A part-level production scheduling method based on forest node segmentation and sorting

Abstract

The present invention relates to the field of intelligent production technology, and in particular to a part-level production scheduling method based on forest node segmentation and sorting, comprising the following steps: setting a forest data structure composed of all task process task nodes, inputting non-consumable production resource stocks, consumable production resource stocks, and an enterprise shift and rest time model, setting a production tact time, and initializing intermediate process variables; placing leaf task nodes c of executable process splits into a node set E, judging whether the number of leaf task nodes in the node set E is 0, and then performing multiple dequeue operations on each queue f according to the queue first-in-first-out principle and the non-consumable resource stocks, placing the remaining nodes in the queue set F back into the node set E, and outputting an output node set G. The present invention divides constraint conditions into consumable resources and non-consumable resources, structures production tasks into data, and realizes a production scheduling algorithm applicable to the part level.

Description

Technical Field

[0001] The present invention relates to the field of intelligent production technology, and in particular to a part-level production scheduling method based on forest node segmentation and sorting. Background Art

[0002] The statements in this section merely provide background information related to the present disclosure and may not constitute prior art.

[0003] Scheduling is the production plan scheduling. The current scheduling methods are to use computers to run scheduling algorithms to achieve more accurate and fast scheduling based on constraints and target conditions. Among them, constraints generally refer to production resource constraints, such as factors such as people, machines, materials, methods, and environment; target conditions generally refer to construction period and task volume. There are many common scheduling methods at present, including limited capacity algorithms based on constraint theory; rule-based algorithms, linear programming, heuristic rule algorithms, and expert systems; there are also algorithms based on genetic algorithms, simulated annealing algorithms, ant colony / particle swarm algorithms, and neural networks; there are also AI algorithms + dynamic adjustment algorithms MAS multi-intelligent agent systems, and algorithms that use MAS multi-agent consultation for dynamic adjustment.

[0004] After years of iterative optimization, these algorithms have performed relatively well in actual production activities and can more accurately and efficiently solve the problem of scheduling confusion in production plans during the production process. Summary of the invention

[0005] Through research, the inventors found that in the process of intelligent production sorting, different characteristics, features and characteristics of some production scenarios are often encountered. For the above contents, specific constraints are formed. However, the currently disclosed technologies have not been properly classified, resulting in limited scenarios for the use of these sorting algorithms.

[0006] The purpose of this disclosure is to provide a part-level production scheduling method based on forest node segmentation and sorting.

[0007] According to one aspect of the present disclosure, there is provided an air traffic control data storage device based on training flight, comprising the following steps: Step 1. Setting a forest data structure composed of all task process task nodes, inputting non-consumable production resource inventory, consumable production resource inventory, enterprise shift and rest time model, setting production cycle time, initializing intermediate process variable node set E, queue set F, and output node set G; Step 2. Putting the leaf task node c of the executable process split into the node set E, and judging whether the number of leaf task nodes c in the node set E is 0, if so, proceeding to step 7; if not, taking T0 as the initial time of the leaf task node c in the node set E, and taking T0 + the time length of the leaf task node c as the end time of the leaf task node c. time, and judge whether there is non-consumable production resource competition between leaf task nodes c in node set E in the same time period. If yes, construct a queue f for leaf task nodes c with the same resource competition respectively. In each queue f, sort the priorities of leaf task nodes c in descending order, and store the nodes in queues from high to low in order of priority to enter step 3. If no, store the nodes without resource competition in a separate queue f; Step 3. According to the queue first-in-first-out principle and the non-consumable resource inventory, perform multiple dequeue operations on each queue f, and allocate non-consumable resources for each node until the allocation of non-consumable resources is completed, and enter step 4; Step 4. Put the dequeued leaf task nodes c into the output node set G, g records the round, node ID, start time, and end time of the leaf task node c, and determines whether the end time of the leaf task node c exceeds the production cycle time. If so, go to step 5. If not, query the material inventory to determine whether the resource consumption demand of each task node in the queue set F can be met within the guaranteed time. If it can be met, subtract the demand from the inventory number and update the inventory. If not, record the leaf task node c that cannot be met, remove the leaf task node c that cannot be met from the output node set G, and put it back into the node set E; step 5. Put the remaining nodes in the queue set F back into the node set E, and determine whether there is a next process task node. If so, find the parent node of the process task node corresponding to the leaf task node c in the output node set G and go to step 6. If not, go to step 6; step 6. Update the start time of the next round of leaf task node c and return to step 2; step 7. Output node set G; step 8. Repeat the above steps 1-6 until the execution is completed.

[0008] In sorting production, from the initial human-assisted screening algorithm to the later intelligent algorithm, it aims to improve the reasonable answers to problems such as sequential processing in sorting production. The inventors found in actual work that because the currently disclosed technical solutions are all directly sorting operations based on the features in the existing sorting scenarios, no classification processing is performed, which leads to chaos in production and has the potential to increase risks.

[0009] In some embodiments of the present disclosure, step 1 further includes: the element in the queue set F is queue f, and the node in the output node set G is g.

[0010] In some embodiments of the present disclosure, the element type in the queue f is a leaf task node c, which is used to store the leaf task nodes c selected according to non-consumable resource constraints.

[0011] In some embodiments of the present disclosure, the node set G uses g to record the filtered leaf task node information, and the attributes of g include round, node ID, start time, and end time.

[0012] In some embodiments of the present disclosure, the node set E is used to store unfiltered leaf task nodes c.

[0013] Compared with the currently disclosed technologies, the present disclosure has the following advantages and beneficial effects: the present disclosure divides the constraints into consumable resources and non-consumable resources. Consumable resources refer to dynamically changing resources that are constantly consumed and replenished in production, and non-consumable resources refer to relatively static resources that are not consumed in production. By structuring the data of production tasks and classifying and disposing consumable resources and non-consumable resources, a production scheduling algorithm applicable to the part level is realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a schematic diagram of the process of the present invention;

[0015] Figure 2 It is a schematic diagram of the task node structure of the present invention;

[0016] Figure 3 This is a schematic diagram of the process of step 1 of the present invention;

[0017] Figure 4 This is a schematic diagram of the process of step 2 of the present invention;

[0018] Figure 5 Schematic diagram of step three of the present invention;

[0019] Figure 6 Schematic diagram of step 4 of the present invention;

[0020] Figure 7 Schematic diagram of step five of the present invention. DETAILED DESCRIPTION

[0021] Please refer to the attached instructions Figure 1 - Instructions attached Figure 7 This embodiment provides a part-level production scheduling method based on forest node segmentation and sorting. The part-level production scheduling method based on forest node segmentation and sorting is already in the actual testing and use stage.

[0022] In the following paragraphs, different aspects of the embodiments are defined in more detail. Each aspect defined in this way can be combined with any other aspect or aspects, unless it is explicitly stated that it cannot be combined. In particular, any feature that is considered to be preferred or advantageous can be combined with one or more other features that are considered to be preferred or advantageous. The terms "first", "second", etc. that appear in the present invention are only for the convenience of description to distinguish different components with the same name, and do not indicate a sequential or primary-secondary relationship.

[0023] Example

[0024] In some embodiments of the present disclosure, according to the attached description Figure 1As shown, it includes the following steps: Step 1. Set the forest data structure composed of all task process task nodes, input the non-consumable production resource inventory, the consumable production resource inventory, the enterprise shift time model, set the production cycle time, initialize the intermediate process variable node set E, the queue set F, and the output node set G; Step 2. Put the leaf task node c of the executable process split into the node set E, and judge whether the number of leaf task nodes c in the node set E is 0. If so, go to step 7; if not, take T0 as the initial time of the leaf task node c in the node set E, take T0 + the time length of the leaf task node c as the end time of the leaf task node c, and judge whether there is a contention for non-consumable production resources between the leaf task nodes c in the node set E in the same time period. If so, construct a queue f for the leaf task nodes c that have the same resource contention, and in each queue f, put the leaf task node c Arrange the priorities in descending order, and store the nodes in the queue in order from high to low priority, and enter step 3. If there is none, store the nodes without resource contention in a separate queue f; Step 3. According to the first-in-first-out principle of the queue, according to the stock of non-consumable resources, perform multiple dequeue operations on each queue f, and allocate non-consumable resources for each node until the non-consumable resources are allocated, and enter step 4; Step 4. Put the dequeued leaf task node c into the output node set G, g records the round, node ID, start time, and end time of the leaf task node c, and judge whether the end time of the leaf task node c exceeds the production cycle time. If so, enter step 5. If not, query the material inventory to judge whether the consumption resource demand of each task node in the queue set F can be met within the guaranteed time. If it can be met, the inventory number is subtracted from the demand number, and the inventory is updated. If not, record the leaf task node c that cannot be met, and the leaf task node c that cannot be met Remove from the output node set G and put it back into the node set E; Step 5. Put the remaining nodes in the queue set F back into the node set E, and determine whether there is a next process task node. If so, find the parent node of the process task node corresponding to the leaf task node c in the output node set G and enter step 6. If not, enter step 6; Step 6. Update the start time of the next round of leaf task nodes c and return to step 2; Step 7. Output node set G; Step 8. Repeat the above steps 1-6 until the execution is completed. The intermediate process variable node set E, queue set F, and output node set G are initialized. The elements in the queue set F are queue f, the nodes in the output node set G are g, and the element type in the queue f is leaf task node c, which is used to store leaf task nodes selected according to non-consumable resource constraints. The node set G uses g to record the selected leaf task node information. The attributes of g include round, node ID, start time, and end time. The node set E is used to store unselected leaf task nodes c.

[0025] It should be noted that, under the existing public technology, the definition of a task is to complete a certain number of goals within a certain period of time. The total production task can also be composed of multiple subtasks, each subtask is regarded as a task node, and the number of goals that can be completed within a time period is divided into a task node, such as Figure 2 As shown in the figure, under the premise of sufficient materials, the target quantity in a time period can be determined by personnel, equipment, and tooling. The target quantity in the time period may be different for different types of parts. The time period should be continuous working time, under this premise, it can be flexibly defined as 4 hours, 8 hours, etc.

[0026] Based on the above, in actual production, there are usually orders, which are equivalent to a general task, including the type of production target and its quantity, total time requirements and other core elements. According to the type of parts, a general task is divided into multiple first-level sub-task nodes. A first-level sub-task node represents the completion of the production of all parts of this type, which is called a category task node; according to the sequence of production processes for each part, the first-level sub-task node will extend multiple levels of task nodes until it extends to the initial process, which is called a process task node; in each process task node, the process node is divided into multiple leaf task nodes according to the total amount of parts and the unit quantity determined by the time granularity.

[0027] The specific steps are as follows: the type of part determines the type of personnel, equipment, and tooling required; the type and quantity of personnel, equipment, and tooling can determine the number of corresponding parts produced in one day (unit quantity determined by time granularity); the total target amount under this part type is divided by the unit quantity to obtain the number of leaf task nodes, and the node division can be realized.

[0028] The attributes of the task node are as follows:

[0029] (1) Category task node: The category task node of a part is the task target under the category part, providing the corresponding data set without splitting and sorting. The relevant information set of the process task node is a subset of the information in the category task node. The attributes of the category task node are: ① node ID; ② part information table; ③ quantity; ④ target completion time; ⑤ process method set; ⑥ bill of materials; ⑦ priority; ⑧ personnel group set; ⑨ equipment group set; ⑩ tooling group set:

[0030] (2) Process task nodes: process task nodes at all levels are different from leaf task nodes. Only after the current first-level process task nodes are divided into leaf task nodes and sorted, can the next-level process task nodes be divided into leaf task nodes and sorted. The combination of leaf task nodes forms a process task node, and the process task nodes can be divided into leaf task nodes by calculating the time granularity. The attributes of the process task node are: ① Node ID; ② Parts information table; ③ Quantity; ④ Process method set; ⑤ Parent process task node ID; ⑥ Child process task node ID; ⑦ Bill of materials; ⑧ Priority; ⑨ Personnel group set; ⑩ Equipment group set; Tooling set collection:

[0031] (3) Leaf task node: The attributes of a leaf task node are: ① Node ID: the identifier of the task node; ② Parts information table: mainly related to basic information such as part name and type; ③ Quantity: how many parts of this type are planned to be produced in this task node, that is, how many parts can be produced in one day under the conditions determined by the personnel group, equipment group, and tooling group for this part type; ④ Time length: the time required to complete the task node; ⑤ Process method: the process method required to produce the part; ⑥ Parent node ID: the previous task node of this task node. From the perspective of the process flow sequence, there is a necessary precedence relationship between the two; if it is an initial task, its parent node is empty; ⑦ Bill of materials: contains information such as the type and quantity of materials required to produce this number of parts in this task node; ⑧ Priority: Task nodes may compete for limited production resources in the same time period. Set a sorting priority for the task nodes, and the ones with higher priorities are sorted first to avoid resource conflicts by staggering. The production of a batch of parts may be divided into multiple task nodes, and the priorities of these task nodes are the same. ⑨ Personnel group: the set of personnel responsible for executing the production task;

[0032] ⑩ Equipment group: the equipment set responsible for executing the production task; Tooling group: the tooling set required for the production task;

[0033] (4) Node relationship description: ① For each category of parts, the parts information tables in the category task node, process task node and leaf task node are the same. If there is a category task node i, there is a process task node j under the category task node i, and there is a leaf task node k in the process task node j, then the leaf task node k. parts information table = process task node j. parts information table = category task node i. parts information table; ② The number of parts in the process task nodes at all levels is equal, and is also equal to the number of parts in the category task node, and is equal to the sum of the number of parts in the leaf task nodes under each process. If there is a category task node i, there is a process task node j under the category task node i, and if the process task node j can be divided into n task nodes, k∈{1, ···, n}, then = process task node j. quantity = category task node i. quantity; ③ The process method set of each leaf task node is the process method set of the corresponding process task node, which is a subset of the process method set in the category task node. If there is a category task node i, under which there is a process task node j, and process task node j has a leaf task node k, then leaf task node k. process method set = process task node j. i. A set of process methods; ④ The time relationship between nodes is the core content of the scheduling algorithm calculation output; ⑤ The parent node of the leaf task node is the corresponding process task node, and the leaf task node has no child nodes; the process task node determines the parent-child node relationship according to the process sequence, and the process task nodes under the same part category have a parent-child relationship. The first executed process node is the child node, and the process executed in the next step is the parent node, and so on; in the algorithm, the last step process task node has no parent node, but in order to understand the division, the parent node of all category task nodes is the total task node; ⑥ The bill of materials of the category task node is decomposed into the bill of materials of the process task node, and the bill of materials of the process task node is structured into the bill of materials of the leaf task node. If there is a category task node i, there is a process task node j under the category task node i, and the process task node j has a leaf task node k, then the leaf task node k. j. i. Bill of materials; ⑦ The priorities of leaf task nodes under the same process are the same, and the representation of the priority of leaf task nodes is composed of the priorities of process task nodes and category task nodes. The priorities of process task nodes are arranged in process order, and the priorities of category task nodes are defined before algorithm calculation. When comparing the priorities of leaf task nodes, first compare the priorities of category task nodes, and then compare the priorities of process task nodes. If there is a category task node i, there is a process task node j under category task node i, and process task node j has a leaf task node k, then leaf task node k.priority = {category task node i.priority, process task node j.priority}; ⑧ The personnel group set of the leaf task node is a subset of the personnel group set of the corresponding process task node, and the personnel group set of each process task node is a subset of the personnel group set of the corresponding category task node. If there is a category task node i, there is a process task node j under category task node i, and process task node j has a leaf task node k, then leaf task node k. j. i. Personnel group set; ⑨ The equipment group set of the leaf task node is a subset of the equipment group set of the corresponding process task node, and the equipment group set of each process task node is a subset of the equipment group set of the corresponding category task node. If there is a category task node i, there is a process task node j under the category task node i, and the process task node j has a leaf task node k, then the leaf task node k. j. i. Equipment group set; ⑩ The tooling group set of the leaf task node is a subset of the tooling group set of the corresponding process task node, and the tooling group set of each process task node is a subset of the tooling group set of the corresponding category task node. If there is a category task node i, there is a process task node j under the category task node i, and the process task node j has a leaf task node k, then the leaf task node k. j. i. Tooling set collection

[0034] Specifically, the production scheduling algorithm process of some embodiments of the present disclosure is as follows: (1) Initial state, assuming:

[0035] ① The set of category task nodes is represented by A, and its number is l. There are category task nodes ai, ai∈A, i∈{1, ···, l}; ② The set of process task nodes under the category task node ai is represented by Bi, and its number is mi. There are process task nodes bij, bij∈Bi, j∈{1, ···, mi}; j is arranged according to the process, and the initial process is bi1; ③ The process task node bij can be divided into nij leaf task nodes, and the set of leaf task nodes is represented by Cij. There are leaf task nodes cijk, cijk∈Cij, k∈{1, ···, nij}; ④ The inventory of non-consumable resources, such as employee resources, equipment list, tooling inventory, etc.; ⑤ The inventory and guarantee plan of consumable resources, such as material inventory and guarantee plan, etc.; ⑥ Initialize the node set E, which is used to store the unscreened leaf task nodes a later; ⑦ Initialize the queue set F, each element in the queue set F is a queue f, and the element type in the queue is the leaf task node a. It is subsequently used to store leaf task nodes selected based on non-consumable resource constraints;⑧ Initialize the node set G, and use element g to record the selected leaf task node information. The attributes of g include round, node ID, start time, and end time;⑨ The enterprise's shift and rest time model;

[0036] The production start time is T0;

[0037] (2) Algorithm flow, ① such as Figure 3 As shown in the figure, the leaf task nodes that can be put into the node set E are the black nodes in the figure, which represent the leaf task nodes that can be started at the same time under the condition of unlimited capacity. ② If there is a conflict of non-consumable resources between some leaf nodes, the nodes that have the same non-consumable resources can be connected by dotted lines. ③ As Figure 4 As shown in the figure, the priorities of nodes that have the same non-consumable resource contention are sorted and grouped into queue f in order. Nodes without resource contention are put into queues separately. The specific numbers represent the priority of the nodes in each group, and 1 is the node with the highest priority. Resources will be allocated to each group of nodes based on the non-consumable resource inventory, and nodes that can meet the requirements will be screened out. The screened out nodes are marked in orange. The screened out nodes are put into the node set G. ④ As shown in the figure Figure 5 As shown in Figure 6, according to the priority and limited consumable resources, the nodes in the node set G that cannot meet the consumable resource allocation are screened out and placed in the node set E; and the remaining nodes in the queue set F are placed back in the node set E. ⑤ As shown in Figure 6, the next executable process task node is split and its leaf task node is placed in the node set E. The subsequent algorithm returns to the first step and starts the second round. Finally, Figure 7As shown, after multiple rounds of iterations, all leaf task nodes split from the process task nodes are placed in the node set G. The g nodes in the node set G only record the round in which the nodes are placed, the node ID, the start time, and the end time.

[0038] 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 and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A part-level production scheduling method based on forest node segmentation and sorting, characterized in that: The following steps are involved: Step 1. Set the forest data structure composed of all task process task nodes, input the non-consumable production resource inventory, the consumable production resource inventory, the enterprise shift and rest time model, set the production cycle time, initialize the intermediate process variable node set E, the queue set F, and the output node set G; Step 2. Put the leaf task node c of the executable process split into the node set E, and determine whether the number of leaf task nodes c in the node set E is 0. If so, proceed to step 7; if not, take T0 as the initial time of the leaf task node c in the node set E, take T0 + the time length of the leaf task node c as the end time of the leaf task node c, and determine whether there is any contention for non-consumable production resources between the leaf task nodes c in the node set E in the same time period. If so, construct a queue f for the leaf task nodes c with the same resource contention. In each queue f, arrange the priorities of the leaf task nodes c in descending order, and store the nodes in the queue in descending order according to the priority, and proceed to step 3. If not, store the nodes without resource contention in a separate queue f. Step 3. According to the first-in-first-out principle of the queue, and based on the stock of non-consumable resources, multiple dequeue operations are performed on each queue f, and non-consumable resources are allocated to each node until the non-consumable resources are allocated, and then step 4 is entered; Step 4. Put the outgoing leaf task node c into the output node set G. G records the round, node ID, start time, and end time of the leaf task node c, and determines whether the end time of the leaf task node c exceeds the production cycle time. If so, proceed to step 5. If not, query the material inventory to determine whether the resource consumption demand of each task node in the queue set F can be met within the guaranteed time. If it can be met, the inventory number is subtracted from the demand number, and the inventory is updated. If not, record the leaf task node c that cannot be met, remove the leaf task node c that cannot be met from the output node set G, and put it back into the node set E. Step 5. Put the remaining nodes in the queue set F back into the node set E, and determine whether there is a next process task node. If so, find the parent node of the process task node corresponding to the leaf task node c in the output node set G and go to step 6. If not, go to step 6; Step 6. Update the start time of the next round of leaf task node c and return to step 2; Step 7. Output node set G; Step 8. Repeat steps 1-6 above until the execution is completed.

2. The part-level production scheduling method based on forest node segmentation and sorting according to claim 1 is characterized in that: The step 1 also includes: the element in the queue set F is queue f, and the node in the output node set G is g.

3. The part-level production scheduling method based on forest node segmentation and sorting according to claim 2 is characterized in that: The element type in the queue f is a leaf task node c, which is used to store the leaf task nodes c selected according to non-consumable resource constraints.

4. The method for part-level production scheduling based on forest node segmentation and sorting according to claim 2, characterized in that: The node set G uses g to record the filtered leaf task node information, and the attributes of g include round, node ID, start time, and end time.

5. The method for part-level production scheduling based on forest node segmentation and sorting according to claim 2, characterized in that: The node set E is used to store unscreened leaf task nodes c.

Images