Mechanical product assembly task allocation method and system

Abstract

The present application relates to the technical field of assembly task allocation, and discloses a mechanical product assembly task allocation method and system, which are used to solve the problems that when mechanical product assembly task allocation is performed, multiple available schemes exist, the optimal scheme cannot be accurately selected, and the scheme cannot be adjusted in real time according to the change of the field state, and comprise the following steps: calculating and evaluating qualified task allocation scheme sets in all preliminary task allocation schemes, calculating a corresponding multi-dimensional normalized margin value for each qualified task allocation scheme in the qualified task allocation scheme sets, and selecting a final task allocation scheme; performing conflict detection on the final task allocation scheme, if there is a conflict, modifying the final task allocation scheme to obtain a final modified task allocation scheme; and if there is no conflict item, directly regarding the final task allocation scheme as the final modified task allocation scheme, thereby effectively improving the continuity and intelligence of the mechanical product assembly process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of assembly task allocation, and more particularly to a mechanical product assembly task allocation method and system. BACKGROUND

[0002] Mechanical automated assembly systems, as important supporting equipment for modern manufacturing industry, have been widely used in the fields of automobile, electronics, precision instruments and intelligent manufacturing. Such systems usually rely on multiple functional modules such as mechanical arm execution, workpiece supply, visual detection, control scheduling and human-computer interaction to realize full-process automation operation from part transmission and positioning to precision assembly. Through high-precision execution of the mechanical arm and real-time detection of the sensor, the assembly system can guarantee the assembly efficiency and consistency in mass production scenarios, providing a key technical foundation for enterprises to improve production capacity and reduce labor costs.

[0003] In the process of executing assembly tasks, in order to realize closed-loop control from workpiece supply, mechanical arm operation, position detection to assembly completion and quality checking, the existing system needs to have the ability to coordinate and respond to assembly sequence, assembly tempo and process parameters, so as to ensure high-precision docking between parts. With the continuous development of concepts such as "task scheduling", "multi-resource collaboration" and "dynamic feedback", mechanical automated assembly systems are gradually evolving towards the direction of flexible task allocation, real-time exception response and state adaptation.

[0004] In recent years, the existing technology generally generates candidate allocation schemes for assembly tasks based on priority rules, resource idle states or preset matching strategies, and selects one scheme as the final execution result. Some methods combine historical data or experience models to optimize the resource scheduling process in order to improve the overall efficiency and success rate of assembly scheduling.

[0005] However, the above-mentioned technologies at least have the following technical problems:

[0006] In the actual execution process of assembly tasks, when there are two or more available schemes at the same time, there is often a lack of clear evaluation basis and systematic comparison mechanism, making it difficult to accurately determine which scheme has higher execution priority and stability. In addition, after the task is issued, if the on-site state changes, most of the existing systems lack the ability to automatically adjust or replace the original allocation scheme according to real-time conditions, resulting in task execution interruption, frequent manual intervention or low system recovery efficiency, which is not conducive to the continuity and intelligent operation of the assembly process. SUMMARY

[0007] In order to overcome the above-mentioned defects of the existing technology, the present application provides a mechanical product assembly task allocation method and system to solve the problems existing in the background art.

[0008] To achieve the above object, the present application provides the following technical solutions:

[0009] A mechanical product assembly task allocation method, comprising the following steps: step 1: receiving assembly task data, the assembly task data including target product model, target product process list, target product structure information and target product planned completion time; step 2: converting the target product structure information in the assembly task data into several assembly processes, extracting the assembly time required by each assembly process, resource type and sequence relationship, and generating a process task list; step 3: collecting assembly resource state information in the current production site, the assembly resources including assembly stations, assembly equipment and personnel, and the state information including idle condition, executable process range and expected idle period; step 4: according to the process task list and the assembly resource state information, calling a preset task scheduling algorithm, matching the process task with the assembly resource, and generating a preliminary task allocation scheme for the assembly task; step 5: calculating and evaluating a qualified task allocation scheme set from all preliminary task allocation schemes, calculating the corresponding multi-dimensional normalized margin value for each qualified task allocation scheme in the qualified task allocation scheme set, and selecting a final task allocation scheme according to the multi-dimensional normalized margin value; step 6: performing conflict detection on the final task allocation scheme to identify station duplication conflict, resource conflict or task timing conflict problems, if there is a conflict, modifying the final task allocation scheme to obtain a final modified task allocation scheme, if there is no conflict item, directly regarding the final task allocation scheme as the final modified task allocation scheme; step 7: sending the final modified task allocation scheme to the corresponding assembly unit, and in the assembly task execution process, real-time monitoring the real-time state change of the multi-dimensional constraint parameters, when detecting that any constraint condition is destroyed or the executability lower bound value becomes negative, automatically triggering a task rollback mechanism, selecting a replacement current task allocation scheme in descending order of the executability lower bound value, reissuing and updating the task execution process.

[0010] Preferably, the obtaining step of the process task list is: reading the target product structure information field in the received assembly task data, the structure information including the part list, assembly nodes, part connection mode and its hierarchical structure, extracting each component composition unit according to the product structure tree, and binding the product number of the target product to all component composition units; based on the assembly sequence and connection logic between components, a product assembly dependent directed graph is constructed to record the upstream components relied on by each component node; according to the product assembly dependent directed graph, each assembly action is decomposed into an executable assembly process according to the component connection logic and operation constraints, and each assembly process is bound to a unique component pair and assembly action type, and the predecessor process number set is automatically identified as the dependency relationship between processes; for each assembly process, the standard assembly time required by the current process is extracted from the process database or the built-in standard process table according to the component complexity, assembly method and historical standard operation data, and the resource type required by the process is identified, including the required workstation type, equipment model, tool number and personnel skill level requirement; for each process, record its process number, product number, bound component pair, assembly action type, predecessor dependency relationship, standard assembly time and required resource type, and generate a structured process task list.

[0011] Preferably, the obtaining step of the preliminary task allocation scheme is: reading the standard assembly time, required resource type and predecessor dependency relationship of each process in the process task list, and obtaining the assembly resource state information of each assembly resource; for each process task, according to its required resource type, filter out the resource units that meet the resource type requirement and cover the process standard assembly time from the current assembly resource state information, record the resource number and available start time of each candidate resource, and call the preset task scheduling algorithm to execute the topological sorting and resource earliest available priority principle according to the predecessor dependency relationship between processes and the idle time window of candidate resources, and sequentially allocate process tasks to candidate resources, and record the expected start time and expected end time of the process, form a complete task allocation sequence, and obtain the preliminary task allocation scheme.

[0012] Preferably, the obtaining step of the qualified task allocation scheme set is: for each preliminary task allocation scheme, obtaining multi-dimensional constraint parameters related to assembly feasibility, including jig opening range, maximum part size, minimum safety gap threshold, target tact time and historical Q90 tact value, process end pose and maximum reachable boundary distance of robot, number of matched personnel qualifications and number of personnel qualifications required by the process; check the compliance of each multi-dimensional constraint parameter, eliminate the preliminary task allocation scheme that does not meet the basic constraint condition, obtain the qualified task allocation scheme, and integrate all qualified task allocation schemes to obtain the qualified task allocation scheme set.

[0013] Preferably, the step of obtaining the multi-dimensional normalized margin value is: obtaining the range of the jig opening used in each process in the current assembly task, the range of the jig opening including the maximum opening size and the minimum opening size of the jig, and obtaining the maximum part size, calculating the size margin value of the process by subtracting the maximum part size from the maximum opening size and the minimum opening size; obtaining the target cycle time and the historical Q90 cycle value of the process, calculating the cycle margin value by subtracting the target cycle time from the historical Q90 cycle value and then dividing by the target cycle time; extracting the last obstacle avoidance distance of the robot in the execution path, obtaining the minimum safety clearance threshold of the path, calculating the path margin value by subtracting the last obstacle avoidance distance from the minimum safety clearance threshold of the path and then dividing by the minimum safety clearance threshold of the path; obtaining the number of personnel qualifications required for the current process and the number of matching personnel qualifications on site, calculating the personnel margin value by subtracting the number of matching personnel qualifications on site from the number of personnel qualifications required for the current process and then dividing by the number of personnel qualifications required for the current process; obtaining the end pose of the process and the maximum reachable boundary distance and the actual required motion distance under the current robot kinematics model, calculating the robot reachable margin value by subtracting the actual required motion distance from the maximum reachable boundary distance and then dividing by the maximum reachable boundary distance; and performing vector normalization on the size margin value, the cycle margin value, the path margin value, the personnel margin value, and the robot reachable margin value to obtain the normalized size margin value, the normalized cycle margin value, the normalized path margin value, the normalized personnel margin value, and the normalized robot reachable margin value.

[0014] Preferably, the step of selecting the final task allocation scheme according to the multi-dimensional normalized margin value is: defining the minimum value in the multi-dimensional normalized margin value of each qualified task allocation scheme as the lower bound of the executability of the scheme; and selecting the qualified task allocation scheme with the maximum lower bound of executability as the final task allocation scheme.

[0015] Preferably, the step of obtaining the final modified task allocation scheme is: step 6.1: for each conflict item, extract the primary resource unit causing the conflict as the current conflict resource, wherein if it is a workstation overlap conflict, the current conflict resource is the overlapping workstation number; if it is a resource occupation conflict, the current conflict resource is the conflict resource number; if it is a task timing conflict, the current conflict resource is the resource number corresponding to the preceding process on the dependent path; step 6.2: for each conflict item, respectively obtain the number of conflict processes involved, the number of downstream dependent processes, the time conflict span, the number of alternative resources of the current conflict resource, and the number of the same type of resources in the total resource pool; calculate the conflict resource scarcity degree according to the number of alternative resources of the current conflict resource and the number of the same type of resources in the total resource pool; normalize the number of conflict processes, the number of downstream dependent processes, the time conflict span, and the conflict resource scarcity degree, and calculate the conflict priority index according to the normalized number of conflict processes, the number of downstream dependent processes, the time conflict span, and the conflict resource scarcity degree; step 6.3: for each conflict item, in its current conflict process, generate three feasible modification paths, namely the replacement modification path, the time modification path, and the order interchanging path; step 6.4: for each feasible modification path, extract the set of processes affected before and after the modification, calculate the start time difference for each process, and accumulate the disturbance value; step 6.5: among all the modifiable paths, select the path with the smallest disturbance value as the final modification method of the current conflict item, and apply it to the task allocation scheme to update the conflict item related fields; step 6.6: for the remaining conflict items, execute steps 6.3 to 6.5 in order from high to low according to the conflict priority index, until all conflicts are processed, and the final modified task allocation scheme is formed.

[0016] Preferably, the step of generating the three feasible modification paths of the replacement modification path, the time modification path and the sequence exchange path is: obtaining a conflict process set in the current conflict item, reading a current conflict resource number and a process standard assembly time bound to each process; according to the resource type of the conflict resource, screening out alternative resources consistent with the current conflict resource type and in an idle state within the process predicted start and end time range from the assembly resource state information to form an alternative resource set; for each conflict process, sequentially assigning it to each resource in the alternative resource set, if the assignment is successful, generating a set of replacement modification paths, and recording the resource number and the corresponding time period modified by the modification path; according to the conflict process set, extracting the precedence relationship and the standard assembly time of each process, under the premise of not changing the resource number, combining the process precedence relationship and the current resource state information, adjusting the predicted start time of the process forward or backward to avoid the occupied time period of the current conflict resource, and updating the predicted end time; judging whether the adjusted time meets the completion time requirement of the precedent process and the resource idle window condition, if it meets, generating a set of time modification paths, and recording the time period difference value adjusted; according to the conflict process set, screening out process pairs located at the same resource number and time with partial overlap but no precedence relationship, for each process pair meeting the condition, trying to exchange the scheduling sequence of it in the task allocation sequence, and recalculating the predicted start time and end time thereof; judging whether the exchanged processes can meet the precedence relationship of each process and the current resource idle time constraint, if it meets, generating a set of sequence exchange paths, and recording the start time change after adjustment.

[0017] Preferably, a mechanical product assembly task allocation system, the system comprising: an assembly task data receiving module for receiving assembly task data, the assembly task data comprising a target product model, a target product process list, target product structure information and a target product planned completion time; a process task list generation module for converting the target product structure information in the assembly task data into a plurality of assembly processes, extracting the assembly time required by each assembly process, the resource type and the sequence relationship, and generating a process task list; an assembly resource information collection module for collecting assembly resource state information in the current production site, the assembly resources including assembly stations, assembly equipment and personnel, and the state information including idle conditions, executable process ranges and estimated idle periods; a preliminary allocation scheme generation module for matching the process tasks and the assembly resources according to the process task list and the assembly resource state information, and generating a preliminary task allocation scheme for the assembly task by calling a preset task scheduling algorithm; a final scheme acquisition module for calculating and evaluating a set of qualified task allocation schemes from all preliminary task allocation schemes, calculating a corresponding multi-dimensional normalized margin value for each qualified task allocation scheme in the set of qualified task allocation schemes, and selecting a final task allocation scheme according to the multi-dimensional normalized margin value; a final scheme correction module for performing conflict detection on the final task allocation scheme, identifying station duplication conflicts, resource conflicts or task timing conflicts, and if there are conflicts, modifying the final task allocation scheme to obtain a final modified task allocation scheme; if there are no conflict items, directly regarding the final task allocation scheme as the final modified task allocation scheme; a cyclic adjustment module for sending the final modified task allocation scheme to the corresponding assembly unit, real-time monitoring the real-time state changes of the multi-dimensional constraint parameters, and when detecting that any constraint condition is destroyed or the executability lower bound value becomes negative, automatically triggering a task rollback mechanism, and selecting a replacement current task allocation scheme in descending order of the executability lower bound value to reissue and update the task execution process.

[0018] Technical effects and advantages of the present application:

[0019] In all preliminary task allocation schemes, a set of qualified task allocation schemes is calculated and evaluated, a multi-dimensional normalized margin value corresponding to each qualified task allocation scheme in the set of qualified task allocation schemes is calculated, and a final task allocation scheme is selected. The final task allocation scheme is subjected to conflict detection, if there are conflicts, the final task allocation scheme is modified to obtain a final modified task allocation scheme; if there are no conflict items, the final task allocation scheme is directly regarded as the final modified task allocation scheme, effectively improving the continuity and intelligence of the mechanical product assembly process. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1A mechanical product assembly task allocation method flow chart is provided for the embodiments of the present application.

[0021] Figure 2 A mechanical product assembly task allocation system structure diagram is provided for the embodiments of the present application. DETAILED DESCRIPTION

[0022] The technical solutions in the present application will be described clearly and completely in combination with the drawings in the present application, and in addition, the forms of each structure described in the following embodiments are only examples, and the mechanical product assembly task allocation method and system involved in the present application are not limited to each structure described in the following embodiments, and all other embodiments obtained by those skilled in the art without making creative efforts belong to the protection scope of the present application.

[0023] The present application provides a mechanical product assembly task allocation method, as shown in the accompanying drawings, comprising the following steps: Figure 1

[0024] Step 1: receiving assembly task data, the assembly task data comprising target product model, target product process list, target product structure information and target product planned completion time, serving as input basis for subsequent task allocation;

[0025] Step 2: converting the target product structure information in the assembly task data into several assembly processes, extracting the assembly time required by each assembly process, resource type and sequence relationship, and generating a process task list;

[0026] In the present embodiment, it needs to be specifically explained that the obtaining step of the process task list is:

[0027] In the received assembly task data, the target product structure information field is read, the structure information comprising a list of parts and components, assembly nodes, component connection modes and hierarchical structure, each component composition unit is extracted according to the product structure tree (such as BOM table or three-dimensional model topology information), and the product number of the target product is bound to all component composition units;

[0028] It needs to be specifically explained that the step of extracting each component composition unit is that, when processing according to the product structure tree, the three-dimensional model topology data of the target product is first read, the part and component nodes are identified one by one, and a node identification set is generated; then the nodes are numbered, a mapping table of parts and components to component units is established, and a structured set of component composition units is obtained.

[0029] Based on the assembly sequence and connection logic between components, a product assembly dependent directed graph is constructed, recording the upstream components depended on by each component node, serving as the generation basis of subsequent process dependent relationship; ​

[0030] Specifically, the step of constructing the product assembly dependency directed graph comprises the following steps: reading the assembly sequence and connection constraints between components while extracting each component unit, abstracting each connection relationship as a directed edge of "starting component-target component", and forming an edge set; combining the node identifier set and the edge set to generate the product assembly dependency directed graph. In the graph, each node represents a component unit, and each directed edge represents the assembly sequence or assembly logic.

[0031] According to the product assembly dependency directed graph, each assembly action is decomposed into an executable assembly process according to the component connection logic and operation constraints, and each assembly process is bound to a unique component pair and assembly action type (such as fitting, screwing, welding, buckling, etc.), and its predecessor process number set is automatically identified as the dependency relationship between processes.

[0032] Specifically, the step of decomposing each assembly action into an executable assembly process comprises the following steps: traversing each edge in the product assembly dependency directed graph, identifying the connection mode between components (such as fitting, welding, screwing, buckling, etc.), and mapping it to the process description of "component pair + action type"; then combining the operation constraints (such as process required resources, tools and skill requirements) to refine the assembly action into an executable process, and automatically generate the process number and predecessor dependency set, and finally form an executable process task set.

[0033] For each assembly process, according to the component complexity, assembly method and historical standard operation data, the standard assembly time required for the current process is extracted from the process database or the built-in standard process table in the system, and the resource type required for the process is identified, including the required workstation type, equipment model, tool number and personnel skill level requirement.

[0034] For each process, record its process number, product number, bound component pair, assembly action type, predecessor dependency relationship, standard assembly time and required resource type, and generate a structured process task list.

[0035] Step 3: Collect the assembly resource state information in the current production site, including assembly workstations, assembly equipment and personnel, and state information including idle condition, executable process range and expected idle period.

[0036] Step 4: According to the process task list and assembly resource state information, call the preset task scheduling algorithm to match the process task with the assembly resource, generate a preliminary task allocation scheme for the assembly task, and the preliminary task allocation scheme includes two or more task allocation scheme sets.

[0037] The preset task scheduling algorithm is a kind of scheduling method commonly used in the prior art for an automated assembly system, usually based on static or semi-dynamic information such as process execution order, resource idle state and task priority, and according to preset rules or heuristic strategies, the matching relationship between process tasks and assembly resources is calculated to generate one or more preliminary task allocation schemes that meet the scheduling constraint conditions. Common task scheduling algorithms include but are not limited to greedy matching algorithm based on priority queue, sliding scheduling algorithm based on resource time window, topological sorting algorithm based on graph structure, constraint matching algorithm based on rule ontology, etc.

[0038] In this embodiment, it needs to be specifically explained that the preliminary task allocation scheme is obtained by:

[0039] The standard assembly time, required resource type and predecessor dependency relationship of each process in the process task list are read, and the assembly resource state information of each assembly resource is obtained;

[0040] For each process task, according to its required resource type, the resource unit that meets the resource type requirement and the predicted idle time period covering the process standard assembly time is filtered from the current assembly resource state information, which is recorded as a candidate resource, and the resource number and available start time of each candidate resource are recorded;

[0041] The preset task scheduling algorithm is called, and according to the predecessor dependency relationship between processes and the idle time window of the candidate resource, the topological sorting and the earliest available resource priority principle are executed, the process tasks are sequentially allocated to the candidate resources, and the predicted start time and predicted end time of the process are recorded to form a complete task allocation sequence as a preliminary task allocation scheme.

[0042] The earliest available resource priority principle is that when allocating process tasks, the available resource with the earliest idle time is preferred to undertake the task. This principle compares the start points of the available time windows of each candidate resource to determine which resource can start executing the process earliest under the premise of meeting the process requirements, thereby minimizing the overall assembly time and maximizing the resource utilization.

[0043] Step 5: Calculate and evaluate the qualified task allocation scheme set from all preliminary task allocation schemes, calculate the corresponding multi-dimensional normalized margin value for each qualified task allocation scheme in the qualified task allocation scheme set, and the multi-dimensional normalized margin value includes but is not limited to size margin value, beat margin value, path margin value, personnel margin value and robot reachable margin value, etc., and select the final task allocation scheme according to the multi-dimensional normalized margin value;

[0044] In this embodiment, it needs to be specifically explained that the qualified task allocation scheme set is obtained by:

[0045] For each preliminary task allocation scheme, obtain multi-dimensional constraint parameters related to assembly feasibility, including fixture opening range, maximum part size, path minimum safety clearance threshold, target cycle time and historical Q90 cycle value, end pose of process and maximum reachable boundary distance of robot, number of matchable personnel qualifications and number of personnel qualifications required for the process, etc.

[0046] The historical Q90 cycle value refers to the 90th percentile obtained by statistical analysis of the cycle time of a certain process or operation in historical assembly execution data. That is, 90% of the data is less than or equal to this value, and only 10% of the execution time exceeds this value.

[0047] It should be specifically pointed out that the fixture opening range is directly read from the fixture design drawing or fixture equipment parameter table;

[0048] The maximum part size is obtained from the part size field in the product process database;

[0049] The path minimum safety clearance threshold is obtained by extracting the key point coordinates of the process path from the trajectory data generated by the robot offline programming software, and calculating the minimum Euclidean distance between the workpiece envelope and the surrounding obstacles as the path minimum safety clearance threshold;

[0050] The target cycle time is obtained by directly reading the process planning time field issued by the assembly task system;

[0051] The historical Q90 cycle is obtained by statistical analysis of the cycle data of the corresponding process in the historical process database and taking the 90th percentile;

[0052] The end pose of the process is obtained by calling the forward kinematics of the robot controller to obtain the three-dimensional position and attitude parameters of the end point of the process as the end pose of the process;

[0053] The maximum reachable boundary distance of the robot is obtained from the robot body parameter file or the kinematics specification provided by the manufacturer;

[0054] The number of matchable personnel qualifications is obtained by statistical analysis of the qualification level information recorded in the production site personnel management system and the current online personnel state;

[0055] The number of personnel qualifications required for the process is obtained from the personnel configuration requirements in the process specification or standard operating procedure.

[0056] Check the compliance of each multi-dimensional constraint parameter, eliminate preliminary task allocation schemes that do not meet the basic constraint conditions, and obtain qualified task allocation schemes. Integrate all qualified task allocation schemes to obtain a set of qualified task allocation schemes.

[0057] Need to be specific, the compliance of each item of the multi-dimensional constraint parameter includes: size constraint, that is, the size of the part must be smaller than the maximum opening of the fixture and larger than the minimum opening; time constraint, that is, the process is expected to complete within the target cycle time, and the deviation from the historical Q90 value is not more than the set proportion; path constraint, that is, the obstacle avoidance distance must be greater than the minimum safety gap threshold of the path; personnel qualification constraint, that is, the actual number of personnel with qualifications must not be less than the number of personnel required by the process; robot accessibility constraint, that is, the distance of the end pose must not exceed the maximum reachable boundary. If any parameter in the initial task allocation scheme does not meet the above constraint conditions, it is determined that the scheme is unqualified.

[0058] In this embodiment, it needs to be specified that the acquisition step of the multi-dimensional normalized margin value is:

[0059] The opening range of the fixture used in each process of the current assembly task is obtained, the opening range of the fixture includes the maximum opening size and the minimum opening size of the fixture, and the maximum part size is obtained. The maximum opening size is subtracted from the maximum part size, and the minimum opening size is subtracted from the maximum opening size to calculate the ratio, and the size margin value of the process is obtained, for example: if the maximum opening of a fixture is 40mm, the minimum opening is 20mm, and the part width is 32mm, the size margin value calculation step is:

[0060] ;

[0061] The size margin value is represented as;

[0062] The target cycle time and the historical Q90 cycle value of the process are obtained, and the historical Q90 cycle value is subtracted from the target cycle time. The ratio of the target cycle time is calculated to obtain the cycle margin value;

[0063] The last obstacle avoidance distance of the robot in the execution path is extracted, the minimum safety gap threshold of the path is obtained, and the last obstacle avoidance distance is subtracted from the minimum safety gap threshold of the path. The ratio of the minimum safety gap threshold of the path is calculated to obtain the path margin value;

[0064] Need to be specific, the last obstacle avoidance distance is recorded in real time by the robot at the last obstacle avoidance event, and the minimum spatial distance between the robot end and the obstacle is taken as the last obstacle avoidance distance.

[0065] The number of personnel with qualifications required by the current process and the number of personnel with qualifications that can be matched on site are obtained, and the number of personnel with qualifications that can be matched on site is subtracted from the number of personnel with qualifications required by the current process. The ratio of the number of personnel with qualifications required by the current process is calculated to obtain the personnel margin value;

[0066] The maximum reachable boundary distance under the current robot kinematics model is subtracted from the actual required motion distance, and the ratio of the result to the maximum reachable boundary distance is calculated to obtain a robot reachable margin value;

[0067] The size margin value, the beat margin value, the path margin value, the personnel margin value, and the robot reachable margin value are subjected to vector normalization processing to obtain normalized size margin value, beat margin value, path margin value, personnel margin value, and robot reachable margin value. It needs to be specifically pointed out that vector normalization processing is prior art, and the specific steps thereof will not be described in detail in this embodiment.

[0068] In this embodiment, it needs to be specifically pointed out that the step of selecting a final task allocation scheme according to the multi-dimensional normalized margin value is:

[0069] The minimum value in the multi-dimensional normalized margin value of each qualified task allocation scheme is defined as the executability lower bound value of the scheme.

[0070] The qualified task allocation scheme with the maximum executability lower bound value is selected as the final task allocation scheme.

[0071] By extracting the minimum value in the multi-dimensional normalized margin value of each qualified task allocation scheme as the executability lower bound value of the scheme, and selecting the scheme with the maximum lower bound value as the final execution object, the overall assembly process can still have maximum execution reliability at the weakest link. Compared with the traditional weighted average or subjective sorting method, this method preferentially selects the "most robust" execution path in the qualified task allocation scheme, strengthens the anti-interference ability of the task scheme from the bottom line safety perspective, and avoids the overall assembly failure caused by resource shortage or condition approaching threshold in individual links.

[0072] Step 6: Perform conflict detection on the final task allocation scheme to identify conflicts such as station repetition conflict, resource conflict, or task timing conflict. If there is a conflict, the final task allocation scheme is modified to obtain a final modified task allocation scheme. If there is no conflict item, the final task allocation scheme is directly regarded as the final modified task allocation scheme.

[0073] It needs to be specifically pointed out that the station repetition conflict is the case that two or more processes are assigned to the same station and the time intervals overlap; the resource occupation conflict is the case that the same resource is occupied by multiple processes in the same time period; and the task timing conflict is the case that the planned start time of any subsequent process is earlier than the end time of its dependent previous process.

[0074] Although the final task allocation scheme has been selected from multiple qualified task allocation schemes through multi-dimensional standardized margin evaluation, and has better scheduling stability and resource matching degree, systematic conflict detection is still needed before actual execution. This is because the assembly station, resource time period and process dependency involved in the task allocation process have complex intersections in the multi-dimensional space. Preliminary evaluation stage may not be able to fully capture all potential conflicts, such as time overlap of station resources, multiple occupation of resources or inconsistency of process timing. By introducing a conflict detection mechanism at this stage, potential scheduling conflicts can be actively discovered and corrected before task execution, thereby ensuring the logical consistency and continuity of the assembly process, significantly reducing the risk of task interruption, manual intervention or rework due to conflicts, and improving the intelligence and automatic recovery capability of the system.

[0075] In this embodiment, it needs to be specifically pointed out that the step of obtaining the final modified task allocation scheme is:

[0076] Step 6.1: For each conflict item, extract the main resource unit causing the conflict as the current conflict resource. If it is a station repetition conflict, the current conflict resource is the overlapping station number. If it is a resource occupation conflict, the current conflict resource is the conflict resource number. If it is a task timing conflict, the current conflict resource is the resource number corresponding to the previous process on the dependency path.

[0077] Step 6.2: For each conflict item, obtain the number of conflict processes involved, the number of downstream dependent processes, the time conflict span (i.e. the length of the overlapping time), the number of alternative resources of the current conflict resource, and the number of the same type of resources in the total resource pool (obtained from the system resource table).

[0078] It needs to be specifically pointed out that the number of alternative resources of the current conflict resource refers to the number of other resources that are in an idle state and have the same behavior ability in the conflict time period under the resource category to which the conflict resource belongs in the current conflict item. For example, in the case of "resource occupation conflict", if a robot R1 is occupied by two processes A and B at the same time, the system identifies the robot R1 as the current conflict resource. At this time, the system further searches the resource pool to find whether there are other robots R2, R3, etc. that are in an idle state in the time period and have the same assembly ability (such as gripping force, reachable range, tool interface, etc.) as R1. If R2 and R3 both meet the conditions, the number of alternative resources of the current conflict resource is 2.

[0079] The conflict resource scarcity degree is calculated according to the number of alternative resources of the current conflict resource and the number of the same type of resources in the total resource pool. The specific obtaining steps are:

[0080] ;

[0081] In the formula, is expressed as the conflict resource scarcity degree, is expressed as the number of alternative resources of the current conflict resource, is expressed as the number of the same type of resources in the total resource pool;

[0082] The conflict process number, the number of downstream dependent processes, the time conflict span, and the conflict resource scarcity degree are normalized. Specifically, the four types of conflict evaluation parameters are combined to form a four-dimensional vector. The norm value of the vector is obtained by calculating the square sum of each component and then taking the square root. Each parameter component is divided by the norm value, and the normalization process is completed. Through the normalization operation, the conflict indicators from different dimensions and different sources have the same order of magnitude in the unified scoring system, avoiding evaluation imbalance or priority deviation caused by parameter value scale differences. It should be noted that the vector normalization method is a prior art, and its processing principle has been widely disclosed and applied in the field of multi-index decision, intelligent scheduling and pattern recognition. Therefore, this embodiment will not describe the specific algorithm process. The conflict priority index is calculated based on the normalized conflict process number, the number of downstream dependent processes, the time conflict span, and the conflict resource scarcity degree. The specific acquisition steps are as follows:

[0083] ;

[0084] In the formula, is expressed as the conflict priority index, is expressed as the normalized number of conflict processes involved, is expressed as the normalized number of downstream dependent processes, is expressed as the normalized time conflict span, is expressed as the normalized conflict resource scarcity degree, , , , is expressed as the weight coefficient of the normalized number of conflict processes involved, the weight coefficient of the normalized number of downstream dependent processes, the weight coefficient of the normalized time conflict span, and the weight coefficient of the normalized conflict resource scarcity degree, , , , obtained by the analytic hierarchy process, and , the analytic hierarchy process (AHP) is a system analysis method based on multi-criteria decision theory. By constructing a multi-level decision structure model, complex problems are decomposed into target layer, criterion layer and index layer, and the relative importance between factors is obtained through pairwise comparison. Specifically, the evaluator constructs a pairwise comparison matrix based on expert experience or actual scene judgment, and then performs consistency check and eigenvalue calculation to extract the normalized weight coefficients of each index relative to the upper layer target.

[0085] It should be specifically pointed out that when using the analytic hierarchy process, first determine the target layer as the "conflict priority index", the criterion layer includes the number of conflict processes involved, the number of downstream dependent processes, the time conflict span and the degree of conflict resource scarcity, and a 4x4 pairwise comparison matrix is constructed. Then calculate the maximum eigenvalue of the matrix and perform consistency check. If the check is passed, extract the corresponding eigenvector and normalize it to obtain the weight coefficients a1, a2, a3, a4.

[0086] Step 6.3: For each conflict item, in its current conflict process, generate three feasible modification paths: replacement modification path, time modification path and sequence exchange path;

[0087] It should be specifically pointed out that the replacement modification path is to replace the current resource with a free resource of the same type within the non-conflict period; the time modification path is to adjust the starting time of the current process within the earliest / latest time range allowed by the resource; and the sequence exchange path is to exchange the execution order with other processes in the same station without violating the process dependency constraint.

[0088] Step 6.4: For each feasible modification path, extract the set of processes affected before and after modification, calculate the starting time difference for each process, and accumulate the disturbance value;

[0089] Step 6.5: In all modifiable paths, select the path with the smallest disturbance value as the final modification method for the current conflict item, and apply it to the task allocation scheme to update the conflict item related fields;

[0090] It should be specifically pointed out that the conflict item related fields include resource number, process expected start time, process expected end time and predecessor dependency.

[0091] It should be specifically pointed out that the specific steps for updating the conflict item related fields are as follows: when the conflict process is assigned to a new resource, replace the original bound resource number; when the process execution order or start and end time is adjusted, modify the start and end time records of the process; if the modification path causes the process order to change, update the predecessor dependency set of the process in the product assembly dependency directed graph to ensure that the dependency logic is consistent with the latest scheduling scheme.

[0092] Step 6.6: Steps 6.3 to 6.5 are performed in order of conflict priority index from high to low for the remaining conflict items until all conflicts are processed, forming the final modified task allocation scheme.

[0093] By constructing the conflict priority index, multiple conflict dimensions are standardized and quantified, and the minimum disturbance correction path is generated based on the priority, which optimizes each item from replacing resources, adjusting time to exchanging order. This mechanism not only systematically identifies and eliminates potential scheduling conflicts, but also maximizes the original task planning structure, reduces the impact on the overall execution rhythm, and improves the stability, feasibility and execution efficiency of task allocation.

[0094] In this embodiment, it needs to be specifically pointed out that the steps of generating three feasible correction paths of replacement correction path, time correction path and order exchange path are:

[0095] Obtain the conflict process set in the current conflict item, read the current conflict resource number and process standard assembly time bound to each process;

[0096] According to the resource type of the conflict resource, filter out the replaceable resources that are consistent with the type of the current conflict resource and in idle state within the process expected start and end time range from the assembly resource state information to form a replaceable resource set;

[0097] For each conflict process, assign it to each resource in the replaceable resource set in turn, if the assignment is successful, generate a set of replacement correction paths, and record the resource number and corresponding time period modified by the correction path;

[0098] It needs to be specifically pointed out that the steps of generating a set of replacement correction paths are, first, obtain the conflict process set in the current conflict item, read the conflict resource number and process standard assembly time bound to each process; then, according to the resource type of the conflict resource, filter out the candidate resources that are consistent with the type and in idle state within the process expected start and end time range from the assembly resource state information to form a replaceable resource set; then try to assign the conflict process to the replaceable resource in turn, if the assignment is successful, generate a set of replacement correction paths, and record the resource number and corresponding time period modified by the correction path.

[0099] According to the conflict process set, extract the precedence relationship and standard assembly time of each process, without changing the resource number, combine the process precedence relationship and the current resource state information, adjust the expected start time of the process forward or backward to avoid the occupied time period of the current conflict resource, and update the expected end time;

[0100] determining whether the adjusted time meets the completion time requirement of the precedent procedure and the resource idle window condition, and if so, generating a set of time correction paths and recording the adjusted time period difference value;

[0101] Specifically, the step of generating a set of time correction paths is to extract the precedent dependency relationship of the conflict procedure and the standard assembly time first, and under the premise of not changing the resource number, the expected start time of the procedure is moved forward or backward in combination with the completion time of the precedent procedure in the product assembly dependency directed graph and the idle time period of the current resource, and the expected end time is updated accordingly; if the adjusted time meets the precedent dependency relationship and the resource idle constraint, a set of time correction paths is generated, and the time difference before and after adjustment is recorded.

[0102] According to the conflict procedure set, a procedure pair with the same resource number, partially overlapping time and no precedent dependency relationship is selected, and for each procedure pair that meets the condition, the scheduling order in the task assignment sequence is tried to be exchanged, and the expected start time and end time are recalculated.

[0103] determining whether the exchanged procedures can meet the precedent dependency relationship and the current resource idle time constraint, and if so, generating a set of order exchange paths and recording the start time change after adjustment.

[0104] Specifically, the step of generating a set of order exchange paths is to exchange the execution order of the procedure pair in the conflict procedure set located at the same resource number and having overlapping time period but no precedent dependency relationship in the product assembly dependency directed graph, and recalculate the expected start time and end time of each procedure; if the exchange still meets the dependency relationship constraint and the resource idle condition, a set of order exchange paths is generated, and the start time change after adjustment is recorded.

[0105] Specifically, three types of correction paths are generated in the correction process according to different situations: when the resource needs to be replaced, the replacement correction path is generated; when the resource is not replaced and only the start and end time needs to be adjusted, the time correction path is generated; when the execution order is adjusted in the same resource, the order exchange path is generated, and the three types of correction paths are independent of each other and complementary to each other.

[0106] By sharing common data sources such as conflict procedure set, standard assembly time, precedent dependency relationship, etc., the problem of repeated retrieval and inconsistent information is avoided. On the other hand, the generation steps of the three types of correction paths are independent of each other and complementary in logic, which can provide diversified solutions from the dimensions of resource replacement, time adjustment and scheduling order, significantly improving the correction flexibility and solution quality under complex conflict situations.

[0107] Step 7: The final revised task allocation scheme is sent to the corresponding assembly unit. During the execution of the assembly task, the real-time state changes of the multi-dimensional constraint parameters are monitored in real time. When it is detected that any constraint condition is broken or the lower bound of executability becomes negative, the task rollback mechanism is automatically triggered. The current task allocation scheme is replaced by a new one selected in descending order of the lower bound of executability, and the task execution process is updated.

[0108] Due to sudden situations such as personnel changes, equipment failures, or rhythm fluctuations in the assembly site, some constraint conditions (such as resource availability, time window restrictions, or dependency relationships) may be broken during execution, resulting in the failure of the task allocation scheme. By monitoring the changes of multi-dimensional constraint parameters in real time and automatically triggering rollback based on the lower bound of executability, the invalid path can be identified and replaced in time. The most feasible alternative scheme is selected in descending order of the lower bound of executability, ensuring that the task flow does not interrupt and the assembly plan can be continuously executed. This effectively reduces the chain effect caused by abnormal propagation and improves the resilience and real-time response capability of the entire assembly system.

[0109] In this embodiment, it needs to be specifically pointed out that, as shown in Figure 2 A mechanical product assembly task allocation system, the system comprises:

[0110] An assembly task data receiving module for receiving assembly task data, the assembly task data including target product model, target product process list, target product structure information, and target product plan completion time;

[0111] A process task list generation module for converting the target product structure information in the assembly task data into a plurality of assembly processes, extracting the assembly time, resource type, and sequence relationship required for each assembly process, and generating a process task list;

[0112] An assembly resource information acquisition module for acquiring assembly resource state information in the current production site, the assembly resources including assembly stations, assembly equipment, and personnel, and the state information including idle conditions, executable process ranges, and expected idle periods;

[0113] A preliminary allocation scheme generation module for matching the process tasks and assembly resources by calling a preset task scheduling algorithm based on the process task list and assembly resource state information, and generating a preliminary task allocation scheme for the assembly task;

[0114] A final scheme acquisition module for calculating and evaluating a set of qualified task allocation schemes from all preliminary task allocation schemes, calculating the corresponding multi-dimensional normalized margin value for each qualified task allocation scheme in the set of qualified task allocation schemes, and selecting a final task allocation scheme based on the multi-dimensional normalized margin value;

[0115] An ultimate solution modification module is configured to perform conflict detection on the ultimate task allocation solution, identify conflicts such as repeated conflicts of stations, resource conflicts or task timing conflicts, and modify the ultimate task allocation solution if there is a conflict to obtain a final modified task allocation solution; if there is no conflict, the ultimate task allocation solution is directly regarded as the final modified task allocation solution.

[0116] A cycle adjustment module is configured to send the final modified task allocation solution to the corresponding assembly unit, monitor real-time state changes of the multi-dimensional constraint parameters in real time, and automatically trigger a task rollback mechanism when any constraint condition is detected to be destroyed or the lower bound of the executability becomes negative, and select a replacement current task allocation solution in descending order of the lower bound of the executability to reissue and update the task execution process.

[0117] Finally, the above is only the preferred embodiment of the present application and is not intended to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0118] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited to this, and any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which shall be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A mechanical product assembly task allocation method characterized by, The method comprises the following steps: Step 1: receiving assembly task data, the assembly task data comprising a target product model, a target product process list, target product structure information and a target product planned completion time; Step 2: converting the target product structure information in the assembly task data into a plurality of assembly processes, extracting the assembly time, resource type and sequence relationship required by each assembly process, and generating a process task list; Step 3: collecting assembly resource state information in the current production site, the assembly resources comprising assembly stations, assembly equipment and personnel, and the state information comprising idle conditions, executable process ranges and predicted idle time periods; Step 4: according to the process task list and the assembly resource state information, calling a preset task scheduling algorithm to match the process tasks and the assembly resources, and generating a preliminary task allocation scheme for the assembly task; Step 5: calculating and evaluating a set of qualified task allocation schemes from all preliminary task allocation schemes, calculating a multi-dimensional normalized margin value for each qualified task allocation scheme in the set of qualified task allocation schemes, and selecting a final task allocation scheme according to the multi-dimensional normalized margin value; Step 6: performing conflict detection on the final task allocation scheme to identify station duplication conflicts, resource conflicts or task timing conflicts, and if there are conflicts, modifying the final task allocation scheme to obtain a final modified task allocation scheme; If there are no conflict items, the final task allocation scheme is directly regarded as the final modified task allocation scheme; Step 7: sending the final modified task allocation scheme to the corresponding assembly unit, and in the assembly task execution process, real-time monitoring of the real-time state changes of the multi-dimensional constraint parameters, and when any constraint condition is detected to be destroyed or the executability lower bound value becomes negative, automatically triggering a task rollback mechanism, and selecting a replacement current task allocation scheme in descending order of the executability lower bound value to reissue and update the task execution process; The multi-dimensional normalized margin value is obtained by: obtaining the range of the opening of the jig used in each process in the current assembly task, the range of the opening of the jig comprising the maximum opening size and the minimum opening size of the jig, and obtaining the maximum part size, calculating the size margin value of the process by the ratio of the maximum opening size minus the maximum part size to the maximum opening size minus the minimum opening size; obtaining the target cycle time and the historical Q90 cycle value of the process, calculating the cycle margin value by the ratio of the historical Q90 cycle value minus the target cycle time to the target cycle time; extracting the last obstacle avoidance distance of the robot in the execution path, obtaining the minimum safety clearance threshold of the path, calculating the path margin value by the ratio of the last obstacle avoidance distance minus the minimum safety clearance threshold of the path to the minimum safety clearance threshold of the path; obtaining the number of personnel qualifications required by the current process and the number of matchable personnel qualifications on site, calculating the personnel margin value by the ratio of the number of matchable personnel qualifications on site minus the number of personnel qualifications required by the current process to the number of personnel qualifications required by the current process; The maximum reachable boundary distance under the current robot kinematics model and the actual required motion distance are obtained at the end of the acquisition process, and the reachable margin value of the robot is obtained by calculating the ratio of the maximum reachable boundary distance minus the actual required motion distance to the maximum reachable boundary distance; The size margin value, the beat margin value, the path margin value, the personnel margin value, and the robot reachable margin value are subjected to vector normalization processing to obtain the normalized size margin value, the beat margin value, the path margin value, the personnel margin value, and the robot reachable margin value; The step of selecting the final task allocation scheme according to the multi-dimensional normalized margin value is: The minimum value in the multi-dimensional normalized margin value of each qualified task allocation scheme is defined as the lower bound of the executability of the scheme; The qualified task allocation scheme with the maximum lower bound of executability is selected as the final task allocation scheme.

2. The method of claim 1, wherein: The step of obtaining the process task list is: In the received assembly task data, the target product structure information field is read, the structure information includes the part list, the assembly node, the part connection mode, and the hierarchical structure, the component composition unit is extracted according to the product structure tree, and the product number of the target product is bound to all component composition units; Based on the assembly sequence and connection logic between components, a product assembly dependent directed graph is constructed, and the upstream components depended on by each component node are recorded; According to the product assembly dependent directed graph, each assembly action is decomposed into an executable assembly process according to the component connection logic and operation constraints, and each assembly process is bound to a unique component pair and assembly action type, and the predecessor process number set is automatically identified as the dependency relationship between processes; For each assembly process, the standard assembly time required by the current process is extracted from the process database or the built-in standard process table according to the component complexity, assembly method, and past historical standard operation data, and the resource type required by the process is identified, including the required workstation type, equipment model, tool number, and personnel skill level requirement; For each process, record its process number, product number, bound component pair, assembly action type, predecessor dependency relationship, standard assembly time, and required resource type, and generate a structured process task list.

3. The mechanical product assembly task allocation method according to claim 1, wherein, The step of obtaining the preliminary task allocation scheme is: Read the standard assembly time, required resource type, and predecessor dependency relationship of each process in the process task list, and obtain the assembly resource state information of each assembly resource; For each process task, according to the required resource type, filter out the resource units that meet the resource type requirement and cover the process standard assembly time from the current assembly resource state information, record the resource number and available start time of each candidate resource, and record the resource number and available start time of each candidate resource. Call the preset task scheduling algorithm, execute topological sorting and resource earliest available priority principle according to the predecessor dependency relationship between processes and the idle time window of candidate resources, and assign process tasks to candidate resources in turn, and record the predicted start time and predicted end time of the process, form a complete task allocation sequence, and take it as the preliminary task allocation scheme.

4. The mechanical product assembly task allocation method according to claim 1, wherein, The obtaining step of the qualified task allocation scheme set is: For each preliminary task allocation scheme, obtain multi-dimensional constraint parameters related to assembly feasibility, including jig opening range, maximum part size, minimum safety gap threshold of path, target cycle time and historical Q90 cycle value, end pose of process and maximum reachable boundary distance of robot, number of matchable personnel qualifications and number of personnel qualifications required by the process; Perform item-by-item compliance check on the multi-dimensional constraint parameters, eliminate preliminary task allocation schemes that do not meet the basic constraint conditions, obtain qualified task allocation schemes, and integrate all qualified task allocation schemes to obtain a qualified task allocation scheme set.

5. The mechanical product assembly task allocation method according to claim 1, characterized in that: The obtaining step of the final modified task allocation scheme is: Step 6.1: For each conflict item, extract the main resource unit causing the conflict as the current conflict resource. If it is a station overlap conflict, the current conflict resource is the overlapping station number. If it is a resource occupation conflict, the current conflict resource is the conflict resource number. If it is a task timing conflict, the current conflict resource is the resource number corresponding to the previous process on the dependent path; Step 6.2: For each conflict item, obtain the number of conflict processes, the number of downstream dependent processes, the time conflict span, the number of alternative resources of the current conflict resource, and the number of the same type of resources in the total resource pool; Calculate the conflict resource scarcity degree according to the number of alternative resources of the current conflict resource and the number of the same type of resources in the total resource pool; Normalize the number of conflict processes, the number of downstream dependent processes, the time conflict span, and the conflict resource scarcity degree, and calculate the conflict priority index according to the normalized number of conflict processes, the number of downstream dependent processes, the time conflict span, and the conflict resource scarcity degree; Step 6.3: For each conflict item, generate three feasible modification paths in its current conflict process: replacement modification path, time modification path, and order interchanging path; Step 6.4: For each feasible modification path, extract the set of processes affected before and after modification, calculate the start time difference for each process, and accumulate to obtain the disturbance value; Step 6.5: In all modifiable paths, select the path with the smallest disturbance value as the final modification method for the current conflict item, and apply it to the task allocation scheme to update the conflict item related fields; Step 6.6: For the remaining conflict items, execute steps 6.3 to 6.5 in order from high to low according to the conflict priority index until all conflicts are handled, forming the final modified task allocation scheme.

6. The method of claim 5, wherein: The step of generating three feasible modification paths: replacement modification path, time modification path, and order interchanging path is: Obtain the conflict process set in the current conflict item, read the current conflict resource number and process standard assembly time bound to each process; According to the resource type of the conflict resource, filter out alternative resources that are consistent with the current conflict resource type and in idle state within the process expected start and end time range from the assembly resource state information to form an alternative resource set; For each conflict process, it is sequentially assigned to each resource in the alternative resource set, and if it can be assigned successfully, a set of replacement correction paths is generated, and the resource number and corresponding time period modified by the correction path are recorded; According to the conflict process set, the precedence relationship and the standard assembly time of each process are extracted, and under the premise of not changing the resource number, the expected start time of the process is adjusted forward or backward combined with the process precedence relationship and the current resource state information, so as to avoid the occupation time period of the current conflict resource, and the expected end time is updated; Determine whether the adjusted time meets the completion time requirement of the previous process and the resource idle window condition, if it meets, a set of time correction paths is generated, and the time period difference value after adjustment is recorded; According to the conflict process set, the process pairs with the same resource number and time overlap but no precedence relationship are screened out, for each process pair that meets the condition, the scheduling order in the task allocation sequence is tried to be exchanged, and the expected start time and end time are recalculated; Determine whether the exchanged processes can meet the precedence relationship and the current resource idle time constraint, if it meets, a set of order exchange paths is generated, and the start time change after adjustment is recorded.

7. A mechanical product assembly task allocation system for implementing a mechanical product assembly task allocation method according to any one of claims 1 to 6, characterized by: The system comprises: An assembly task data receiving module for receiving assembly task data, the assembly task data including target product model, target product process list, target product structure information, and target product planned completion time; A process task list generation module for converting the target product structure information in the assembly task data into a plurality of assembly processes, extracting the assembly time, resource type, and sequence relationship required by each assembly process, and generating a process task list; An assembly resource information acquisition module for acquiring the state information of the assembly resources in the current production site, the assembly resources including assembly stations, assembly equipment, and personnel, and the state information including idle condition, executable process range, and expected idle period; A preliminary allocation scheme generation module for matching the process tasks and assembly resources according to the process task list and assembly resource state information, and generating a preliminary task allocation scheme for the assembly task by calling a preset task scheduling algorithm; A final scheme acquisition module for calculating and evaluating a set of qualified task allocation schemes from all preliminary task allocation schemes, calculating the corresponding multi-dimensional normalized margin value for each qualified task allocation scheme in the set of qualified task allocation schemes, and selecting a final task allocation scheme according to the multi-dimensional normalized margin value; A final scheme correction module for performing conflict detection on the final task allocation scheme, identifying process station duplication conflicts, resource conflicts, or task timing conflicts, and if there are conflicts, modifying the final task allocation scheme to obtain a final modified task allocation scheme; if there are no conflict items, the final task allocation scheme is directly regarded as the final modified task allocation scheme. A circulation adjustment module is configured to send the final modified task allocation scheme to the corresponding assembly units, monitor the real-time state changes of the multi-dimensional constraint parameters in real time, and automatically trigger a task rollback mechanism when detecting that any constraint condition is destroyed or the lower bound of the executability becomes negative, and select an alternative current task allocation scheme in descending order of the lower bound of the executability to reissue and update the task execution process.

Images