CN-121543999-B - Robot production scheduling method and system based on task priority

CN121543999BCN 121543999 BCN121543999 BCN 121543999BCN-121543999-B

Abstract

The invention belongs to the technical field of industrial intelligent management, and particularly relates to a robot production scheduling method and system based on task priority; the method comprises the steps of obtaining and arranging production task information, constructing a task scheduling topological graph containing priority, time sequence and resource constraint, generating a plurality of feasible scheduling paths by using a graph searching algorithm, performing simulation execution to identify bottleneck tasks, dynamically adjusting the priority of the bottleneck tasks based on a delay scale factor, performing iterative optimization to generate a final scheduling scheme, and issuing corresponding scheduling instructions to a robot for execution. The corresponding system comprises a task acquisition module, a graph construction module, a path search module, a scheme evaluation module, a scheme optimization module and an instruction issuing module, and achieves high-priority task priority response and overall task delay minimization.

Inventors

LI TAOTAO
SHAN KAIYUAN
SHEN MENGJIE
ZHU WEN
Zuo Minna

Assignees

杭州硕泰科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260119

Claims (8)

1. A method for scheduling production of a robot based on task priority, comprising: s1, acquiring production task information, namely acquiring the production task information and sequencing priorities to form a task queue to be scheduled; the production task information comprises a task number, a task predicted starting time, a task predicted finishing time, task time consumption and task priority, and the task sets are prioritized according to the task time consumption and the task priority to form a task queue to be scheduled; the task priority comprises five levels, namely a first-level emergency task, a second-level high-priority task, a third-level medium-high-priority task, a fourth-level general task and a fifth-level low-high-priority task, wherein the task queues to be scheduled are ordered according to the following rules, namely the task queues to be scheduled are ordered from high to low according to priority among different levels; S2, constructing a task scheduling topological graph, namely constructing the task scheduling topological graph according to the task queue to be scheduled and the robot resource pool; The topological graph is a directed acyclic graph and comprises nodes and edges, wherein the nodes correspond to production tasks, the edges represent task dependency relationships, and three constraint conditions of priority constraint, time sequence constraint and resource constraint are simultaneously considered during construction, wherein the priority constraint refers to that a high-priority task cannot be arranged after a low-priority task, the time sequence constraint refers to that a safe interval time is required to be reserved between continuous tasks of the same robot, and the resource constraint refers to that a single robot cannot execute a plurality of tasks simultaneously in any period; s3, searching feasible scheduling paths and recording execution schemes, namely searching all the feasible paths meeting constraints on a task scheduling topological graph by adopting a graph searching algorithm to obtain a plurality of scheduling schemes, and recording task execution sequences and estimated execution total time corresponding to each scheme; S4, simulating execution and identifying bottleneck task nodes, namely arranging the feasible paths according to the estimated total execution time ascending sequence, and selecting a scheduling scheme corresponding to the first K paths from the feasible paths for simulating execution to obtain a virtual execution record; S5, optimizing bottleneck task nodes and generating a scheduling scheme in an iterative mode, namely adjusting the execution starting time of the bottleneck task nodes to the forefront position of an available time period according to the virtual execution record, regenerating a task queue to be scheduled, performing optimization iteration until the overall task delay is minimized and the utilization rate of the robot meets a preset threshold value, and outputting a final task scheduling scheme; And S6, generating a scheduling instruction set and issuing and executing, namely generating the scheduling instruction set according to a final task scheduling scheme, wherein the scheduling instruction set comprises a task execution sequence, a robot number, a starting time and target station information, and issuing the scheduling instruction to a corresponding robot to execute a task.
2. The method of task priority based robotic production scheduling of claim 1, wherein: in S3, the graph search algorithm employs a breadth-first search algorithm, and each node expands only if resource constraints are met, and no loop structure is included in the feasible paths.
3. The method of task priority based robotic production scheduling of claim 1, wherein: In S4, the virtual execution record comprises a task execution sequence, a robot utilization rate, a task delay condition and a bottleneck task identifier; The bottleneck task identification in the virtual execution record is determined according to the rule that task nodes with task delay time longer than the median of all task delay time exist in a task scheduling path, and at least one task node with time consuming time for the task to be executed, which is longer than the task itself due to resource occupation, exists in a task sequence behind the task nodes.
4. The method of task priority based robotic production scheduling of claim 1, wherein: in S5, if no new bottleneck task node is detected in the optimization iteration process, and the overall task delay drop amplitude is smaller than the preset convergence threshold, stopping the optimization iteration and outputting the final scheme.
5. The method of task priority based robotic production scheduling of claim 1, wherein: after the task scheduling scheme is generated, the following steps are added: s7, monitoring the execution state of task scheduling, wherein the monitored parameters comprise a task completion identifier, actual starting time, actual time consumption and robot state feedback information; S8, performing dynamic rearrangement after abnormality, namely if abnormal task execution or robot occurrence faults are monitored, regenerating a task queue to be scheduled according to the completed task and incomplete task information, and performing task priority sequencing and scheduling again, wherein the incomplete task needs to be inserted into the head of a new queue preferentially.
6. The method of task priority based robotic production scheduling of claim 1, wherein: in S2, the task scheduling topology graph adopts a directed acyclic graph structure, wherein nodes are marked as task numbers, the weight of edges is the expected completion time of the task, the degree of entry of each node does not exceed M, and M is the number of robots that can be simultaneously controlled in the same time period.
7. The method of task priority based robotic production scheduling of claim 1, wherein: In S4, the feasible paths are sorted in ascending order according to the estimated total execution time, where K is a positive integer, and the value of K is dynamically determined according to the distribution of the priorities of the current tasks, and the higher the duty ratio of the task with high priority is, the smaller the value of K is.
8. A system for robotic production scheduling based on task priority for implementing the method of any of claims 1-7, comprising: the task acquisition module is used for acquiring production task information and sequencing according to task priority and time consumption to form a task queue to be scheduled; The diagram construction module is used for constructing a task scheduling topological diagram according to the task queue to be scheduled and the robot resource pool; The path searching module is used for executing a graph searching algorithm on the task scheduling topological graph to search all feasible paths so as to form a scheduling scheme set; the scheme evaluation module is used for performing virtual execution on the feasible paths and recording bottleneck tasks and key indexes; The scheme optimizing module is used for adjusting and optimizing iteration of the task starting time based on the bottleneck task record and outputting a final task scheduling scheme; the instruction issuing module is used for generating a scheduling instruction set containing the task sequence, the robot number and the execution parameters, and issuing the set to the corresponding robot to execute the task.

Description

Robot production scheduling method and system based on task priority Technical Field The invention belongs to the technical field of industrial intelligent management, and particularly relates to a robot production scheduling method and system based on task priority. Background In modern intelligent manufacturing environments, robotic systems are widely used in various links of a production line, such as assembly, handling, inspection, and other tasks. Along with the continuous improvement of the structural complexity and order flexibility of the product, how to effectively schedule various production tasks executed by the robot becomes a key factor affecting the production efficiency and the resource utilization rate. The traditional static scheduling method is difficult to adapt to the production requirement of dynamic change, and the intellectualization, instantaneity and expandability of a scheduling system become research hotspots. The existing robot production scheduling method mostly adopts technical paths based on heuristic rules, genetic algorithms, ant colony algorithms or discrete event simulation modeling and the like to realize task allocation and time scheduling. And modeling the precedence dependence relationship among the tasks by constructing a task graph model by a part of the system, and solving the optimal scheduling path of the task by adopting an optimization algorithm. However, these methods mostly ignore the priority relation between tasks, and often take the minimum total time or the device idle rate as an objective function, and fail to embody the priority execution requirement of the critical tasks. Meanwhile, when the practical problems of multi-robot cooperation, task sudden change, temporary occupation of resources and the like are faced, the flexibility and the strain capacity are lacked, and the problems of task conflict, resource contention or scheduling failure and the like are easy to occur. In addition, part of the scheduling system depends on a centralized control architecture, single-point bottlenecks exist, and real-time response and concurrency coordination under multi-station and heterogeneous task scenes are difficult to adapt. Therefore, a method for dynamically identifying task priorities and adjusting scheduling strategies accordingly is needed to improve the scheduling efficiency and adaptability of the robot system in complex production environments. Disclosure of Invention In view of the above problems, the present invention provides a method for scheduling production of a robot based on task priority, comprising the following steps: S1, acquiring production task information, wherein the production task information comprises task numbers, task estimated starting time, task estimated finishing time, task time consumption and task priority, and the task sets are prioritized according to the task time consumption and the task priority to form a task queue to be scheduled; s2, constructing a task scheduling topological graph, namely constructing the task scheduling topological graph according to a task queue to be scheduled and a robot resource pool, wherein the topological graph is a directed acyclic graph and comprises nodes and edges, the nodes correspond to production tasks, the edges represent task dependency relationships, and three constraint conditions of priority constraint, time sequence constraint and resource constraint are simultaneously considered during construction; s3, searching feasible scheduling paths and recording execution schemes, namely searching all the feasible paths meeting constraints on a task scheduling topological graph by adopting a graph searching algorithm to obtain a plurality of scheduling schemes, and recording task execution sequences and estimated execution total time corresponding to each scheme; S4, simulating execution and identifying bottleneck task nodes, namely arranging the feasible paths according to the estimated total execution time ascending order, and selecting a scheduling scheme corresponding to the previous K paths from the feasible paths to perform simulation execution to obtain a virtual execution record, wherein the virtual execution record comprises a task execution sequence, a robot utilization rate, a task delay condition and bottleneck task identification; S5, optimizing bottleneck task nodes and generating a scheduling scheme in an iterative mode, namely adjusting the execution starting time of the bottleneck task nodes to the forefront position of an available time period according to the virtual execution record, regenerating a task queue to be scheduled, performing optimization iteration until the overall task delay is minimized and the utilization rate of the robot meets a preset threshold value, and outputting a final task scheduling scheme; And S6, generating a scheduling instruction set and issuing and executing, namely generating the scheduling instruction set according to a final task schedulin