Intelligent scheduling optimization method and system for multi-channel magnetic levitation production lines

Abstract

This invention discloses an intelligent scheduling optimization method and system for multi-channel magnetic levitation production lines, belonging to the field of scheduling optimization technology. The method includes: constructing a multi-objective model; outputting the initial task execution plan for the moving trolleys of the multi-channel magnetic levitation production line; constructing an event-driven synchronization log linked list for each scheduling group; analyzing the consistency risk score of each scheduling group; determining the scheduling advancement strategy for each scheduling group; when the scheduling advancement strategy is to initiate local prediction, statistically analyzing the lag time series and step deviation series of each predicted scheduling group; determining the scheduling optimization strategy for each predicted scheduling group; and thereby executing the intelligent scheduling optimization for each predicted scheduling group. This invention achieves calculable, predictable, and convergent control of scheduling synchronization deviations, improving the real-time consistency, scheduling robustness, and system safety of multi-channel magnetic levitation production lines under complex parallel processes.

Description

Technical Field

[0001] This invention relates to the field of scheduling optimization technology, and in particular to an intelligent scheduling optimization method and system for multi-channel magnetic levitation production lines. Background Technology

[0002] In traditional discrete manufacturing, production lines are mostly driven by discrete cycles such as belts, chains, and roller conveyors, with workpieces occupying equipment and workstations sequentially at fixed positions. Typical scheduling is centered on a job-machine model, using rules (such as EDD, SPT, FIFO) or intelligent algorithms (such as heuristics, metaheuristics, and reinforcement learning) to solve the trade-offs between resource constraints (equipment capacity, changeover time, work-in-process inventory limits) and objectives (delivery time, equipment utilization, energy consumption, and cycle stability). The common approach is to map the process sequence to equipment occupancy and queuing relationships, forming intra-station / inter-station scheduling; to generate solutions layer by layer across multiple time scales—seconds, minutes, and hours—combining station release (second-level response), work-in-process balancing / energy efficiency coordination (minute-level), and production line coordination (hour-level); and to rely on centralized or hierarchical control, using near real-time state-driven strategy updates and loop corrections.

[0003] For example, Chinese invention patent CN114611897B discloses an intelligent production line adaptive dynamic scheduling strategy selection method, which includes three steps: (1) a selection scheduling framework model is proposed, including a workshop production environment module, an algorithm module, and a scheduling rule module. (2) Based on the framework proposed in step (1), strategy selection is implemented using an algorithm based on deep reinforcement learning. (3) Based on the strategy selection in step (2), an adaptive model for complex dynamic environment changes is established, and an adaptive rule scheduling selector is designed, including a scheduling rule selection scheduler and a job machine pair selection scheduler.

[0004] For example, Chinese invention patent CN112465333B discloses a method for intelligent production line scheduling optimization based on multiple time scales, which includes: modeling based on the process flexibility, processing flexibility, and processing sequence flexibility existing in flexible process planning; establishing a second-level equipment capacity configuration optimization strategy with optimal equipment utilization as the optimization objective; establishing a minute-level process energy efficiency optimization strategy with minimum energy consumption as the optimization objective; establishing an hour-level production line joint optimization strategy with optimal system scheduling efficiency as the optimization objective; integrating the quantitative relationship between variables and objectives under the three time scales, and starting from the actual physical system, performing the final optimization solution by assigning certain optimization weights to the three levels.

[0005] With increasing demands for cleanliness, flexibility, and high responsiveness, production lines have evolved from traditional conveyor systems to magnetic levitation production lines centered on linear motors. However, conventional production lines operate at the workstation level, emphasizing discrete events such as arrival, occupancy, and departure. Path stability is limited, branching is minimal, queuing relationships are clear, and buffering is primarily confined to fixed workstations / queues. In contrast, magnetic levitation achieves low friction and high responsiveness through stator-motor linear drive. Position, speed, acceleration, and dwell time become continuously adjustable variables. Traffic flow following, braking, and platooning constraints enter the scheduling domain, resulting in parallel channels and frequent branching. Motors can selectively bypass, cut in line, or stop, requiring dynamic assessment of channel capacity and conflict zone occupancy. Therefore, magnetic levitation production lines adopt a distributed architecture combining local control (workstations / branching) and central scheduling. Status and instructions are asynchronously published and subscribed to on the network through event-driven mechanisms. In the context of multiple vehicles operating in parallel, frequent lane switching, and high cycle time, the system inevitably experiences brief windows of weak consistency. For example, a node may have updated its process status (step completion, parameter switching, workstation release, or conflict zone clearing), but other nodes involved in the same decision may remain on the old version. If scheduling and release / rerouting decisions still default to near real-time global consistency (continuing the implicit assumption of ordinary production lines), conflicting decisions will arise within the weak consistency window: one side releases traffic, while the other restricts or preempts it. This can lead to minor issues like mis-driving, delays, or unnecessary yielding causing cycle time jitter, or, in severe cases, amplified in complex topologies into system-level process disorder, cycle time drift, and decreased energy efficiency. Summary of the Invention

[0006] To address the aforementioned technical problems in the existing technology, embodiments of the present invention provide an intelligent scheduling optimization method and system for multi-channel magnetic levitation production lines. The technical solution is as follows:

[0007] On the one hand, an intelligent scheduling optimization method for multi-channel magnetic levitation production lines is provided, including:

[0008] A multi-objective model is constructed, multi-dimensional feature vectors are collected, and the initial task execution scheme of the moving trolley on the multi-channel maglev production line is output, thereby dividing the moving trolley on the multi-channel maglev production line into multiple scheduling groups.

[0009] Construct event-driven synchronization log linked lists for each scheduling group, analyze the consistency risk score of each scheduling group, and determine the scheduling advancement strategy for each scheduling group.

[0010] When the scheduling advancement strategy is to initiate local prediction, the corresponding scheduling group is recorded as the prediction scheduling group. The prediction scheduling group is statistically analyzed, and the lag time series and step deviation series of each prediction scheduling group are analyzed to determine the scheduling optimization strategy of each prediction scheduling group. Then, the intelligent scheduling optimization of each prediction scheduling group is executed.

[0011] On the other hand, an intelligent scheduling and optimization system for multi-channel magnetic levitation production lines is provided. The system includes an initial task execution plan generation module, a scheduling advancement strategy determination module, and a scheduling optimization strategy determination and execution module.

[0012] The initial task execution plan generation module is used to construct a multi-objective model, collect multi-dimensional feature vectors, and output the initial task execution plan of the moving trolley on the multi-channel maglev production line, thereby dividing the moving trolley on the multi-channel maglev production line into multiple scheduling groups.

[0013] The scheduling advancement strategy determination module is used to construct the event-driven synchronization log linked list for each scheduling group, analyze the consistency risk score of each scheduling group, and determine the scheduling advancement strategy for each scheduling group.

[0014] The scheduling optimization strategy determination and execution module is used to record the corresponding scheduling group as the predicted scheduling group when the scheduling advancement strategy is to start local prediction, to count each predicted scheduling group, to analyze the lag time series and step deviation series of each predicted scheduling group, to determine the scheduling optimization strategy of each predicted scheduling group, and to execute the intelligent scheduling optimization of each predicted scheduling group.

[0015] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0016] 1. The intelligent scheduling optimization method for multi-channel magnetic levitation production lines provided by this invention can achieve dynamic identification and adaptive optimization of state consistency risks in a distributed scheduling system through the coordinated execution of each step. Specifically, by constructing an event-driven synchronous log linked list for each scheduling group, the data reported by the moving trolley and the process switching event nodes are arranged in chronological order to form a traceable time-series linked list. This maintains the integrity and timing accuracy of event records within the group in an asynchronous communication environment, providing a quantifiable basis for subsequent consistency risk score calculation. By analyzing the consistency risk scores of each scheduling group, the system can quantify the degree of synchronization deviation of each scheduling group in different process advancement stages in real time, thereby accurately identifying potential out-of-synchronization risks. Furthermore, through a dynamic correction mechanism for the consistency risk threshold, a frequency-risk correction tightening table and a loosening table driven by historical anomaly frequency are introduced, making the consistency risk threshold more flexible and sustainable. The value can be adaptively adjusted according to the stability of the scheduling group, which can avoid false alarms and enhance the system's sensitivity to abnormal groups. When the consistency risk score exceeds the consistency risk threshold, the system automatically performs local prediction, calculates the probability of time lag and the probability of step deviation, and quantifies the synchronization risk level of multiple vehicles in the group in terms of process progress and event response. When the predicted scheduling group triggers the step threshold tightening strategy, the allowable process step difference within the group is dynamically adjusted, so that each moving vehicle in the group executes the next process after the data is not fully synchronized, thereby avoiding false triggers caused by asynchronous reporting or network latency. Normal progress is resumed after the step criterion is met, so that the scheduling rhythm is always within a controllable risk range. Through the coordinated execution of the above steps, this invention achieves calculable, predictable, and convergent control of scheduling synchronization deviation, improving the real-time consistency, scheduling robustness, and system safety of multi-channel magnetic levitation production lines under complex parallel processes.

[0017] 2. This invention determines the scheduling advancement strategy for each scheduling group. After determining the scheduling advancement strategy based on the comparison results of consistency risk score and consistency risk threshold, it executes two strategies for different groups: continuing to advance according to the set scheduling plan or initiating local prediction. This prevents the scheduling system from advancing at a globally uniform pace, but rather adaptively adjusts the advancement pace according to the real-time synchronization status of each scheduling group. Maintaining the established scheduling plan ensures the stability and efficiency of the production rhythm; triggering the local prediction process enables the local isolation and synchronization status repair of abnormal groups. Furthermore, it allows for the adjustment of the scheduling pace of high-risk groups without affecting the advancement of other normal scheduling groups. While ensuring the overall throughput of the system, it reduces the risk of out-of-order flow and process interference caused by cross-node state asynchrony, achieving a dynamic balance between scheduling stability, sensitivity, and robustness.

[0018] 3. This invention determines the scheduling optimization strategy for each predictive scheduling group. Based on the time lag probability and step deviation probability calculated from the lag time series and step deviation series, it automatically judges the persistence and severity of anomalies and generates the optimal scheduling optimization strategy accordingly. For predictive scheduling groups with a high proportion of synchronization anomalies, the system implements a step threshold tightening strategy, which dynamically reduces the maximum allowable step difference of the moving trolleys within the group in subsequent process advancement, thereby achieving local timing correction without completely relying on global synchronization signals. For the corresponding scheduling groups that continue to advance according to the set scheduling scheme, it can avoid production line rhythm fluctuations caused by ineffective intervention. In summary, a closed-loop optimization path can be formed at the predictive group level, enabling scheduling to have the ability to respond in advance to asynchronous risks and gradually converge, ensuring that multi-vehicle process flows can still maintain timing consistency, process continuity, and high reliability of instruction execution under dynamic operating conditions. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the intelligent scheduling and optimization method for a multi-channel magnetic levitation production line provided in an embodiment of the present invention.

[0021] Figure 2 This is a structural diagram of an intelligent scheduling and optimization system for a multi-channel magnetic levitation production line provided in an embodiment of the present invention.

[0022] Figure 3 This is a flowchart for determining consistency risk scoring and scheduling advancement strategies.

[0023] Figure 4 This is a flowchart for determining the scheduling optimization strategy for predictive scheduling groups.

[0024] Figure 5 This is a reference diagram of a single-channel magnetic levitation production line. Detailed Implementation

[0025] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0026] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0027] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0028] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0029] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0030] like Figure 5 The diagram shows a reference image of a single-channel magnetic levitation production line. This line employs a closed-loop track layout, primarily consisting of straight and curved sections. Multiple fixed supports and connecting components are distributed above the track to ensure its stability and precision. The magnetically levitated actuator (not shown in the diagram) can perform high-speed, low-friction, and contactless precise material handling along the guide rail, enabling efficient flow of materials or workpieces on the production line. This type of single-channel design is commonly used in automated production scenarios such as electronic assembly, semiconductors, and precision components, where high cleanliness, response speed, and flexible layout are required. It provides a foundation for subsequent expansion into multi-channel or complex network structures.

[0031] like Figure 1 The flowchart shown here illustrates an intelligent scheduling optimization method for a multi-channel magnetic levitation production line, including:

[0032] A multi-channel magnetic levitation production line refers to a flexible automated production system that uses magnetic levitation drive technology to achieve simultaneous coordinated operation of multiple parallel or intersecting work channels (tracks). This system is widely used in precision assembly, electronics manufacturing, and material handling in modern manufacturing. Unlike traditional mechanical chain conveyors or rail-based transport, multi-channel magnetic levitation production lines arrange several independently dispatchable transport units on multiple independent or interconnected magnetic levitation tracks, enabling efficient, synchronous, and flexible flow of products, components, tooling, and other materials along complex process paths. This significantly improves production line throughput, changeover flexibility, and automation levels. In this system, the following term "moving trolley" refers to the core carrier unit running on the magnetic levitation track, often called a magnetic levitation tray or moving unit. The moving trolley integrates magnetic levitation coils or permanent magnet components, moving on the track without contact and with high precision using a magnetic levitation drive system. Each moving trolley is typically equipped with an independent control unit, sensors, communication modules, and a power system, capable of automatically completing actions such as starting, stopping, acceleration, deceleration, positioning, track changing, and obstacle avoidance according to dispatching instructions. The trolley not only undertakes the physical transportation of products or components, but also enables online assembly, barcode scanning, and testing through the integration of fixtures, tools, or RFID reading and writing modules. It is a key equipment foundation for realizing highly flexible and intelligent production on magnetic levitation production lines.

[0033] A multi-objective model is constructed, multi-dimensional feature vectors are collected, and the initial task execution scheme of the moving trolley on the multi-channel maglev production line is output, thereby dividing the moving trolley on the multi-channel maglev production line into multiple scheduling groups.

[0034] In this embodiment, the multi-objective model is a mathematical modeling tool used to optimize task allocation and path scheduling for trolleys on a magnetic levitation production line. Based on actual production needs, the model comprehensively considers multiple objective functions, including maximizing production efficiency, minimizing energy consumption, balancing equipment utilization, minimizing path conflicts, minimizing task response latency, synchronizing process batches, and optimizing equipment maintenance windows. It is constructed using weighted multi-objective optimization algorithms (such as multi-objective integer programming, weighted linear combination, or Pareto front analysis). To support model optimization, the system needs to collect multi-dimensional feature vectors for each trolley and related processes in real-time or periodically. The collected data includes, but is not limited to: the trolley's unique identifier ID, current physical location coordinates (e.g., track segment number, distance, workstation number), currently assigned process task type, task priority, and remaining process steps, equipment health status (temperature, battery level, maintenance cycle), trolley speed, acceleration, direction of travel, path history (completed task trajectories), process batch number, and associated material information. This data is collected in real-time through sensors, control units, and MES / SCADA system interfaces, and aggregated in a scheduling server or central database as input features for the model. Based on the aforementioned multi-objective model and the collected multi-dimensional features, the system invokes a solution engine (such as integer linear programming, genetic algorithm, particle swarm optimization, reinforcement learning scheduling, etc.) to calculate the optimal or suboptimal initial task allocation and path scheme for each moving vehicle under the current production situation. The output includes: the start and end points of each moving vehicle's task; task type and priority allocation; optimal / feasible path (including the workstations along the route and the estimated arrival time); and the estimated start / end times of task execution. The system supports batch computation and rolling optimization (refreshing the scheme in real time according to the actual situation).

[0035] Based on the initial task execution plan, the system schedules and groups the moving vehicles according to business rules. Common grouping principles include, but are not limited to: process type grouping (grouping moving vehicles performing the same type of process (such as assembly, handling, and inspection) into the same group), path / area grouping (grouping moving vehicles traveling on the same track segment, in the same area, or with highly overlapping paths into the same group, facilitating path conflict coordination and refined scheduling), batch synchronization grouping (grouping multiple moving vehicles in the same process batch that need to be coordinated into one group, improving process consistency), and equipment characteristic grouping (grouping according to differences in moving vehicle performance (such as load capacity and speed). The grouping scheme is automatically generated by the scheduling algorithm, and manual intervention and rule configuration are also supported. The system assigns a unique group number to each group and records the list of moving vehicles belonging to it, serving as the basis for subsequent core processes such as group scheduling, anomaly detection, and consistency analysis.

[0036] See Figure 3The diagram shows the flowchart for determining the consistency risk score and scheduling advancement strategy. First, an event-driven synchronization log linked list is constructed for each scheduling group to record the event and state change information of each moving trolley in the multi-channel maglev production line. Then, the consistency risk score of each scheduling group is analyzed, and the scheduling consistency risk of each group is assessed by calculating the weighted result of the maximum event lag and step deviation. Next, the consistency risk threshold of each scheduling group is analyzed, and the threshold level of each group is determined by combining the basic consistency risk threshold and the correction factor of historical anomaly frequency. Afterward, the system performs a judgment: whether the consistency risk score is less than or equal to the consistency risk threshold. If the judgment result is yes, the scheduling advancement strategy of the corresponding scheduling group is recorded as "continue to advance with the set scheduling scheme"; if the judgment result is no, the scheduling advancement strategy of the corresponding scheduling group is recorded as "start local prediction" and anomalies are synchronously marked, thereby realizing dynamic hierarchical control of the scheduling groups of the multi-channel maglev production line.

[0037] Construct event-driven synchronization log linked lists for each scheduling group, analyze the consistency risk score of each scheduling group, and determine the scheduling advancement strategy for each scheduling group.

[0038] Furthermore, an event-driven synchronization log linked list for each scheduling group is constructed. The specific construction process is as follows:

[0039] Establish an independent log list structure for each scheduling group.

[0040] In this embodiment of the invention, establishing an independent log linked list structure for each scheduling group specifically involves: allocating a unique storage space for each scheduling group in the distributed database of the production line main control system, initializing it as an empty linked list structure. Each linked list node is used to store event nodes within a single collection cycle. The linked list structure uses pointers or indexes to achieve ordered connections between nodes, ensuring that data can be appended sequentially according to the collection order and the order in which events occur, and allowing for rapid backtracking.

[0041] In each scheduling group, data reported by the assigned moving trolley is received according to the preset collection cycle. The data includes at least the unique identifier of the moving trolley, the current position, the process step number, the local clock, the current task number, and the data version number.

[0042] In a specific embodiment, the data acquisition period refers to the time interval at which the system periodically triggers the moving trolleys to report their operating status data. In actual deployment, the acquisition period can be set based on factors such as the complexity of the production line process, the real-time requirements of the equipment, and communication bandwidth. Specifically, during parameter initialization, the system sets an acquisition period parameter for each scheduling group, such as 1 second, 500 milliseconds, or other specific values ​​(which can be set empirically by experts). This period parameter can be preset and adjusted through the operation interface, configuration file, or remote parameter distribution. Driven by the synchronous clock signal of the system's main control server, the moving trolley controllers within the scheduling group periodically acquire their own status according to the set acquisition period and send data packets to the main control server according to protocol requirements. After data acquisition, the acquired data is sent to the production line main control server via wired or wireless network, according to the specified data packet format (such as TCP / IP, EtherCAT, CAN, or a custom protocol). The main control server is equipped with a communication interface module that listens to and parses data packets from each moving trolley in real time, identifies the corresponding scheduling group, and archives the data into the log linked list structure of the corresponding group, ensuring data integrity and timing accuracy.

[0043] When any event in the preset process switching event set occurs on the trolley, an event node is generated that includes the event type, event number, trolley unique identifier, and local clock.

[0044] The process changeover event set is a predefined set of key event types by engineers based on the actual needs of the production process during system initialization or production line configuration. It can be stored in the main control server's configuration file, database table, or firmware parameters of the embedded controller. Specific event types typically include trolley arrival at workstation, process step switching, task assignment, task completion, emergency stop, stall, and communication failure. The event set can be added, deleted, or adjusted via an interface configuration tool. During runtime, the system determines whether the status change reported by the trolley belongs to a process changeover event based on the predefined event set. If it does, an event node is generated and added to the log chain for management.

[0045] Data and event nodes are written to the corresponding log linked list in chronological order. Each node includes all the above fields, forming a log node sequence arranged in ascending order of time.

[0046] Traverse each scheduling group to obtain the event-driven synchronization log linked list for each scheduling group.

[0047] Furthermore, the consistency risk scores of each scheduling group are analyzed. The specific analysis process is as follows:

[0048] Obtain the historical anomaly frequency for each scheduling group.

[0049] In this embodiment of the invention, the historical anomaly frequency refers to the ratio of the number of times each scheduling group experiences anomalies to the total number of analyses within a specified historical analysis period. Specifically, the anomaly determination criterion is: if the consistency risk score of each scheduling group is greater than the consistency risk threshold of that scheduling group within a risk analysis cycle, then an anomaly is marked. After statistical analysis over multiple cycles, the historical anomaly frequency can be defined as: Historical Anomaly Frequency = Number of Anomalies / Total Number of Analyses. For example, if a scheduling group is determined to be anomaly 15 times in the most recent 100 analysis cycles, then its historical anomaly frequency is 0.15. This frequency is used to reflect the activity level of the group's consistency risk within the historical window.

[0050] Using the historical anomaly frequency of each scheduling group as the query index, the number of tracing steps for each scheduling group can be obtained.

[0051] To dynamically adjust the tracing steps of the log linked list based on the historical anomaly frequency of a scheduling group, the system pre-defines a mapping table. This table uses anomaly frequency ranges as indexes, with each range corresponding to a specific number of tracing steps. Specifically, during initialization or operation, the system sets several anomaly frequency thresholds, such as ranges [0, 0.05), [0.05, 0.15), [0.15, 0.30), and [0.30, 1.00], corresponding to tracing steps N1, N2, N3, and N4 respectively. In practical applications, the system finds the range of a scheduling group based on its currently calculated historical anomaly frequency and then retrieves the corresponding tracing steps. For example, if the anomaly frequency is 0.12, falling within the range [0.05, 0.15), then the tracing steps are N2 (e.g., N2 = 10 steps). The higher the historical anomaly frequency, the longer the corresponding tracing steps are extracted, allowing for automatic extension of the backtracking window for high-risk scheduling groups and improving the sensitivity of consistency risk detection.

[0052] Based on the number of tracing steps for each scheduling group, the event reporting records of each scheduling group are traced backward from the event-driven synchronization log chain of each scheduling group to obtain the event reporting timestamp sequence and event reporting version number sequence of each scheduling group.

[0053] The maximum event lag for each scheduling group is obtained based on the event reporting timestamp sequence of each scheduling group.

[0054] Maximum event lag refers to the maximum time difference between the local clock timestamps reported by different moving vehicles when the same type of process event (such as process step switching) occurs within the same tracing step. Specifically, the timestamps of all moving vehicles at the time of a certain process event are taken, and the difference between the maximum and minimum values ​​is calculated as the lag for that event. This calculation is performed separately for all critical events within the tracing window, and the maximum value is taken to obtain the maximum event lag.

[0055] The step deviation of each scheduling group is obtained based on the sequence of event reporting version numbers of each scheduling group.

[0056] Step deviation refers to the maximum difference between the event version numbers reported by all moving vehicles in the same scheduling group when the same process event occurs within a specified traceability window. That is, for the same event, the version numbers reported by all moving vehicles are collected, and the difference between the maximum and minimum values ​​is calculated. This value reflects the degree of deviation in the synchronization of the process progress of each moving vehicle within the scheduling group. The larger the step deviation, the greater the difference in the process progress of the moving vehicles within the group, indicating a potential synchronization risk.

[0057] In one specific embodiment, a consistency risk score analysis is performed on a particular scheduling group. Within the last 100 risk analysis periods, this scheduling group had 4 instances where its consistency risk score exceeded the consistency risk threshold; therefore, the historical anomaly frequency is 0.04. According to the system's preset anomaly frequency-tracing step mapping table, a frequency of 0.04 corresponds to 3 tracing steps. Taking this scheduling group as the object, we statistically analyzed its three most recent event reporting records. Assuming that the scheduling group contains 5 moving trolleys, the timestamps and version numbers for these three critical process events are as follows: Event 1, reporting timestamps are 102.1 seconds, 103.6 seconds, 104.0 seconds, 104.2 seconds, and 103.3 seconds, and reporting version numbers are 51, 52, 52, 53, and 51; Event 2, reporting timestamps are 111.9 seconds, 112.4 seconds, 113.2 seconds, 112.8 seconds, and 112.0 seconds, and reporting version numbers are 53, 53, 54, 52, and 53; Event 3, reporting timestamps are 120.5 seconds, 121.7 seconds, 120.8 seconds, 122.0 seconds, and 121.2 seconds, and reporting version numbers are 54, 55, 54, 55, and 54. Calculate the maximum event lag and step deviation for each of the three events: For the first event, the maximum event lag is 104.2 - 102.1 = 2.1 seconds, and the step deviation is 53 - 51 = 2 seconds; for the second event, the maximum event lag is 113.2 - 111.9 = 1.3 seconds, and the step deviation is 54 - 52 = 2 seconds; for the third event, the maximum event lag is 122.0 - 120.5 = 1.5 seconds, and the step deviation is 55 - 54 = 1 second. Therefore, the maximum event lag is 2.1 seconds, and the maximum step deviation is 2 seconds.

[0058] Extract the maximum event lag consistency risk measure factor and the step deviation consistency risk measure factor from the database.

[0059] It should be noted that the database pre-defines multiple risk measurement factors for consistency risk analysis to reflect the weight distribution of different consistency indicators (maximum event lag and step deviation) in the overall consistency risk score. These risk measurement factors are stored in the database as a normalized weight parameter table, scientifically configured based on historical scheduling anomaly statistics and synchronization deviation impact analysis results from actual production processes. When calculating the consistency risk score, the system can directly retrieve the maximum event lag consistency risk measurement factor and the step deviation consistency risk measurement factor from the database. All of these factors are real numbers ranging from 0 to 1, and their sum is required to equal 1. This ensures that the weight allocation of each indicator in the consistency risk score is both reasonable and rigorous, thereby helping to improve the scheduling system's comprehensive ability to identify multi-source consistency anomalies and the robustness of scheduling strategies.

[0060] Extract the preset reference maximum event lag and reference step deviation from the database.

[0061] A consistency risk metric is introduced, and the normalized results of the maximum event lag and step deviation of each scheduling group are weighted and coupled to obtain the consistency risk score of each scheduling group.

[0062] In a specific embodiment, the consistency risk score for each scheduling group is defined as follows: , where A i Let a be the consistency risk score for the i-th scheduling group, where j=1 is the maximum event lag, j=2 is the step deviation, and a i,1 For the maximum event delay of the i-th scheduling group, a i,2 Let a be the step deviation of the i-th scheduling group. 10 To reference the maximum event lag, a 20 For reference step deviation, x1 is the maximum event lag consistency risk measure factor, and x2 is the step deviation consistency risk measure factor.

[0063] Furthermore, the scheduling advancement strategy for each scheduling group is determined, and the specific determination process is as follows:

[0064] Analyze the consistency risk threshold for each scheduling group.

[0065] If the consistency risk score of a scheduling group is less than or equal to the consistency risk threshold of that scheduling group, it indicates that the process progress, event synchronization, and overall operating status of each moving vehicle within the current scheduling group are all within a controllable range, with no significant synchronization lag or step deviation. The stability and consistency of the scheduling process meet the established process requirements and risk tolerance standards. In this case, no additional abnormal intervention measures are required, and the scheduling advancement strategy for that scheduling group is recorded as continuing to advance according to the set scheduling scheme.

[0066] If the consistency risk score of a certain scheduling group is greater than the consistency risk threshold of that scheduling group, it indicates that there are obvious progress differences or synchronization anomalies among the moving trolleys in the current scheduling group in terms of process advancement and event reporting. Its consistency risk exceeds the normal tolerance range of the system, and there is a potential risk of scheduling out-of-sync, process backlog or anomaly propagation. In this case, the scheduling advancement strategy of the scheduling group is recorded as starting local prediction and the synchronization is marked as abnormal.

[0067] Furthermore, the consistency risk threshold of each scheduling group is analyzed. The specific analysis process is as follows:

[0068] Extract the preset basic consistency risk threshold and anomaly frequency threshold from the database.

[0069] It should be explained that the basic consistency risk threshold is a benchmark value used to initially determine the consistency risk level of scheduling groups. It is typically set by the system developer or process engineer before system deployment, based on historical scheduling anomaly data and synchronization fluctuation statistics from the actual production line. This value, based on statistical analysis, empirical judgment, and simulation evaluation of a large amount of historical data, reflects the upper limit of acceptable risk for group synchronization fluctuations under ideal or steady-state conditions. After system deployment, the basic consistency risk threshold can be dynamically adjusted or optimized according to actual operating conditions. The anomaly frequency threshold is a boundary indicator used to distinguish the anomaly risk status of groups. It is also determined through statistical analysis of the frequency of anomaly events occurring in historical scheduling groups within a certain period. Generally, the maximum anomaly probability that the system can tolerate is selected as the threshold, for example, set to 0.05, 0.10, etc., depending on the risk tolerance of different production lines. Both of these thresholds are stored in the system database in the form of parameter configuration tables, supporting subsequent adjustments and queries as needed.

[0070] The historical anomaly frequency of each scheduling group is obtained. If the historical anomaly frequency of a certain scheduling group is less than or equal to the anomaly frequency threshold, it indicates that the probability of anomaly occurrence in the past operation cycle of the scheduling group is low, the overall scheduling synchronization performance is within a controllable range, the system considers the risk status of the group to be relatively robust, there is no need to relax the basic risk threshold, and it can be appropriately tightened to further improve the scheduling accuracy. Then, the historical anomaly frequency of the scheduling group is used as the query index to query the consistency risk threshold correction factor of the scheduling group from the frequency-risk correction tightening table.

[0071] If the historical anomaly frequency of a certain scheduling group is greater than the anomaly frequency threshold, it indicates that the probability of anomalies occurring in that group within the historical period has exceeded the system's expected risk tolerance limit, suggesting a high scheduling consistency risk for the current group. In this case, to prevent frequent misjudgments of anomalies due to overly tight parameters, the system will correspondingly relax the risk threshold for that group to enhance the system's fault tolerance for anomaly groups and reduce the false alarm rate. The consistency risk threshold correction factor for that scheduling group is obtained by querying the frequency-risk correction relaxation table using the historical anomaly frequency as the query index.

[0072] In a specific embodiment, two mapping tables are pre-established in the database: a frequency-risk correction tightening table and a frequency-risk correction relaxation table. Both tables are indexed by historical abnormal frequencies, with each historical abnormal frequency interval corresponding to a specific consistency risk threshold correction factor. When threshold correction is needed for a specific scheduling group, the historical abnormal frequency of that group is first obtained. If the historical abnormal frequency is less than or equal to the abnormal frequency threshold, the frequency-risk correction tightening table is searched using the historical abnormal frequency of the scheduling group as the index, and the corresponding tightening factor is retrieved as the consistency risk threshold correction factor. For example, if the abnormal frequency is 0.03, the correction factor is 0.8. If the historical abnormal frequency is greater than the abnormal frequency threshold, the relaxation factor is retrieved from the frequency-risk correction relaxation table using the historical abnormal frequency as the index, such as a relaxation factor of 1.3 for an abnormal frequency of 0.15. This lookup method can use interval mapping (such as segmented intervals: [0,0.02], (0.02,0.05], (0.05,0.1], etc.) or continuous variable interpolation to ensure that each abnormal frequency can uniquely correspond to the corresponding correction factor.

[0073] By iterating through each scheduling group, the consistency risk threshold correction factor for each scheduling group is obtained.

[0074] The basic consistency risk threshold is corrected (multiplicative) based on the consistency risk threshold correction factor of each scheduling group to obtain the consistency risk threshold of each scheduling group.

[0075] See Figure 4The diagram shows the flowchart for determining the scheduling optimization strategy for predictive scheduling groups. First, each predictive scheduling group is statistically analyzed to identify the group that initiates local prediction in the current multi-channel maglev production line. Then, the time lag probability and step deviation probability of each predictive scheduling group are calculated. By statistically analyzing the proportion of out-of-limit data points in the lag time series and step deviation series, probability indicators reflecting scheduling synchronization and step consistency are obtained. Next, the system determines whether the time lag probability is greater than or equal to the time lag probability threshold and whether the step deviation probability is greater than or equal to the step deviation probability threshold. If both conditions are met, the step threshold is tightened to optimize the process. The scheduling accuracy is improved. If the above conditions are not met simultaneously, it is further determined whether the time lag probability is greater than or equal to the time lag probability threshold or the step deviation probability is greater than or equal to the step deviation probability threshold. If either condition is met, an over-limit prompt message is generated, and it is determined whether an over-limit prompt message already exists. If the judgment result is yes, the step threshold tightening is performed again. If the judgment result is no, the scheduling scheme is continued. If neither the time lag probability nor the step deviation probability reaches the threshold, the scheduling scheme is directly continued, thereby realizing the adaptive scheduling optimization decision of predictive scheduling grouping.

[0076] When the scheduling advancement strategy is to initiate local prediction, the corresponding scheduling group is recorded as the prediction scheduling group. The prediction scheduling group is statistically analyzed, and the lag time series and step deviation series of each prediction scheduling group are analyzed to determine the scheduling optimization strategy of each prediction scheduling group. Then, the intelligent scheduling optimization of each prediction scheduling group is executed.

[0077] Furthermore, the lag time series and step deviation series of each forecast scheduling group are analyzed. The specific analysis process is as follows:

[0078] The scheduling group whose scheduling advancement strategy is to initiate local prediction is denoted as the prediction scheduling group, and statistics are compiled for each prediction scheduling group.

[0079] For a predictive scheduling group, extract the traceback step count of the predictive scheduling group, and extract the historical reporting records of process switching events sequentially backward from the event-driven synchronization log chain of the predictive scheduling group in chronological order.

[0080] For each process switch event, extract the reporting timestamps and event version numbers of all moving trolleys in the predicted scheduling group under the process switch event from the historical reporting records.

[0081] For each extracted process switching event, the reporting timestamps of all moving trolleys in the prediction scheduling group are statistically analyzed, and the difference between the maximum and minimum timestamps is calculated as the lag time of the process switching event.

[0082] The event version number of each moving trolley is synchronously counted, and the difference between the maximum version number and the minimum version number is calculated as the step deviation of the process switching event.

[0083] The lag time and step deviation of each process changeover event are arranged in chronological order to form a lag time sequence and a step deviation sequence.

[0084] In a specific embodiment, taking predictive scheduling group A as an example, the system sets the traceback step count for this group to 6. The system extracts the historical reporting records of the 6 most recent process switch events sequentially from the event-driven synchronization log chain of predictive scheduling group A. Assuming that predictive scheduling group A contains 5 trolleys, the reporting timestamps and event version numbers of each trolley in these 6 process switch events are as follows: For the first process switch event, the trolley reporting timestamps are 100.5 seconds, 100.5 seconds, 100.5 seconds, 100.5 seconds, and 100.5 seconds, with reporting version numbers of 25, 25, 25, 25, and 25. For the second process switch event, the trolley reporting timestamps are 110.2 seconds, 110.4 seconds, 110.2 seconds, 110.3 seconds, and 110.4 seconds, with reporting version numbers of 26, 28, 27, 26, and 29. For the third process changeover event, the moving trolley reported timestamps of 120.8 seconds, 120.8 seconds, 120.8 seconds, 120.8 seconds, and 120.8 seconds, with version numbers of 29, 29, 29, 29, 29. For the fourth process changeover event, the moving trolley reported timestamps of 131.5 seconds, 132.1 seconds, 132.1 seconds, 131.5 seconds, 132.1 seconds, and version numbers of 30, 30, 31, 31, 31. For the fifth process changeover event, the moving trolley reported timestamps of 140.7 seconds, 140.7 seconds, 140.7 seconds, 140.7 seconds, and 140.7 seconds, with version numbers of 32, 35, 32, 35, 35. In the 6th process switchover event, the moving trolley reported timestamps of 151.0 seconds, 151.4 seconds, 151.0 seconds, 151.0 seconds, and 151.0 seconds, with version numbers of 35, 39, 36, 35, and 37. The statistical results for each process switchover event are as follows: First event: Maximum timestamp 100.5 seconds, minimum 100.5 seconds, lag time 0 seconds; maximum version number 25, minimum 25, step deviation 0. Second event: Maximum timestamp 110.4 seconds, minimum 110.2 seconds, lag time 0.2 seconds; maximum version number 29, minimum 26, step deviation 3. Third event: Maximum timestamp 120.8 seconds, minimum 120.8 seconds, lag time 0 seconds; maximum version number 29, minimum 29, step deviation 0. Fourth time: Maximum timestamp 132.1 seconds, minimum 131.5 seconds, lag time 0.6 seconds; maximum version number 31, minimum 30, step deviation 1. Fifth time: Maximum timestamp 140.7 seconds, minimum 140.7 seconds, lag time 0 seconds; maximum version number 35, minimum 32, step deviation 3. Sixth time: Maximum timestamp 151.4 seconds, minimum 151.0 seconds, lag time 0.4 seconds; maximum version number 39, minimum 35, step deviation 4. Finally, the lag time sequence of the six process switching events is [0 seconds, 0.2 seconds, 0 seconds, 0.6 seconds, 0 seconds, 0.4 seconds], and the step deviation sequence is [0, 3, 0, 1, 3, 4].

[0085] By iterating through each prediction scheduling group, the lag time series and step deviation series of each prediction scheduling group are obtained.

[0086] Furthermore, the scheduling optimization strategy for each predicted scheduling group is determined, and the specific determination process is as follows:

[0087] The time lag probability and step deviation probability of each predictive scheduling group are obtained based on the lag time series and step deviation series of each predictive scheduling group.

[0088] Furthermore, the time lag probability and step deviation probability of each prediction scheduling group are analyzed in detail as follows:

[0089] For a predictive scheduling group, based on the lag time series of the predictive scheduling group, data points with lag times greater than or equal to a preset lag time threshold are extracted and recorded as abnormal lag data points.

[0090] Based on the step deviation sequence of the prediction scheduling group, data points with step deviations greater than or equal to the preset step deviation threshold are extracted and recorded as abnormal step deviation data points.

[0091] In specific embodiments, the lag time threshold and the step deviation threshold are used to determine the maximum allowable time lag and the upper limit of process step difference between the moving trolleys within the scheduling group for each process switching event, respectively. These are generally set by engineering technicians in conjunction with production process requirements and historical data statistics. For example, the lag time threshold can be selected based on the production line's actual tolerance for process synchronization, such as setting it to 0.5 seconds or 1 second, ensuring that most synchronization fluctuations do not affect downstream processes; the step deviation threshold can be set to 2 steps, reflecting the maximum acceptable range for differences in process progress across multiple trolleys. Specific values ​​can be determined through long-term collected scheduling data distribution statistics, correlation analysis of product quality fluctuations and synchronization indicators, and experience optimization during trial operation. Finally, these values ​​are stored in the database as parameter configurations to achieve dynamic recall and maintainability.

[0092] Count the number of abnormally lagging data points and the number of abnormal step deviation data points.

[0093] The total number of data points in the statistical time series is counted, and combined with the number of abnormally lagged data points, the time lag probability of the predicted scheduling group is obtained.

[0094] The total number of data points in the step deviation sequence is counted, and combined with the number of abnormal step deviation data points, the step deviation probability of the predicted scheduling group is obtained.

[0095] Continuing with the above-described predictive scheduling group A as a specific example, the lag time series is [0 seconds, 0.2 seconds, 0 seconds, 0.6 seconds, 0 seconds, 0.4 seconds], and the step deviation series is [0, 3, 0, 1, 3, 4]. Assume the system's preset lag time threshold is 0.5 seconds and the step deviation threshold is 2. Anomaly detection is performed on the lag time series. Each data point is compared with the lag time threshold of 0.5 seconds: the first lag time is 0 seconds, less than 0.5 seconds, not an anomaly; the second is 0.2 seconds, less than 0.5 seconds, not an anomaly; the third is 0 seconds, less than 0.5 seconds, not an anomaly; the fourth is 0.6 seconds, greater than 0.5 seconds, recorded as an abnormal lag data point; the fifth is 0 seconds, less than 0.5 seconds, not an anomaly; the sixth is 0.4 seconds, less than 0.5 seconds, not an anomaly. If the number of abnormally lagging data points is 1 and the total number is 6, then the predicted time lag probability of scheduling group A is approximately 1 / 6 ≈ 0.167. Anomaly detection is performed on the step deviation sequence. Each data point is compared with the step deviation threshold 2: the first step deviation is 0, less than 2, not an anomaly; the second is 3, greater than 2, recorded as an abnormal step deviation data point; the third is 0, less than 2, not an anomaly; the fourth is 1, less than 2, not an anomaly; the fifth is 3, greater than 2, recorded as an abnormal step deviation data point; the sixth is 4, greater than 2, recorded as an abnormal step deviation data point. The number of abnormal step deviation data points is 3 and the total number is 6, therefore the predicted step deviation probability of scheduling group A is 3 / 6 = 0.5.

[0096] By iterating through each prediction scheduling group, the time lag probability and step deviation probability of each prediction scheduling group are obtained.

[0097] Extract the preset time lag probability threshold and step deviation probability threshold from the database.

[0098] In this embodiment, both the time lag probability threshold and the step deviation probability threshold are threshold parameters used to determine the severity of scheduling synchronization anomalies. Their preset values ​​are typically set based on statistical analysis of historical big data from the production line, combined with process tolerance requirements and risk management standards. Specifically, engineers statistically analyze the distribution of time lag probabilities and step deviation probabilities for each group within a large number of scheduling cycles. Based on the process's tolerance for anomalies (e.g., 5%, 10%, 20%), they select probability values ​​that reflect the safety boundary of group synchronization as thresholds. Finally, these two thresholds are stored in a database as a parameter table and can be dynamically adjusted and optimized according to actual operational results.

[0099] If the time lag probability of a certain predictive scheduling group is greater than or equal to the time lag probability threshold, and the step deviation probability is greater than or equal to the step deviation probability threshold, it indicates that the lag phenomenon of synchronization timing and the abnormal proportion of the dispersion of process step progress are both high. The predictive scheduling group has insufficient synchronization and there is a high risk of scheduling out-of-sync or production anomalies. Therefore, the scheduling optimization strategy of the predictive scheduling group is directly recorded as step threshold tightening.

[0100] If the time lag probability of a certain predictive scheduling group is greater than or equal to the time lag probability threshold, or the step deviation probability is greater than or equal to the step deviation probability threshold, an over-limit warning message is generated. The system then checks whether the predictive scheduling group already has an over-limit warning message. If it does, it means that the group has previously been detected as having an anomaly and an over-limit alarm has been triggered, and this anomaly is either continuous or recurring. The system determines this as a secondary over-limit, indicating that the problem may worsen or not be alleviated in the short term. Therefore, the scheduling optimization strategy for this predictive scheduling group is recorded as tightening the step threshold. Otherwise, it indicates that the anomaly in this predictive scheduling group is a first-time over-limit and is not considered high-risk. The scheduling optimization strategy for this predictive scheduling group is recorded as continuing to advance with the set scheduling scheme.

[0101] If the time lag probability of a certain predictive scheduling group is less than the time lag probability threshold and the step deviation probability is less than the step deviation probability threshold, it means that the synchronization timing and process step progress of the current predictive scheduling group are both within the acceptable range of the system, no problem of excessively high abnormality rate is found, the overall production process is stable and the scheduling is orderly, then the scheduling optimization strategy of the predictive scheduling group is directly recorded as continue to advance with the set scheduling scheme.

[0102] Furthermore, the step threshold is tightened, and the specific execution process is as follows:

[0103] The predicted scheduling group whose scheduling optimization strategy is step threshold tightening is denoted as the step tightening scheduling group, and each step tightening scheduling group is statistically obtained.

[0104] For a step-tightening scheduling group, the time lag probability of the step-tightening scheduling group is subtracted from the time lag probability threshold to obtain the time lag deviation probability of the step-tightening scheduling group. The step deviation probability of the step-tightening scheduling group is subtracted from the step deviation probability threshold to obtain the step deviation probability of the step-tightening scheduling group.

[0105] Set the time lag bias probability or step bias probability, which is less than zero, to zero to obtain the final time lag bias probability and the final step bias probability.

[0106] It's important to explain that in the calculation of the step threshold tightening, the time lag deviation probability and the step deviation probability are obtained by the difference between the current probability and the corresponding probability threshold. If the current probability is lower than the threshold, the difference is negative, indicating that the abnormal proportion of the current scheduling group has not yet exceeded the system's set tolerance limit. In scheduling optimization design, a negative deviation does not mean that further tightening of synchronization control is needed; rather, it reflects that the system is in a normal or safe range better than the threshold. To avoid negative deviation values ​​misleading the step threshold tightening algorithm or introducing unreasonable tightening trends, a lower limit zeroing process is adopted, forcibly setting all deviation probabilities less than zero to zero. This ensures that only the excess portion (i.e., the abnormal probability exceeds the corresponding threshold) participates in subsequent step threshold adjustments, improving the targeting and scientific nature of the tightening measures.

[0107] The set of deviation probabilities, consisting of the final time lag deviation probability and the final step deviation probability, is used as the query index to query the step threshold tightening value. Step threshold tightening is then performed, and after the step criterion is met, the scheduled plan continues to be implemented.

[0108] In this embodiment, a multi-dimensional mapping table is set up for step threshold tightening, using the set of deviation probabilities—time lag deviation probability and step deviation probability—as a joint query index. Specifically, the system forms ordered binary pairs of the final time lag deviation probability and the final step deviation probability, and uses these pairs to look up the preset step threshold tightening configuration table in the database. This table can be organized using interval segmentation, continuous interval interpolation, or hierarchical blocks. For example, the rows and columns of the mapping table correspond to two deviation probabilities within different interval ranges, and the corresponding step threshold tightening value is stored at the table's intersection. When the deviation probability input by the system falls within a certain partition, the suggested tightening range can be retrieved, achieving precise matching and dynamic adjustment.

[0109] Continuing with the aforementioned predictive scheduling group A as a specific example, the calculated time lag probability for group A is 0.167, the time lag probability threshold is 0.15, and the difference between the two is 0.017. The step deviation probability is 0.5, the step deviation probability threshold is 0.3, and the difference between the two is 0.2. These two differences are used as the time lag deviation probability and the step deviation probability, respectively. Since both deviation probabilities are positive, no zeroing is required. The system uses these two values ​​to form a deviation probability set as a joint index, i.e., (0.017, 0.2). Using this pair as a parameter, the system queries the preset step threshold tightening configuration table in the database. For example, the configuration table specifies that when the time lag deviation probability is in the range [0, 0.05) and the step deviation probability is in the range [0.15, 0.25), the step threshold tightening value is set to 1 step, which requires that the maximum allowable step difference of the process progress of each moving trolley within the group be tightened from the original 3 steps to 2 steps. Based on this, the system adjusts the stepping threshold of predictive scheduling group A, reducing the maximum stepping difference from 3 steps to 2 steps, and mandating that the maximum process stepping difference for all moving carriages within the group must not exceed 2 steps during the process. Subsequently, the system continues to monitor the process progress of each moving carriage in predictive scheduling group A, and simultaneously determines whether the stepping criterion has been met (i.e., the stepping difference of all moving carriages has returned to the allowable range after tightening). Once the criterion is met, the system automatically resumes production according to the set scheduling plan, achieving a dynamic balance between tightened scheduling and controllable risk.

[0110] like Figure 2 The diagram shown is a structural diagram of an intelligent scheduling and optimization system for a multi-channel maglev production line. The system includes: an initial task execution plan generation module, a scheduling advancement strategy determination module, and a scheduling optimization strategy determination and execution module.

[0111] The initial task execution plan generation module is used to construct a multi-objective model, collect multi-dimensional feature vectors, and output the initial task execution plan of the moving trolley on the multi-channel maglev production line, thereby dividing the moving trolley on the multi-channel maglev production line into multiple scheduling groups.

[0112] The scheduling advancement strategy determination module is used to construct the event-driven synchronization log linked list for each scheduling group, analyze the consistency risk score of each scheduling group, and determine the scheduling advancement strategy for each scheduling group.

[0113] The scheduling optimization strategy determination and execution module is used to record the corresponding scheduling group as the predicted scheduling group when the scheduling advancement strategy is to start local prediction, to count each predicted scheduling group, to analyze the lag time series and step deviation series of each predicted scheduling group, to determine the scheduling optimization strategy of each predicted scheduling group, and to execute the intelligent scheduling optimization of each predicted scheduling group.

[0114] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the flow or function according to the embodiments of the present invention is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. A computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. Semiconductor media can be solid-state drives.

[0115] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0116] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0117] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0118] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0119] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0120] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0121] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0122] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for intelligent scheduling optimization for a multi-lane magnetic levitation production line, characterized in that, The method comprises the following steps: a multi-target model is constructed, a multi-dimensional feature vector is collected, and an initial task execution scheme of a mover trolley of a multi-channel magnetic suspension production line is output, so that the mover trolleys on the multi-channel magnetic suspension production line are divided into multiple scheduling groups; an event-driven synchronization log linked list of each scheduling group is constructed, a consistency risk score of each scheduling group is analyzed, and a scheduling promotion strategy of each scheduling group is determined; when the scheduling promotion strategy is starting local prediction, the corresponding scheduling group is recorded as a prediction scheduling group, each prediction scheduling group is counted, a lag time sequence and a step deviation sequence of each prediction scheduling group are analyzed, a scheduling optimization strategy of each prediction scheduling group is determined, and intelligent scheduling optimization of each prediction scheduling group is performed; the determination of the scheduling promotion strategy of each scheduling group is specifically as follows: the consistency risk threshold of each scheduling group is analyzed; if the consistency risk score of a scheduling group is less than or equal to the consistency risk threshold of the scheduling group, the scheduling promotion strategy of the scheduling group is recorded as continuing to promote with the set scheduling scheme; if the consistency risk score of a scheduling group is greater than the consistency risk threshold of the scheduling group, the scheduling promotion strategy of the scheduling group is recorded as starting local prediction and synchronously marking an exception; the analysis of the consistency risk threshold of each scheduling group is specifically as follows: a preset basic consistency risk threshold and an exception frequency threshold in a database are extracted; the historical exception frequency of each scheduling group is obtained, if the historical exception frequency of a scheduling group is less than or equal to the exception frequency threshold, the historical exception frequency of the scheduling group is taken as a query index, and a consistency risk threshold correction factor of the scheduling group is obtained from a frequency-risk correction tightening table through query; if the historical exception frequency of a scheduling group is greater than the exception frequency threshold, the historical exception frequency of the scheduling group is taken as a query index, and a consistency risk threshold correction factor of the scheduling group is obtained from a frequency-risk correction relaxation table through query; the consistency risk threshold correction factors of each scheduling group are obtained through traversal; the basic consistency risk threshold is corrected based on the consistency risk threshold correction factors of each scheduling group, and the consistency risk threshold of each scheduling group is obtained.

2. The intelligent scheduling optimization method for multi-lane magnetic levitation production line according to claim 1, characterized in that, the construction of the event-driven synchronization log linked list of each scheduling group is specifically as follows: an independent log linked list structure is established for each scheduling group; in each scheduling group, data reported by the mover trolley is received according to a preset collection period, and the data at least includes a mover trolley unique identifier, a current position, a process step number, a local clock, a current task number, and a data version number; when the mover trolley belongs to any event in a preset process switching event set, an event node containing an event type, an event number, a mover trolley unique identifier, and a local clock is generated; the data and the event node are sequentially written into the corresponding log linked list in chronological order, each node includes all the reported data, and a log node sequence arranged in chronological order is formed; the event-driven synchronization log linked list of each scheduling group is obtained through traversal.

3. The intelligent scheduling optimization method for multi-lane magnetic levitation production line according to claim 1, characterized in that, The consistency risk score of each scheduling group is analyzed, and the specific analysis process is as follows: Obtain the historical abnormal frequency of each scheduling group; Use the historical abnormal frequency of each scheduling group as a query index to query the traceable step number of each scheduling group; Based on the traceable step number of each scheduling group, the event reporting record of each scheduling group is traced forward from the event-driven synchronization log chain table of each scheduling group, and the event reporting timestamp sequence and event reporting version number sequence of each scheduling group are obtained; Based on the event reporting timestamp sequence of each scheduling group, the maximum event lag of each scheduling group is obtained; Based on the event reporting version number sequence of each scheduling group, the step deviation of each scheduling group is obtained; A consistency risk measurement factor is introduced, and the normalized results of the maximum event lag and the step deviation of each scheduling group are coupled by weighting to obtain the consistency risk score of each scheduling group.

4. The intelligent scheduling optimization method for multi-lane magnetic levitation production line according to claim 1, characterized in that, The lag time sequence and step deviation sequence of each predicted scheduling group are analyzed, and the specific analysis process is as follows: Scheduling groups with a scheduling promotion strategy of starting local prediction are recorded as predicted scheduling groups, and each predicted scheduling group is counted; For a predicted scheduling group, the traceable step number of the predicted scheduling group is extracted, and the historical reporting records of process switching events are extracted in time sequence and sequentially from the event-driven synchronization log chain table of the predicted scheduling group; For each process switching event, the reporting timestamp and event version number of all mobile carriages in the predicted scheduling group at the process switching event are extracted from the historical reporting records; For each extracted process switching event, the reporting timestamps of all mobile carriages in the predicted scheduling group are counted, and the difference between the maximum timestamp and the minimum timestamp is calculated as the lag time of the process switching event; The event version numbers of each mobile carriage are synchronously counted, and the difference between the maximum version number and the minimum version number is calculated as the step deviation of the process switching event; The lag time and step deviation of each process switching event are arranged in time sequence to form the lag time sequence and step deviation sequence; Each predicted scheduling group is traversed to obtain the lag time sequence and step deviation sequence of each predicted scheduling group.

5. The intelligent scheduling optimization method for multi-lane magnetic levitation production line according to claim 1, characterized in that, The scheduling optimization strategy of each predicted scheduling group is determined, and the specific determination process is as follows: Based on the lag time sequence and step deviation sequence of each predicted scheduling group, the time lag probability and step deviation probability of each predicted scheduling group are obtained; The preset time lag probability threshold and step deviation probability threshold in the database are extracted; If the time lag probability of a predicted scheduling group is greater than or equal to the time lag probability threshold, and the step deviation probability is greater than or equal to the step deviation probability threshold, the scheduling optimization strategy of the predicted scheduling group is directly recorded as step threshold tightening. If the time lag probability of a predicted scheduling group is greater than or equal to the time lag probability threshold value, or the step deviation probability is greater than or equal to the step deviation probability threshold value, generate an out-of-limit prompt information, and query whether the predicted scheduling group already has an out-of-limit prompt information, if the predicted scheduling group already has an out-of-limit prompt information, record the scheduling optimization strategy of the predicted scheduling group as step threshold tightening, otherwise record the scheduling optimization strategy of the predicted scheduling group as continuing to advance with the set scheduling scheme; If the time lag probability of a predicted scheduling group is less than the time lag probability threshold value, and the step deviation probability is less than the step deviation probability threshold value, directly record the scheduling optimization strategy of the predicted scheduling group as continuing to advance with the set scheduling scheme.

6. The intelligent scheduling optimization method for multi-lane magnetic levitation production line according to claim 5, characterized in that, The time lag probability and the step deviation probability of each predicted scheduling group are specifically analyzed as follows: For a predicted scheduling group, based on the lag time sequence of the predicted scheduling group, extract data points with lag time greater than or equal to a preset lag time threshold value, and record them as abnormal lag data points; Synchronously based on the step deviation sequence of the predicted scheduling group, extract data points with step deviation greater than or equal to a preset step deviation threshold value, and record them as abnormal step deviation data points; Count the number of abnormal lag data points and the number of abnormal step deviation data points; Count the total number of data points in the lag time sequence, and combine the number of abnormal lag data points to obtain the time lag probability of the predicted scheduling group; Count the total number of data points in the step deviation sequence, and combine the number of abnormal step deviation data points to obtain the step deviation probability of the predicted scheduling group; Iterate through each predicted scheduling group to obtain the time lag probability and the step deviation probability of each predicted scheduling group.

7. The intelligent scheduling optimization method for multi-lane magnetic levitation production line according to claim 5, characterized in that, The step threshold tightening is specifically executed as follows: Record the predicted scheduling group with the scheduling optimization strategy of step threshold tightening as a step tightening scheduling group, and count to obtain each step tightening scheduling group; For a step tightening scheduling group, difference process the time lag probability of the step tightening scheduling group from the time lag probability threshold value to obtain the time lag deviation probability of the step tightening scheduling group, and difference process the step deviation probability of the step tightening scheduling group from the step deviation probability threshold value to obtain the step deviation deviation probability of the step tightening scheduling group; Set the time lag deviation probability or the step deviation deviation probability less than zero to zero to obtain the final time lag deviation probability and the final step deviation deviation probability; Use the deviation probability set composed of the final time lag deviation probability and the final step deviation deviation probability as a query index to query the step threshold tightening value, thereby executing the step threshold tightening, and continuing to advance with the set scheduling scheme after the step criterion is met.

8. An intelligent scheduling optimization system for a multi-lane maglev production line for implementing the intelligent scheduling optimization method for a multi-lane maglev production line according to any one of claims 1-7, characterized in that, The system comprises an initial task execution scheme generation module, a scheduling advancement strategy determination module, and a scheduling optimization strategy determination and execution module; The initial task execution scheme generation module is configured to construct a multi-objective model, collect a multi-dimensional feature vector, and output an initial task execution scheme of a mover trolley of a multi-channel magnetic suspension production line, thereby dividing the mover trolley on the multi-channel magnetic suspension production line into multiple scheduling groups. The scheduling promotion strategy determination module is configured to construct an event-driven synchronization log linked list of each scheduling group, analyze a consistency risk score of each scheduling group, and determine a scheduling promotion strategy of each scheduling group. The scheduling optimization strategy determination and execution module is configured to, when the scheduling promotion strategy is starting local prediction, mark a corresponding scheduling group as a prediction scheduling group, count each prediction scheduling group, analyze a lag time sequence and a step deviation sequence of each prediction scheduling group, determine a scheduling optimization strategy of each prediction scheduling group, and thereby execute intelligent scheduling optimization of each prediction scheduling group.

Images