas, integrating the basic constraints and auxiliary constraints into the production scheduling model, and setting the multi-objective function as the objective function of the production scheduling model.
[0008] Optionally, in a third implementation of the first aspect of the present invention, the hard constraints of the constraint programming technology are set, and the production scheduling model is solved and optimized by the constraint programming technology to obtain multiple initial solutions that meet the constraints, including: setting the hard constraints of the constraint programming technology, and the hard constraints include dependency order constraints, time constraints and production resource overlapping constraints; converting basic constraints, auxiliary constraints and hard constraints into constraint matrices; applying the constraint propagation algorithm of the constraint propagation technology to process the constraint matrix, and in the constraint propagation process, identifying the constraint matrix that can be processed in parallel; synchronously processing the constraint matrix that can be processed in parallel through a parallel computing framework; collecting all initial solutions that meet the hard constraints, basic constraints and auxiliary constraints obtained in the constraint propagation process and the parallel computing process to obtain an initial solution set including multiple initial solutions.
[0009] Optionally, in a fourth implementation of the first aspect of the present invention, the search for an approximate solution corresponding to an initial solution based on a heuristic search strategy includes: selecting a simulated annealing strategy to guide a search process based on a domain search algorithm; selecting any initial solution from a set of initial solutions, and generating a new approximate solution through a neighborhood search based on the selected initial solution; calculating an objective function value of the generated new approximate solution, and comparing the quality of the newly generated approximate solution with that of the selected initial solution; when the quality of the newly generated approximate solution is better than that of the selected initial solution, replacing the selected initial solution with the newly generated approximate solution, and returning to execute the generation of a new approximate solution through the domain search; when the quality of the selected initial solution is better than that of the newly generated approximate solution, returning to execute the selection of any initial solution from the combination of initial solutions; and when a preset stop condition is met, outputting multiple approximate solutions corresponding to the initial solution.
[0010] Optionally, in a fifth implementation of the first aspect of the present invention, the combination of the branch and bound algorithm and the backtracking search strategy optimizes the approximate solution to obtain the optimal solution, including: setting an initial branch and bound tree based on the multiple approximate solutions obtained, and selecting a node from the branch and bound tree for expansion; applying basic constraints, auxiliary constraints and hard constraints to the selected node to generate all possible child nodes; when the boundary value of the selected node cannot meet the optimal solution condition, deleting the selected node and all its child nodes; when the boundary value of the selected node meets the optimal solution condition, calculating the corresponding boundary for each newly generated child node, and determining whether the generated new child node is added to the branch and bound tree based on the calculated boundary; when the generated child node cannot meet the preset better solution condition, backtracking to the previous child node of the selected node based on the backtracking search strategy; when the search stop condition is met, outputting the optimal solution.
[0011] Optionally, in a sixth implementation method of the first aspect of the present invention, the steps are also included: when an order adjustment request is received, weight information and auxiliary information corresponding to the adjusted production scheduling target are obtained according to the adjusted production scheduling target included in the order adjustment request, and an adjustment multi-objective function is generated based on the adjusted production scheduling target and its weight information; in-production order information is obtained, and order locking constraints are generated based on the in-production order information; an adjustment production scheduling model is constructed based on basic production information, auxiliary information, adjustment multi-objective functions and order locking constraints; the production scheduling model is replaced with the adjustment production scheduling model, and the hard constraints of the constraint programming technology are returned to execute, and the constraint programming technology is used to solve the combinatorial optimization of the production scheduling model to obtain multiple initial solutions that meet the constraints.
[0012] The second aspect of the present invention provides a multi-objective production scheduling device, including: an acquisition module, used to obtain the demand production scheduling target, and obtain weight information and auxiliary information corresponding to the demand production scheduling target, and generate a multi-objective function based on the demand production scheduling target and its weight information; a construction module, used to obtain basic production information, and construct a production scheduling model based on the basic production information, auxiliary information and multi-objective function; a setting module, used to set the hard constraints of the constraint programming technology, and use the constraint programming technology to solve the production scheduling model for combinatorial optimization to obtain multiple initial solutions that meet the constraints; a search module, used to find an approximate solution corresponding to the initial solution based on a heuristic search strategy; an optimization module, used to optimize the approximate solution in combination with the branch and limit algorithm and the backtracking search strategy to obtain the optimal solution.
[0013] The third aspect of the present invention provides a multi-target production scheduling device, which includes: a memory and at least one processor, wherein the memory stores instructions; at least one of the processors calls the instructions in the memory to enable the multi-target production scheduling device to execute each step of the multi-target production scheduling method described in any one of the above items.
[0014] A fourth aspect of the present invention provides a computer-readable storage medium having instructions stored thereon, and when the instructions are executed by a processor, the various steps of any of the multi-objective production scheduling methods described above are implemented.
[0015] In the technical solution of the present invention, a production scheduling model is constructed by generating a multi-objective function based on the acquired basic production information, auxiliary information and demand production scheduling targets and their weight information; hard constraints of constraint programming technology are set, and constraint programming technology is used to solve the production scheduling model for combinatorial optimization to obtain multiple initial solutions that meet the constraints; approximate solutions corresponding to the initial solutions are searched based on a heuristic search strategy; approximate solutions are optimized by combining a branch and limit algorithm and a backtracking search strategy to obtain the optimal solution; the method disclosed in the present application combines a variety of algorithms to realize the generation of multi-objective production scheduling plans, which can effectively improve the accuracy and rationality of the generated production scheduling plans, greatly reduce the time consumption of production scheduling, and improve resource utilization. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 A first flow chart of a multi-objective production scheduling method provided by an embodiment of the present invention; Figure 2 A second flow chart of the multi-objective production scheduling method provided by an embodiment of the present invention; Figure 3 A third flow chart of the multi-objective production scheduling method provided by an embodiment of the present invention; Figure 4 A fourth flow chart of the multi-objective production scheduling method provided by an embodiment of the present invention; Figure 5 A fifth flow chart of the multi-objective production scheduling method provided by an embodiment of the present invention; Figure 6 A sixth flow chart of the multi-objective production scheduling method provided by an embodiment of the present invention; Figure 7 A seventh flow chart of the multi-objective production scheduling method provided by an embodiment of the present invention; Figure 8 A schematic diagram of the structure of a multi-objective production scheduling device provided in an embodiment of the present invention; Fig. 9 A schematic diagram of the structure of a multi-objective production scheduling device provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0017] The present invention provides a multi-objective production scheduling method, apparatus, device and storage medium. In the present invention, the terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0018] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 , an embodiment of the multi-objective production scheduling method in the embodiment of the present invention includes: 101. Obtain a demand scheduling target, and obtain weight information and auxiliary information corresponding to the demand scheduling target, and generate a multi-objective function based on the demand scheduling target and its weight information; In this embodiment, by obtaining the demand scheduling target and its corresponding weight information, it can be ensured that the scheduling model is more in line with actual business needs. At the same time, the introduction of weights enables the model to prioritize different targets according to importance, thereby improving the accuracy and rationality of the subsequently generated scheduling plan.
[0019] 102. Obtain basic production information, and build a production scheduling model based on the basic production information, auxiliary information and multi-objective functions; In this embodiment, the multi-objective function generated based on the demand scheduling target and weight information can comprehensively consider multiple production goals, such as cost, time, resource utilization, etc., to achieve more comprehensive production plan optimization.
[0020] 103. Set hard constraints of constraint programming technology, use constraint programming technology to solve the production scheduling model for combinatorial optimization, and obtain multiple initial solutions that meet the constraints; In this embodiment, by setting hard constraints and using constraint programming technology to solve the production scheduling model, it can be ensured that the obtained solution meets various constraints in actual production, such as equipment capacity, human resources, material supply, etc.; and it can provide a good starting point for subsequent optimization, which helps to improve the overall solution efficiency.
[0021] 104. Find an approximate solution corresponding to the initial solution based on a heuristic search strategy; In this embodiment, by searching for an approximate solution corresponding to an initial solution based on a heuristic search strategy, a high-quality solution can be quickly located in the solution space, thereby improving the solution speed and the quality of the solution.
[0022] 105. Combine the branch and limit algorithm and the backtracking search strategy to optimize the approximate solution and obtain the optimal solution; In this embodiment, the approximate solution is optimized using a branch and bound algorithm and a backtracking search strategy, which can further improve the accuracy of the solution and ensure that the final solution is the optimal solution, satisfies all constraints and reaches the optimal value of the objective function.
[0023] The present application discloses a multi-objective production scheduling method, which constructs a production scheduling model by generating a multi-objective function based on the acquired basic production information, auxiliary information and demand production scheduling targets and their weight information; sets hard constraints of constraint programming technology, adopts constraint programming technology to solve and optimize the production scheduling model, and obtains multiple initial solutions that meet the constraints; searches for approximate solutions corresponding to the initial solutions based on a heuristic search strategy; optimizes the approximate solutions by combining a branch and limit algorithm and a backtracking search strategy to obtain the optimal solution; the method disclosed in the present application combines a variety of algorithms to realize the generation of multi-objective production scheduling plans, which can effectively improve the accuracy and rationality of the generated production scheduling plans, greatly reduce the time consumption of production scheduling, and improve resource utilization.
[0024] See also Figure 2 The second embodiment of the multi-objective production scheduling method in the embodiment of the present invention includes: 201. Obtain a demand production scheduling target, where the demand production scheduling target includes one or more production scheduling targets; In this embodiment, the production scheduling target may be order punctuality rate, resource load rate, lowest production cost, fastest delivery rate, shortest production cycle, lowest number of equipment adjustments, etc.
[0025] 202. According to the production scheduling target, a keyword matching algorithm is used to match auxiliary information corresponding to the production scheduling target from a pre-built auxiliary information database; In this embodiment, a synonym matching algorithm is used to match auxiliary information corresponding to the production scheduling target from a pre-built auxiliary information library; for example, when the demand scheduling target includes the lowest production cost, it is necessary to additionally import the production cost information of each resource as auxiliary information; when the demand scheduling target includes the lowest number of equipment transfers, it is necessary to provide the equipment transfer information when switching between different production task orders as auxiliary information.
[0026] 203. When the production scheduling target includes multiple ones, obtain a weight value corresponding to each production scheduling target, and combine multiple weight values to obtain weight information; 204. Using the production scheduling target as a variable and the weight value corresponding to the production scheduling target as a coefficient, a multi-objective function is constructed; In this embodiment, the formula of the multi-objective function is expressed as: ; Among them, W is the weight value, T is the production scheduling target, and X is the target number.
[0027] See also Figure 3 The third embodiment of the multi-objective production scheduling method in the embodiment of the present invention includes: 301. Obtain basic production information, generate basic constraints based on the basic production information, and generate auxiliary constraints based on the auxiliary information; In this embodiment, the basic production information includes basic production task orders, resource capacity, equipment lists and other information.
[0028] 302. Construct a production scheduling model based on the idea of operations research, integrate basic constraints and auxiliary constraints into the production scheduling model, and set the multi-objective function as the objective function of the production scheduling model; In this embodiment, for example, when the demand scheduling target is to minimize production cost and minimize production time, its scheduling model is expressed as: ; Among them, O is the set of task orders, R is the set of production resources, and M is the set of production tools. It represents the time when task order i uses tool k in production resource j; represents the time cost of task order i using tool k in production resource j; Indicates whether task order i is scheduled to production resource j and uses tool k; Represents the weight parameter.
[0029] In this embodiment, the production scheduling model constructed based on operations research ideas aims to simulate and optimize the production process through mathematical methods to achieve goals such as improving production efficiency, reducing costs, and meeting customer needs; the constructed production scheduling model can comprehensively consider various production resources and constraints, such as equipment capacity, material supply, personnel allocation, time windows, etc., so as to formulate a reasonable production plan.
[0030] See also Figure 4 The fourth embodiment of the multi-objective production scheduling method in the embodiment of the present invention includes: 401. Setting hard constraints of constraint programming technology, wherein the hard constraints include dependency order constraints, time constraints, and production resource overlap constraints; In this embodiment, the dependency order constraint condition is expressed as: ; in, Indicates the start time of task order a. Indicates the end time of task order b; The time constraint is expressed as: ; in, represents the start time of task order i using tool k on production resource j; represents the waiting time for task order i to use tool k on production resource j; represents the operation time of task order i using tool k on production resource j; represents the end time of task order i using tool k on production resource j; The production resource overlap constraint is expressed as: ; in, Indicates the start time of production task order a of production resource j, Indicates the end time of production task order b of production resource j.
[0031] 402. Convert the basic constraint conditions, auxiliary constraint conditions and hard constraint conditions into a constraint matrix; In this embodiment, the basic constraints, auxiliary constraints and hard constraints are converted into constraint matrices to facilitate processing and optimization in computer programs; the constraint matrix provides a structured and standardized way to represent the constraints in the problem, so that the constraint propagation algorithm can calculate and iterate more efficiently; through polarity conversion, the mathematical modeling process of the problem can be simplified, and the efficiency of solving linear programming, integer programming or other optimization problems can be improved.
[0032] 403. Applying a constraint propagation algorithm of the constraint propagation technique to process the constraint matrix, and in the constraint propagation process, identifying the constraint matrix that can be processed in parallel; 404. Through the parallel computing framework, the constraint matrix that can be processed in parallel is processed synchronously; In this embodiment, the MPI parallel computing framework may be used to implement parallel processing of the constraint matrix.
[0033] In this embodiment, the constraint propagation algorithm of the constraint propagation technology is applied to process the constraint matrix, which can effectively identify the constraints that can be processed in parallel, thereby improving the overall processing efficiency; the constraint matrix that can be processed in parallel is synchronously processed by the parallel computing framework, which can significantly shorten the calculation time and speed up the problem solving speed; parallel processing can make full use of computing resources, avoid the waste of resources in single-threaded processing, and realize the optimal allocation and use of resources.
[0034] 405. Collect all initial solutions satisfying hard constraints, basic constraints and auxiliary constraints obtained in the constraint propagation process and the parallel computing process to obtain an initial solution set including multiple initial solutions.
[0035] See also Figure 5 The fifth embodiment of the multi-objective production scheduling method in the embodiment of the present invention includes: 501. Selecting a simulated annealing strategy to guide the search process based on the domain search algorithm; In this embodiment, a simulated annealing strategy is selected to guide the search process based on the domain search algorithm. The simulated annealing algorithm allows some solutions of poor quality to be accepted during the search process by introducing a probabilistic acceptance criterion, which helps to escape from the local optimal solution and increase the probability of finding the global optimal solution.
[0036] 502. Select any initial solution from the initial solution set, and generate a new approximate solution through neighborhood search based on the selected initial solution; In this embodiment, domain search is a method for finding a solution to a problem, which attempts to improve the current solution by searching for a new solution in the neighborhood of the current solution; this neighborhood can be defined based on a certain metric (such as Euclidean distance) or based on a problem-specific structure; generating a new approximate solution through domain search means exploring in the neighborhood of the current solution to find a better solution.
[0037] 503. Calculate the objective function value of the generated new approximate solution, and compare the quality of the newly generated approximate solution with the selected initial solution; 504. When the quality of the newly generated approximate solution is better than the quality of the selected initial solution, the newly generated approximate solution is used to replace the selected initial solution, and the process returns to generate a new approximate solution through the domain search; when the quality of the selected initial solution is better than the newly generated approximate solution, the process returns to select any initial solution from the combination of initial solutions; In this embodiment, when the quality of the newly generated approximate solution is better than the initial solution, the newly generated approximate solution is selected to replace the original initial solution, and then the field search is continued based on the newly generated approximate solution to generate more high-quality approximate solutions; on the contrary, if the quality of the newly generated solution is not as good as the initial solution, another initial solution is selected from the initial solution set as the starting point, and the iterative process is continued; the quality of the obtained approximate solution can be gradually improved, effectively avoiding falling into the local optimal solution, thereby increasing the possibility of finding the global optimal approximate solution corresponding to the initial solution.
[0038] 505. When a preset stop condition is met, output a plurality of approximate solutions corresponding to the initial solution; In this embodiment, the preset stopping condition is reaching a maximum number of iterations, that is, after reaching a preset maximum number of initial solution selections, the iteration is stopped and a plurality of approximate solutions corresponding to the initial solutions are output.
[0039] See also Figure 6 The sixth embodiment of the multi-objective production scheduling method in the embodiment of the present invention includes: 601. Setting an initial branch and bound tree according to the obtained multiple approximate solutions, and selecting a node from the branch and bound tree for expansion; In this embodiment, by setting an initial branch and bound tree, the solution space can be systematically explored to ensure that no possible solution is missed.
[0040] 602. Apply basic constraints, auxiliary constraints, and hard constraints to the selected node to generate all possible child nodes; In this embodiment, the basic constraints, auxiliary constraints and hard constraints are applied to effectively prune the nodes, reduce unnecessary node expansion and improve search efficiency.
[0041] 603. When the boundary value of the selected node cannot satisfy the optimal solution condition, the selected node and all its child nodes are deleted; In this embodiment, for nodes whose boundary values do not meet the optimal solution conditions, the nodes and all their child nodes are deleted in time, thus avoiding unnecessary calculations and saving resources.
[0042] 604. When the boundary value of the selected node meets the optimal solution condition, a corresponding boundary is calculated for each newly generated child node, and it is determined whether the generated new child node is added to the branch and bound tree according to the calculated boundary; In this embodiment, for nodes whose boundary values satisfy the optimal solution condition, the boundaries of newly generated child nodes are calculated, and it is decided whether to add the branch and bound tree, thereby ensuring the accuracy of the search process.
[0043] 605. When the generated child node cannot satisfy the preset better solution condition, backtrack to the previous child node of the selected node based on the backtracking search strategy; In this embodiment, when a child node cannot meet the conditions for a better solution, a backtracking search strategy is adopted to quickly return to the previous valid node and continue searching, thereby ensuring the flexibility of the search; the failure to meet the preset better solution conditions means that when a child node cannot calculate a better solution, at this time, based on the backtracking search strategy, backtrack to the previous child node of the selected node to try other branches.
[0044] 606. When the search stop condition is met, output the optimal solution; In this embodiment, the preset search stop condition may be that the search time exceeds a limit. When the search stop condition is met, the currently found optimal solution is output.
[0045] See also Figure 7 The seventh embodiment of the multi-objective production scheduling method in the embodiment of the present invention includes: 701. When receiving an order adjustment request, obtaining weight information and auxiliary information corresponding to the adjusted production scheduling target according to the adjusted production scheduling target included in the order adjustment request, and generating an adjustment multi-objective function based on the adjusted production scheduling target and its weight information; 702. Obtain the in-production order information, and generate the order locking constraint condition based on the in-production order information; In this embodiment, when the user changes the original production scheduling conditions by inserting orders, canceling orders, locking orders, adjusting order priorities, adjusting production resources, etc., the orders that are already in production, that is, the obtained order information in production, can be used to provide locking constraints (specify variable values and selection order) for adjusting the production scheduling model by locking orders. Specifically, when locking orders, the values of certain variables can be clearly specified based on the order information in production, such as product quantity, production time, production line selection, etc. These specified values will be used as input conditions of the production scheduling model to generate a production scheduling plan that meets the requirements; in some cases, it may be necessary to select resources or perform operations in a certain order during the production scheduling process. The locking constraints can set this selection order to ensure the rationality and efficiency of the production scheduling plan.
[0046] 703. Construct an adjustment scheduling model based on basic production information, auxiliary information, adjustment multi-objective functions and order locking constraints; 704. Replace the production scheduling model with the adjusted production scheduling model, and return to execute the hard constraints of the constraint programming technology, use the constraint programming technology to solve the combined optimization of the production scheduling model, and obtain multiple initial solutions that meet the constraints; specifically, return to execute steps 103 to 105 to obtain the optimal solution corresponding to the adjusted production scheduling model.
[0047] In this embodiment, when the user changes the original production scheduling conditions by inserting orders, canceling orders, locking orders, adjusting order priorities, adjusting production resources, etc., the orders that are already in production can provide constraints for adjusting the production scheduling model by locking the orders (specifying the values of variables and the selection order), and the unproduced production orders are rescheduled; that is, the method disclosed in the present application can realize dynamic adjustment of the production scheduling plan, can dynamically adjust and reschedule based on the existing production scheduling plan, flexibly handle complex and changeable production environments, reduce the impact of plan adjustments on the original plan, and meet the needs of real-time adjustment of the production scheduling plan.
[0048] The above describes the multi-objective production scheduling method in the embodiment of the present invention. The following describes the multi-objective production scheduling device in the embodiment of the present invention. Figure 8 , an embodiment of a multi-objective production scheduling device in an embodiment of the present invention includes: An acquisition module 801 is used to acquire a demand scheduling target, and acquire weight information and auxiliary information corresponding to the demand scheduling target, and generate a multi-objective function based on the demand scheduling target and its weight information; A construction module 802 is used to obtain basic production information and construct a production scheduling model based on the basic production information, auxiliary information and a multi-objective function; The setting module 803 is used to set hard constraint conditions of constraint programming technology, and use constraint programming technology to solve the production scheduling model for combinatorial optimization to obtain multiple initial solutions that meet the constraint conditions; A search module 804, used to search for an approximate solution corresponding to the initial solution based on a heuristic search strategy; The optimization module 805 is used to optimize the approximate solution by combining the branch and bound algorithm and the backtracking search strategy to obtain the optimal solution.
[0049] Figure 8 The structure of the multi-objective production scheduling device shown does not constitute a limitation on the multi-objective production scheduling device, and can implement the steps of the multi-objective production scheduling method provided by the above-mentioned method embodiments.
[0050] above Figure 8 The multi-objective production scheduling device in the embodiment of the present invention is described in detail from the perspective of modular functional entities, and the multi-objective production scheduling device in the embodiment of the present invention is described in detail from the perspective of hardware processing.
[0051] Fig. 9It is a structural schematic diagram of a multi-target production scheduling device provided by an embodiment of the present invention. The multi-target production scheduling device 900 may have relatively large differences due to different configurations or performances, and may include one or more processors 910 and memories 920, and one or more storage media 930 (for example, one or more massive storage devices) storing application programs 933 or data 932. Among them, the memory 920 and the storage medium 930 may be short-term storage or permanent storage. The program stored in the storage medium 930 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the multi-target production scheduling device 900. Furthermore, the processor 910 may be configured to communicate with the storage medium 930, and execute a series of instruction operations in the storage medium 930 on the multi-target production scheduling device 900 to implement the steps of the multi-target production scheduling method provided by the above-mentioned method embodiments.
[0052] The multi-target production scheduling device 900 may also include one or more power supplies 940, one or more wired or wireless network interfaces 950, one or more input and output interfaces 960, and / or one or more operating systems 931, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be appreciated by those skilled in the art that Fig. 9 The structure of the multi-objective production scheduling device shown does not constitute a limitation of the multi-objective production scheduling device, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.
[0053] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the multi-target production scheduling method.
[0054] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described system, device, or unit can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.
[0055] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the whole or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program code.
[0056] Finally, it should be noted that the above description is only a preferred example of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A multi-objective production scheduling method, characterized in that: include: Obtaining a demand scheduling target, and obtaining weight information and auxiliary information corresponding to the demand scheduling target, and generating a multi-objective function based on the demand scheduling target and its weight information; Obtain basic production information, and build a production scheduling model based on basic production information, auxiliary information, and multi-objective functions; Set hard constraints of constraint programming technology, use constraint programming technology to solve the production scheduling model for combinatorial optimization, and obtain multiple initial solutions that meet the constraints; Find an approximate solution corresponding to the initial solution based on a heuristic search strategy; The branch and bound algorithm and the backtracking search strategy are combined to optimize the approximate solution and obtain the optimal solution.
2. The multi-objective production scheduling method according to claim 1, characterized in that: The step of obtaining the demand scheduling target, obtaining weight information and auxiliary information corresponding to the demand scheduling target, and generating a multi-objective function based on the demand scheduling target and its weight information includes: Obtaining a demand scheduling target, wherein the demand scheduling target includes one or more scheduling targets; According to the production scheduling target, a keyword matching algorithm is used to match the auxiliary information corresponding to the production scheduling target from the pre-built auxiliary information library; When there are multiple production scheduling targets, a weight value corresponding to each production scheduling target is obtained, and multiple weight values are combined to obtain weight information; The production scheduling target is used as a variable, and the weight value corresponding to the production scheduling target is used as a coefficient to construct a multi-objective function.
3. The multi-objective production scheduling method according to claim 1, characterized in that: The obtaining of basic production information and building a production scheduling model based on the basic production information, auxiliary information and a multi-objective function include: Obtaining basic production information, generating basic constraints based on the basic production information, and generating auxiliary constraints based on the auxiliary information; The production scheduling model is constructed based on the idea of operations research, and the basic constraints and auxiliary constraints are integrated into the production scheduling model, and the multi-objective function is set as the objective function of the production scheduling model.
4. The multi-objective production scheduling method according to claim 3, characterized in that: The hard constraints of the constraint programming technology are set, and the constraint programming technology is used to solve the production scheduling model for combinatorial optimization to obtain multiple initial solutions that meet the constraints, including: Setting hard constraints of constraint programming technology, wherein the hard constraints include dependency order constraints, time constraints, and production resource overlap constraints; Convert basic constraints, auxiliary constraints and hard constraints into constraint matrices; The constraint propagation algorithm of the constraint propagation technique is applied to process the constraint matrix. During the constraint propagation process, the constraint matrix that can be processed in parallel is identified; Through the parallel computing framework, the constraint matrix that can be processed in parallel is processed synchronously; All initial solutions satisfying hard constraints, basic constraints and auxiliary constraints obtained in the constraint propagation process and the parallel computing process are collected to obtain an initial solution set including multiple initial solutions.
5. The multi-objective production scheduling method according to claim 4, characterized in that: The step of searching for an approximate solution corresponding to the initial solution based on a heuristic search strategy includes: The simulated annealing strategy is selected to guide the search process based on the domain search algorithm; Select any initial solution from the initial solution set, and generate a new approximate solution through neighborhood search based on the selected initial solution; Calculate the objective function value of the generated new approximate solution and compare the quality of the newly generated approximate solution with the selected initial solution; When the quality of the newly generated approximate solution is better than that of the selected initial solution, the newly generated approximate solution is used to replace the selected initial solution, and the process returns to generate a new approximate solution through domain search; when the quality of the selected initial solution is better than that of the newly generated approximate solution, the process returns to select any initial solution from the combination of initial solutions; When the preset stop condition is met, multiple approximate solutions corresponding to the initial solution are output.
6. The multi-objective production scheduling method according to claim 5, characterized in that: The method of combining the branch and bound algorithm with the backtracking search strategy to optimize the approximate solution and obtain the optimal solution includes: An initial branch-and-bound tree is set according to the obtained multiple approximate solutions, and a node is selected from the branch-and-bound tree for expansion; Apply basic constraints, auxiliary constraints and hard constraints to the selected nodes to generate all possible child nodes; When the boundary value of the selected node cannot meet the optimal solution condition, the selected node and all its child nodes are deleted; When the boundary value of the selected node meets the optimal solution condition, the corresponding boundary is calculated for each newly generated child node, and it is decided whether the generated new child node is added to the branch and bound tree according to the calculated boundary; When the generated child node cannot meet the preset better solution conditions, the search strategy is backtracked to the previous child node of the selected node; When the search stop condition is met, the optimal solution is output.
7. The multi-objective production scheduling method according to claim 1, characterized in that: Also includes: When receiving an order adjustment request, obtaining weight information and auxiliary information corresponding to the adjusted production scheduling target according to the adjusted production scheduling target included in the order adjustment request, and generating an adjustment multi-objective function based on the adjusted production scheduling target and its weight information; Obtain the information of in-production orders and generate order locking constraints based on the in-production order information; Construct an adjustment scheduling model based on basic production information, auxiliary information, adjustment multi-objective functions and order locking constraints; The production scheduling model is replaced by adjusting the production scheduling model, and the hard constraints of the constraint programming technology are returned to execute. The constraint programming technology is used to solve the combined optimization of the production scheduling model to obtain multiple initial solutions that meet the constraints.
8. A multi-objective production scheduling device, characterized in that: include: An acquisition module is used to acquire a demand scheduling target, and acquire weight information and auxiliary information corresponding to the demand scheduling target, and generate a multi-objective function based on the demand scheduling target and its weight information; The construction module is used to obtain basic production information and build a production scheduling model based on the basic production information, auxiliary information and multi-objective functions; The setting module is used to set the hard constraint conditions of the constraint programming technology, and use the constraint programming technology to solve the combined optimization of the production scheduling model to obtain multiple initial solutions that meet the constraint conditions; A search module, used for searching an approximate solution corresponding to an initial solution based on a heuristic search strategy; The optimization module is used to optimize the approximate solution by combining the branch and bound algorithm and the backtracking search strategy to obtain the optimal solution.
9. A multi-objective production scheduling device, characterized in that: The multi-target production scheduling device comprises: a memory and at least one processor, wherein instructions are stored in the memory; At least one of the processors calls the instructions in the memory to enable the multi-objective production scheduling device to execute each step of the multi-objective production scheduling method as described in any one of claims 1-7.
10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the various steps of the multi-objective production scheduling method as described in any one of claims 1-7 are implemented.
Citation Information
Patent Citations
Self-defined instruction automatic identification method based on constraint planning
CN110333857A
Chip high multilayer ceramic packaging substrate production line multi-constraint automatic scheduling method
CN117391423A
Multi-target intelligent optimization APS production scheduling method and system
CN118229039A
Intelligent scheduling method and intelligent scheduling system for cargo transportation
CN118278844A
Cited By
Carbon powder production line energy consumption optimization method and system
CN120725195A
Intelligent scheduling multi-objective optimization method based on pulse propagation algorithm
CN120725403A
Intelligent scheduling multi-objective optimization method based on pulse propagation algorithm
CN120725403B
