RGV guide vehicle intelligent scheduling method

Abstract

The application provides an RGV guide vehicle intelligent scheduling method, comprising: initializing intelligent machining system parameters, selecting a scheduling mode according to a material process type, generating a candidate machine tool set and constructing a candidate path set, evaluating the total time consumption of each candidate path based on a path time consumption calculation model, selecting the path with the shortest time consumption as the scheduling scheme, if a CNC machine tool fault is detected, removing the faulty machine tool from the candidate set, and re-including the machine tool in the scheduling sequence after maintenance is completed, controlling the RGV guide vehicle to perform movement, feeding and discharging, and cleaning operations according to the selected path, and cyclically performing until a system time threshold is reached; the application optimizes the scheduling strategy through dynamic programming and the single-parent genetic algorithm for single-process, double-process, and machine tool fault scenarios, realizes efficient path planning of the RGV guide vehicle, reasonably allocates machine tool resources, reduces machine tool idle time, and improves material yield, to solve the problems of low efficiency, insufficient resource utilization, and weak interference response in traditional scheduling.

Description

Technical Field

[0001] This invention relates to the field of workshop scheduling technology, and in particular to an intelligent scheduling method for RGV guided vehicles. Background Technology

[0002] With the advancement of industrial development, intelligent manufacturing has become the core direction for the transformation and upgrading of the manufacturing industry. Among them, the scheduling optimization of intelligent processing systems is a key link to improve production efficiency. Early industrial production relied on manual scheduling, with workers manually operating material handling and machine tool start-up and shutdown. This not only resulted in high labor costs but also easily led to efficiency fluctuations and accuracy errors due to fatigue. At the end of the 20th century, automation technology was introduced into the industrial field. The combination of conveyor belts and robotic arms initially realized the automation of the production process. However, the scheduling logic was simple and relied on fixed timing, making it difficult to cope with complex scenarios involving multiple machine tools and multiple processes.

[0003] In recent years, intelligent scheduling technology has developed rapidly, giving rise to scheduling models based on rules and heuristic algorithms (such as genetic algorithms and simulated annealing). Rule-based scheduling (such as "first-come, first-served" and "shortest processing time priority") is simple to implement, but it is prone to getting trapped in local optima. Heuristic algorithms have achieved certain results in job shop scheduling problems by optimizing scheduling schemes through global search. However, their adaptability is still insufficient when facing dynamically changing production environments (such as machine tool failures and process switching).

[0004] In current intelligent manufacturing systems, RGV guided vehicles serve as the core of material transfer, and their scheduling efficiency directly determines the overall performance of the production system. Existing technologies suffer from three prominent problems: First, in single-process scheduling, RGV path planning only considers immediate tasks and fails to anticipate future demands, leading to wasted movement time and high machine tool idle rates. Second, in dual-process scheduling, machine tool process allocation relies on experience and fails to optimize the quantity and location distribution based on processing time ratios, resulting in poor process connections, stockpiling of semi-finished materials, or idle machines in the second process. Third, they have weak fault-response capabilities; traditional scheduling often employs simple replacement strategies when machine tools malfunction, without dynamically adjusting paths and process allocations, easily causing production interruptions and affecting the stability of finished product output. Furthermore, most scheduling models lack efficiency evaluation and robustness verification mechanisms, making it difficult to quantify their practical application value.

[0005] These problems restrict the production efficiency and anti-interference ability of intelligent processing systems. Therefore, there is an urgent need in this field for a scheduling method that takes into account dynamic programming, intelligent optimization and fault adaptation. Summary of the Invention

[0006] This invention aims to provide an intelligent scheduling method for RGV guided vehicles. Targeting scenarios such as single-process, dual-process, and machine tool failures, it optimizes scheduling strategies through dynamic programming and a parent genetic algorithm to achieve efficient path planning for RGV guided vehicles, rationally allocate machine tool resources, reduce machine tool idle time, increase finished product output, and enhance the system's robustness under fault conditions. This addresses the problems of low efficiency, insufficient resource utilization, and weak interference response in traditional scheduling, thus meeting the high-efficiency production requirements of intelligent machining systems.

[0007] This invention provides an intelligent scheduling method for RGV (Remote Guided Vehicle) vehicles, comprising the following steps:

[0008] Step 1: Initialize the intelligent machining system parameters, including the RGV guide car position, CNC machine tool status, material process type, and system time threshold;

[0009] Step 2: Select the scheduling mode according to the material process type. For single-process material processing, a dynamic shortest path planning model is used. For dual-process material processing, the optimal quantity ratio and position distribution of machine tools for the two types of processes are first determined based on the single-parent genetic distribution optimization model, and then an improved dynamic path planning model is used.

[0010] Step 3: When the RGV guide vehicle is in a stopped and waiting state, collect the remaining task time of each CNC machine tool in real time, generate a candidate machine tool set and construct a candidate path set;

[0011] Step 4: Evaluate the total time of each candidate path based on the path time calculation model, and select the path with the shortest time as the scheduling scheme.

[0012] Step 5: If a CNC machine tool malfunction is detected, the malfunctioning machine tool is removed from the candidate set and will be reinstated into the scheduling sequence after repairs are completed.

[0013] Step six: Control the RGV guide vehicle to perform movement, loading and unloading, and cleaning operations along the selected path. After completion, return to step three and repeat until the system time threshold is reached.

[0014] Compared with the prior art, the beneficial effects of this application are as follows:

[0015] 1. This application uses a dynamic shortest path planning model to predict the k fastest-finishing machine tools and generate the optimal path, avoiding blind movement of the RGV guide car; in single-process scenarios, the machine tool is cyclically scheduled according to its number to reduce ineffective movement time; in dual-process scenarios, the machine tool distribution and process ratio are optimized based on a single-parent genetic algorithm, combined with alternating scheduling to reduce the average idle time of the machine tools, improve resource utilization, and increase the output of finished products.

[0016] 2. This application introduces a prediction mechanism, in which the RGV moves to the next task point in advance when there is no demand signal, saving waiting time; when multiple machine tools send signals at the same time, the optimal solution is selected through the path consumption time calculation model to avoid local optima; when a fault occurs, the faulty machine tool is automatically removed and its state is reset, and it is reinstated after repair. 100 fault simulation experiments show that the fluctuation of the finished product output is less than 5%, and the robustness is significantly better than that of traditional static scheduling.

[0017] 3. The single-parent genetic algorithm in this application uses operators such as elite retention and rearrangement mutation to iteratively optimize the distribution of dual-process machine tools and output the optimal solution that meets the time ratio constraint. This solves the problem of reliance on experience in manual allocation, makes the process connection smoother, and improves the efficiency of semi-finished material transfer.

[0018] 4. The model parameters (such as movement time and processing time) of this application can be adjusted according to the actual system and are compatible with CNC machine tools and RGV equipment of different specifications; the efficiency evaluation process quantifies the performance through average idle time and fault simulation, providing a clear reference for actual production, and can be directly applied to smart workshops in multiple fields such as automobile manufacturing and electronic processing.

[0019] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0020] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0021] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0022] Figure 1 This is a flowchart illustrating an intelligent scheduling method for RGV guided vehicles provided in an embodiment of this application;

[0023] Figure 2 This is a schematic diagram illustrating the underutilization of lathe resources in an RGV guided vehicle intelligent scheduling method provided in this application embodiment;

[0024] Figure 3 This is a flowchart of the dynamic shortest path planning in a single-process case in an RGV guided vehicle intelligent scheduling method provided in an embodiment of this application;

[0025] Figure 4 This is a basic flowchart of the two-stage material processing operation in an RGV guided vehicle intelligent scheduling method provided in this application embodiment;

[0026] Figure 5 This is a schematic diagram illustrating the simulation results of the number of finished materials in one process of an RGV guided vehicle intelligent scheduling method provided in this application embodiment;

[0027] Figure 6 This is a schematic diagram illustrating the simulation results of the number of finished materials in two processes of an RGV guided vehicle intelligent scheduling method provided in this application embodiment. Detailed Implementation

[0028] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0029] Example 1:

[0030] Please refer to Figure 1-4 This invention provides an intelligent scheduling method for RGV (RGV) guided vehicles, comprising the following steps:

[0031] Step 1: Initialize the intelligent machining system parameters, including the RGV guide car position, CNC machine tool status, material process type, and system time threshold;

[0032] The system parameters include RGV travel time parameters. CNC machine tool process time Loading and unloading time Cleaning time and the maximum working time of each shift; the CNC machine tool status is determined by the remaining task time model. Dynamic updates include processing status, idle status, and fault repair status.

[0033] Specifically, the intelligent machining system includes 8 CNC machine tools, 1 RGV guided vehicle, a loading / unloading conveyor belt, and a linear track. During initialization, the RGV guided vehicle is in its initial position. (correspond and (Right in the middle); The system time threshold is set to the working time of one shift, i.e., 8 hours (28800 seconds);

[0034] Scenario 1: Inefficient use of lathe resources

[0035] First, let's consider a special case: before the RGV guide car finishes feeding the i-th machine tool, the CNC machine tool that has already been loaded has completed its work and sent a demand signal. In order to maximize work efficiency, i.e., the number of materials produced, the RGV guide car will return to the CNC machine tool in front. That is, the RGV guide car will continuously cycle and schedule among the adjacent CNC machine tools, while the machine tools behind will remain idle.

[0036] like Figure 2 As shown, the RGV guided vehicle first loads material onto CNC1 machine tool, while the other machine tools remain idle and waiting to be loaded. Subsequently, the RGV guided vehicle loads material onto the CNC machine tools sequentially according to their numbers. While loading material onto CNC6 machine tool, CNC1 machine tool completes its processing and simultaneously sends a demand signal to the RGV guided vehicle. Finally, because the system needs to process as much material as possible, after completing the loading task onto CNC6 machine tool, the RGV guided vehicle will move to CNC1 machine tool for loading and unloading operations, instead of moving to CNC7 machine tool for loading. Furthermore, while loading material onto CNC1 machine tool, CNC2 machine tool completes its processing.

[0037] The RGV will repeat the above steps, scheduling between the CNC1 and CNC6 machines, while the CNC7 and CNC8 machines will remain idle.

[0038] Based on the above analysis, when CNC1 completes material processing and sends a demand signal to the RGV guided vehicle, the RGV will immediately return after completing its current task. The system will continuously perform cyclic scheduling, and subsequent machine tools will remain idle. Therefore, before CNC1 completes material processing, the RGV has not yet been scheduled between CNC7 and CNC8. The required relationships for each variable are calculated as follows:

[0039]

[0040] in The floor number where the RGV guide vehicle is located. =1,2,3 The time required for the RGV guided vehicle to load and unload materials for the CNC1 machine tool. The time required for the RGV guided vehicle to load and unload materials for the CNC2 machine tool. For material cleaning time, The time required to process materials in one process.

[0041] Scenario 2: Simultaneous receipt of demand signals

[0042] After the intelligent machining system is powered on and started, the RGV guided vehicle begins to load materials onto the CNC machine tools in sequence according to their numbers. At some point, before the RGV guided vehicle finishes a series of loading and unloading operations on a certain CNC machine tool, it may receive demand signals from multiple CNC machine tools. Therefore, when the RGV guided vehicle finishes its work, the question of which CNC machine tool to select as the next work target arises, i.e., the decision-making problem of the work target.

[0043] Scenario 3: The RGV guide vehicle predicts the route.

[0044] Under normal operating conditions, if a CNC machine tool is idle, it sends a loading request signal to the RGV guided vehicle; otherwise, the CNC machine tool is in machining operation and sends a request signal to the RGV guided vehicle immediately upon completion of the machining operation. Therefore, when no CNC machine tool sends a request signal to the RGV guided vehicle, the RGV guided vehicle will wait in place for instructions before moving to the target location. During this process, the movement time is actually wasted time in the production process. Figure 3 As shown, if the RGV guided vehicle can make certain "predictions," that is, predict the next task target when there is no demand signal and move to the corresponding CNC machine tool in advance, it can save a certain amount of time and improve production efficiency.

[0045] The intelligent machining system schedules the RGV guided vehicle based on the working status of each machine tool. In the process of dynamic scheduling, two scheduling subjects are involved—the RGV guided vehicle and the CNC machine tool. The model needs to conduct a reasonable analysis of the status of these two subjects in order to construct a reasonable scheduling model. Therefore, we take the RGV guided vehicle and the CNC machine tool as the center, analyze the variables that will affect the scheduling results in the dynamic scheduling process, and parameterize the variables to establish a mathematical model.

[0046] The specific system parameters are: RGV movement time parameters. (Time required to move i units, CNC machine tool process time includes single-process time. The first step of the two-stage process takes a long time. The second time Loading and unloading time (The loading and unloading times for machine tools with different numbers differ, such as the difference between odd-numbered and even-numbered machine tools); Cleaning time (Time for cleaning and transporting finished materials);

[0047] More specifically, because the RGV guided vehicle can only perform one action at a time (waiting, moving left and right, loading and unloading, cleaning), and can only perform loading and unloading and cleaning operations when it moves in front of the CNC machine tool.

[0048] (1) The RGV guide vehicle "stops waiting".

[0049] When the RGV guided vehicle is in the "stop waiting" state, it can respond to the demand command in a timely manner and move to the corresponding CNC machine tool. When it is not in the "stop waiting" state, the demand command issued by the CNC machine tool will not affect the state of the RGV guided vehicle. The RGV guided vehicle will still execute the work task corresponding to the accepted command until the task is completed before responding to the demand command.

[0050] When performing intelligent scheduling, if calculations are performed in the "non-stop waiting" state of the RGV guide vehicle, the RGV guide vehicle cannot respond to the instruction in a timely manner. Furthermore, new demand instructions may be generated at any time in the "non-stop waiting" state, affecting the scheduling result. Therefore, performing calculations in the "non-waiting" state of the RGV guide vehicle will result in a waste of computing resources. In the model, it is assumed that the scheduling instruction can only be executed when the RGV guide vehicle is in the "stop waiting" state. The dynamic scheduling of the RGV guide vehicle does not need to consider the situation in the "non-stop waiting" state of the RGV guide vehicle.

[0051] (2) RGV guide vehicle movement

[0052] The RGV guide vehicle can move left or right. The direction from CNC1 to CNC7 (moving right) is defined as positive, and the direction from CNC7 to CNC1 (moving left) is defined as negative. Therefore, when the RGV guide vehicle receives the instruction to "move i units left / right" at time t, the next "stop waiting" state of the RGV guide vehicle will be at the following time point:

[0053]

[0054] in, The time required for the RGV guide vehicle to move i units, and ;exist At the corresponding time point, the location of the guide vehicle was:

[0055]

[0056] Clearly, the movement of the RGV guide vehicle should be within the working area, that is: , =0,1,2,3;

[0057] (3) RGV guided vehicle loading and unloading, cleaning work

[0058] The different processing steps and parameter values ​​required for a given task will affect the entire scheduling process. Therefore, the following discussion will focus on the general case that meets the following conditions:

[0059] The time taken for a CNC machine tool to process raw materials is much longer than the time taken for an RGV guided vehicle to move.

[0060] The time difference between loading odd-numbered and even-numbered machine tools at the same location is much smaller than the time it takes for the guide car to move.

[0061] At this point, we only need to classify the tasks of loading, unloading, and cleaning the RGV guided vehicle according to the number of processing steps.

[0062] 1) Material processing operations of a single process

[0063] When the shift begins, the RGV guide vehicle is located After the first loading is completed on an odd-numbered machine, there are two possible options for the next loading:

[0064] l Load the even-numbered machine tool at the same position.

[0065] After moving 1 unit, feed the material onto the odd-numbered machine tool.

[0066] This means calculating the minimum of the two loading times to decide on the next task. The specific expression is:

[0067]

[0068] Since the model assumes that the time difference between loading and unloading odd-numbered machine tools at the same location is much smaller than the time taken by the moving guide car, the first option can achieve the shortest loading time between two adjacent loading operations. If this assumption is not met, the RGV guide car will always choose the odd-numbered guide car as the next task target after each loading operation, and the even-numbered guide car will always be idle, which does not meet the actual situation. Therefore, this assumption is reasonable.

[0069] Organizing the above ideas, we obtain the shortest material loading solution for the RGV guide vehicle under this condition:

[0070] S1, the RGV guide car grabs raw material from the feeding conveyor belt and feeds it to the odd-numbered machine tool at the current position. After completion, it grabs raw material from the feeding conveyor belt again and feeds it to the even-numbered machine tool at the current position. No cleaning time is generated during this round.

[0071] S2, if the CNC1 machine tool has not completed the loading task at this time, the trolley will move one step forward to the right;

[0072] S3, repeat S1 and S2 until the trolley reaches the position, otherwise the RGV guide car returns to the middle of CNC1 and CNC2 (i.e., returns to the initial position, at which point the first round of loading is completed);

[0073] S4: Grab raw material from the feeding conveyor belt, then unload and load (simultaneously) the odd-numbered machine tool at the current position. Afterward, clean the processed material and place it onto the unloading conveyor belt. Perform the same operation on the even-numbered machine tool at the current position.

[0074] S5, execute the path scheduling algorithm (see step two);

[0075] S6, having reached the maximum working time limit for the shift, the RGV guide vehicle stops moving.

[0076] 2) Material processing operations in both processes

[0077] When the shift begins, the RGV guide vehicle is located Since the two processing tasks are performed on different machine tools, the scheduling algorithm is highly related to the number of machine tool types and their location distribution.

[0078] If only one machine tool is used for the first process, multiple machine tools will be idle and waiting to be processed. Furthermore, if the only machine tool is located at the very edge, excessive material handling time will be consumed. These issues will be considered in a later model, and will not be elaborated here.

[0079] S1, the RGV guide car moves to the nearest first CNC machining center. i In front of the machine tool;

[0080] S2, at this time the RGV guide vehicle grabs raw material from the feeding conveyor belt and feeds it to the machine tool in S1. At this time there is no material on S1, so there is no need to unload or clean it.

[0081] S3, if there are still first-process machine tools that have not been loaded with materials, then repeat S1 and S2;

[0082] S4, when all the first-process machine tools have finished loading, the trolley goes to the first-process machine tool that has finished processing the earliest, and waits for unloading, loading, and cleaning of the finished product;

[0083] S5, execute the path scheduling algorithm (see step two);

[0084] S6, having reached the maximum working time limit for the shift, the RGV guide vehicle stops moving.

[0085] At each step in the process, the RGV guided vehicle needs to perform operations such as grabbing raw materials, unloading, loading, moving to the second process machine tool for unloading (if there are finished materials on the machine tool), loading, cleaning the finished materials, and placing the finished materials when performing tasks on the machine tool of the first process.

[0086] Because the processing time for materials varies across different processes, and the processes of cleaning and transferring materials result in the RGV guided vehicle receiving an increasing number of "first process task completion signals" and "second process task completion signals," making reasonable decisions regarding the RGV guided vehicle's tasks, taking into account the constraints of different processes, is crucial for intelligent scheduling.

[0087] CNC machine tool status is determined by the remaining task time model. Dynamic update: During processing, Decrease over time (decreases by 1 per unit of time); during idle state, During fault repair, Replaced by repair time During the maintenance period, updates will be made according to the processing status rules until... This indicates that the repair is complete; the material processing type is divided into single process (which can be completed by any CNC machine tool) and double process (which requires two different CNC machine tools to process in sequence).

[0088] More specifically, CNC machine tools only exist in two states: working and stopped / waiting. The working state of the CNC machine tool directly affects the scheduling scheme of the RGV guided vehicle. Therefore, we establish parameters related to the RGV guided vehicle scheduling process based on the working state of the CNC machine tool.

[0089] (1) Working status of CNC machine tool

[0090] Indicates CNC i The remaining time for the machine tool to execute its current task at time i; if the machine tool is not currently executing a machining task or the machining task has been completed, then... =0; if the machine tool malfunctions, the repair time will be used. Remaining time for the current machining task on the alternative machine tool:

[0091] (2) Changes in the working status of CNC machine tools

[0092] 1) RGV guided vehicle loading operation

[0093] Only when the machine tool's status code When =0, CNC i The machine tool is not performing a task or the current task has been completed. At this time, the CNC... i The machine tool sends a demand signal to the RGV guided vehicle. At time t, if the RGV guided vehicle assigns the j-th machining operation to the i-th machine tool, then the remaining time for the CNCi machine tool to complete the current machining task is... for:

[0094]

[0095] Among them, when the current shift performs a material processing operation for one process, the following settings are made: ;

[0096] 2) Changes in the working status of CNC machine tools over time

[0097] Indicates CNC i The remaining time for the machine tool to perform its current task at time i; the initial state is as follows: , =0,1,2,3 The time initially decreases with time t, indicating that the machine tool is currently performing a task. At time t+1, the remaining time for the machine tool to complete the current machining task. Update according to the following rules:

[0098]

[0099] in, Indicates CNC i The remaining time for the machine tool to complete the current machining task at time t;

[0100] 3) CNC machine tool malfunction

[0101] If the machine tool malfunctions at time t, then in the subsequent... During this time, the machine tools will undergo manual maintenance. At this time, the CNC... i The remaining time for the machine tool to complete the current machining task is:

[0102]

[0103] When repairing this machine tool, the repair work is also considered as a machining task of the machine tool, and is still treated as such. Update the remaining processing time of the machine tool when... When it becomes zero, it is considered CNC. i Machine tool repair completed.

[0104] Step 2: Select the scheduling mode according to the material process type. For single-process material processing, a dynamic shortest path planning model is used. For dual-process material processing, the optimal quantity ratio and position distribution of machine tools for the two types of processes are first determined based on the single-parent genetic distribution optimization model, and then an improved dynamic path planning model is used.

[0105] The dynamic shortest path planning model for single-process materials includes: selecting k fastest-completing CNC machine tools based on the remaining task time model to form a candidate machine tool set; generating a candidate path set by performing full permutations on the candidate machine tools, with each path containing an ordered sequence of machine tool operations.

[0106] The single-parent genetic distribution optimization model for dual-process materials includes: representing the process type allocation scheme of each CNC machine tool through chromosome encoding; using the maximum material output within a shift as the fitness function, iteratively optimizing the machine tool distribution using rearrangement mutation operators, rotation operators, and probability exchange operators; and outputting the optimal machine tool distribution scheme that satisfies the process time ratio constraint.

[0107] The operation of the single-parent genetic distribution optimization model includes: the elite retention mechanism replicates the individual with the highest fitness to the next generation; the chromosome coding order is randomly adjusted by rearrangement mutation operator; a new solution is generated by circular shifting by rotation operator; and the process type codes of any two machine tools are swapped by probability swap operator.

[0108] The improved dynamic path planning model includes two scheduling modes: a 1-2-1 scheduling mode, which uses the first process machine tool as the starting and ending point of the path and inserts the second process machine tool in the middle; and a 2-1-2 scheduling mode, which uses the second process machine tool as the starting and ending point of the path and inserts the first process machine tool in the middle; the scheduling mode is automatically selected according to the ratio of the number of machine tools of the two types of processes.

[0109] Specifically, during single-process machining, the dynamic shortest path planning model collects the remaining time of each CNC machine tool at time t. The candidate set is formed by selecting the k minimum values ​​of the machine tools (i.e., the machine tools corresponding to the k minimum remaining task times). (e.g., k=3), generate all permutations of machine tools within the candidate set. Candidate path set Each path is an ordered sequence of machine tool operations (e.g., ).

[0110] In dual-process machining, the machine tool distribution is first determined using a single-parent genetic distribution optimization model: In chromosome coding, if the gene value belongs to... The corresponding machine tool is the first process ( (This refers to the number of machine tools used in the first process), otherwise it refers to the number of machine tools used in the second process. Fitness is the maximum number of units produced within a given shift. During iteration, 20% of elite individuals are retained for the next generation. The distribution is optimized using rearrangement operators (randomly rearranging the coding order), rotation operators (circularly cutting the coding to generate new solutions), and probabilistic swap operators (randomly swapping two codes with a 50% probability of continuing the swap). The final output satisfies... The optimal distribution.

[0111] In the improved dynamic path planning model, the 1-2-1 pattern is used in the first machine tool operation. and Insert a second machine tool for the intermediate process to make the path Minimum processing time; 2-1-2 mode on the second machine tool. and Insert the first machine tool in between to make the path The shortest time, according to and Choose the optimal ratio mode.

[0112] Step 3: When the RGV guide vehicle is in a stopped and waiting state, collect the remaining task time of each CNC machine tool in real time, generate a candidate machine tool set and construct a candidate path set;

[0113] The candidate machine tool set is generated using a minimum value selection model, where k is an integer greater than 1; the candidate path set contains all permutations and combinations of candidate machine tools. The value of k ranges from 3 to k to 5.

[0114] Specifically, after the RGV guide vehicle completes its current task, it enters a stop-and-wait state, at which point the remaining time of the eight CNC machine tools is collected in real time. The three machine tools with the shortest remaining time are selected as the candidate machine tool set using the minimum value selection model of the first k (k=3). Perform full permutations (6 combinations) on the 3 machine tools in the candidate set to generate a candidate path set. Each path is an ordered sequence of operations for 3 machine tools (e.g., ), ensuring that all possible job sequences are covered.

[0115] Step 4: Evaluate the total time of each candidate path based on the path time calculation model, and select the path with the shortest time as the scheduling scheme.

[0116] Specifically, the path time calculation model evaluates the total time of each candidate path in the following way: for the j-th machine tool in the path Calculate RGV from the current position Time to move to the machine tool position ( (For the machine tool location), then calculate the total working time at that machine tool. (Including waiting, loading / unloading, and cleaning time), the total time for each path is... The sum of the total time. Select the path with the shortest total time as the scheduling scheme. If there are multiple optimal paths (such as those with the same time), prioritize the path with the odd-numbered machine tool.

[0117] More specifically, at time t, if the RGV guide vehicle is in a "stop waiting" state or in the next second, then path planning can be performed on the RGV guide vehicle, and instructions can be issued through the scheduling system to make the RGV guide vehicle immediately go to the next location to perform the task.

[0118] If we only consider the next machine tool that is about to finish its work and schedule the RGV guide car in advance, we are likely to fall into a "local optimum". Therefore, this embodiment uses dynamic programming to predict the k machine tools that will finish their processing tasks first, and then finds the shortest path among them. The starting position of the shortest path is used as the target position for scheduling the RGV guide car. After each RGV guide car completes its task, the previous steps are repeated. This approach can largely avoid falling into a "local optimum".

[0119] like Figure 4 As shown, the construction process of the single-process path planning model is as follows:

[0120] definition for Location of the lathe:

[0121]

[0122] At time t, the remaining time for each machine tool to complete its current machining task is as follows: , ,..., The candidate set is formed by selecting the k smallest values ​​(the k machine tools that complete the current machining task the fastest). ;

[0123] This represents the set of the first k minimum values ​​(including duplicate values). If these k times are all the same, it means that at a certain point in time, multiple machine tools will complete the task simultaneously.

[0124] Candidate set at time t The machine tools in the candidate set are arranged arbitrarily and formed in sequence. There are 10 different candidate paths, which together form a candidate path set. ;

[0125] At time t, the RGV guide vehicle selects the i-th path from the candidate path set, denoted as . And record for The j-th machine tool in the path. The decision objects of this model are these candidate paths.

[0126] Based on model analysis, the decision objective is to select a candidate path that minimizes the time required for the RGV guided vehicle to complete these machine tool tasks. The first machine tool on the shortest path becomes the next work point for the RGV guided vehicle. The relevant expressions for the decision objective include:

[0127] The objective function is calculated as follows: the RGV guide vehicle follows the candidate path. Time required to complete the task;

[0128] Candidate paths The j-th machine tool in the process;

[0129] The time required for the RGV guide vehicle to move to the j-th position on the decision path;

[0130] The total time for the RGV guide vehicle to wait and perform operations at location j;

[0131] To guide the car from Position moved to The unit that needs to be moved from the location of the machine tool;

[0132] Move the RGV guide vehicle The time required by the unit;

[0133] The time when the RGV guide vehicle moves to the j-th location.

[0134] for machine tools in The remaining time to complete the current processing task;

[0135] This refers to the time required for one loading and unloading operation of the machine tool.

[0136] Finished material cleaning and transportation time;

[0137] At time t, the RGV guide vehicle departs from... Location to depart Time required to locate the machine tool: ;

[0138] Subsequently Total time spent waiting and performing operations at the machine tool: ;

[0139] At this point, the system uptime is updated to: ;

[0140] During time updates, relevant parameters are updated synchronously. Iteration begins from j=1, yielding the following expression for the objective function: ;

[0141] For each RGV guide vehicle in a "stop waiting" state, or entering a "stop waiting" state in the next second, the next work point is predicted according to the above model. The goal of the prediction is to ensure that: ;

[0142] However, the optimal solution is not unique in many cases. For situations where multiple solutions reach the optimal solution, the following rule is specified: if the only difference between the optimal solutions is the order of loading and unloading materials before and after the odd and even numbers, then the odd-numbered solutions will be processed first.

[0143] like Figure 5 As shown, the construction process of the single-process path planning model is as follows:

[0144] Determine the appropriate number / proportion of machine tool types; the arrangement of different types of machine tools on the production line; and maximize the utilization rate of machine tool resources for processing the second step.

[0145] The ratio of machine tools for the two types of processes is closely related to the processing time of the two processes. For example, if the processing time for the first process is longer than that for the second process, then more machine tools of the first process type should be allocated. Under stable conditions, the ratio of machine tools for the two types of processes should be the ratio of their processing times. However, since the ratio of processing times is not necessarily an integer, and the material transfer process also takes time when the number of machine tools of the two types is unequal, therefore, let... The first process of machining on a Taiwanese machine tool. The second process of machining on the Taiwanese machine tool is suitable. and The following relationship should be satisfied:

[0146]

[0147] In this operating system, the working efficiency of the machine tool in the second process directly affects the output of finished products. The working path of the RGV guided vehicle should alternate between the first and second processes. To fully utilize machine tool resources and increase output, the single-process path planning model described above is extended to a 1-2-1 scheduling model and a 2-1-2 scheduling model. For different numbers / proportions of machine tools in different processes, the model with better performance is selected from these two scheduling models.

[0148] The set of machine tools that perform the first operation is The set of machine tools that perform the second process is ; The two machine tools with the shortest remaining time to complete the current machining task are respectively , ,and ; The two machine tools with the shortest remaining time to complete the current machining task are respectively , And there are

[0149] The single-process path planning model described above can be viewed as a 1-1-1 scheduling model, which selects three machine tools for the first process to minimize the time required to complete the path. The 1-2-1 scheduling model, on the other hand, aims to process as much material as possible in the first process, inserting a second-process machine tool between two first-process machine tools to minimize the time required to complete the path. Select a second machine tool to complete the path. The total time taken was the shortest.

[0150] Similar to the 1-2-1 scheduling model, the 2-1-2 scheduling model can be viewed as having the planning objective of processing as much material as possible in the second process. This is considered as inserting a machine tool for the first process between two machines working on the second process, minimizing the time required to complete the path. That is, in... Select a machine tool for the first process to complete the path. The total time taken was the shortest.

[0151] These paths are used as the candidate path set in the single-process path planning model described above. The optimal path is determined by applying this dynamic programming model.

[0152] During iterative path planning, the starting point of the optimal path is the endpoint of the path decided in the previous stage. Therefore, for each decided path, the RGV guided vehicle will perform tasks on the first two machine tools in the optimal path. After that, the RGV guided vehicle will be idle again and dynamic planning will be performed again.

[0153] Furthermore, since the number of machine tools for the two types of processes are respectively and Then there are a total of There are several possible distribution scenarios. When the total number of machine tools is small, each possible distribution can be enumerated and the optimal total number of processed materials under that machine tool arrangement scheme within a shift can be calculated to select a better arrangement scheme. However, when the total number of machine tools and processes increases, the enumeration method is no longer applicable. Therefore, this embodiment also proposes to use a single-parent genetic algorithm for optimal distribution planning:

[0154] For each individual X, the code in each dimension determines the machine tool's working type based on its value range. If... This indicates that the CNC i The machine tool is for the first process; otherwise, it is for the second process.

[0155] Fitness is the maximum number of finished products that a machine tool can process in a shift under the current arrangement. The higher the fitness, the better the distribution plan.

[0156] At the start of the algorithm, initial values ​​(initial allocation schemes) are randomly generated. These allocation schemes are then simulated and scheduled to calculate the fitness function value (number of units produced). The better allocation schemes are selected and retained. These better allocation schemes have a certain probability of mutation and swapping, thus generating new solutions. This process is repeated until the termination condition is met. At this point, the algorithm returns an optimal solution that has stabilized after multiple generations of evolution. This optimal solution represents the optimal allocation scheme.

[0157] Step 5: If a CNC machine tool malfunction is detected, the malfunctioning machine tool is removed from the candidate set and will be reinstated into the scheduling sequence after repairs are completed.

[0158] The fault handling includes replacing the remaining task time of the faulty machine tool with a maintenance time model. During maintenance, the machine tool will not participate in path planning, and will be reset to its initial idle state after maintenance is completed.

[0159] Specifically, when the fault occurs, its remaining time Replaced by repair time (10-20 minutes) During maintenance, the machine tool will not be included in the candidate set, and its existence will be ignored during path planning. After maintenance is completed ( ), reset it to idle state ( If the faulty machine tool is in the current candidate path, the RGV guide car will replan the path and adjust the scheduling mode (such as switching between 1-2-1 and 2-1-2 modes) to adapt to the change in the proportion of machine tools when it stops and waits again.

[0160] Step six: Control the RGV guide vehicle to perform movement, loading and unloading, and cleaning operations along the selected path. After completion, return to step three and repeat until the system time threshold is reached.

[0161] Specifically, the RGV guided vehicle moves sequentially to each machine tool position according to the optimal path and performs the following operations: In a single operation, it first unloads (if finished material is available) and loads (raw material) onto the current machine tool, then cleans the finished material and sends it to the unloading conveyor belt; in a double operation, it unloads (semi-finished material) and loads (raw material) onto the first operation machine tool, then moves to the second operation machine tool to unload (finished material) and load (semi-finished material), cleans the finished material, and sends it to the unloading conveyor belt. After completing the operation of all machine tools in the path, the RGV guided vehicle enters a stop waiting state again, returns to step three to regenerate the candidate set and path, until the system time reaches 28800 seconds (shift ends). In single-operation scheduling, the machine tools are operated in a cyclical manner according to their numbers. In double-operation scheduling, the first round loads the first operation machine tool according to its number, and subsequent rounds prioritize scheduling the machine tool that completes the first operation the fastest, then schedule the nearest second operation machine tool, and repeat the operation.

[0162] Example 2,

[0163] Based on Example 1, the method also includes an efficiency evaluation step: namely, statistically analyzing the average idle time of machine tools within a shift as a system efficiency indicator; and verifying the robustness of the scheduling scheme based on a random fault simulation model.

[0164] For details, please refer to Figure 5-6 In efficiency evaluation, the average idle time of machine tools is calculated as follows: for each machine tool, record the remaining time for each task. The difference between the moment when the non-zero value changes to zero (task completion) and the moment when the value changes to non-zero (new task start) is the loading time + processing time + waiting time. Subtracting the loading time and processing time gives the single waiting time. The waiting time of all machine tools is accumulated and the average value is taken.

[0165] Based on the random failure simulation model, 100 independent experiments were conducted on single-process machining, with the machine tool and time of failure controlled by random numbers in each experiment. For dual-process machining, two machine tools were randomly selected to fail, and the same 100 experiments were conducted. The number of finished products processed in each experiment was statistically analyzed. The results showed that the number of finished products fluctuated within a small range and was not significantly different from the case without failure, verifying the high robustness of the scheduling scheme, that is, the failure of a single machine tool will not significantly affect the overall system efficiency.

[0166] More specifically, because the timing and location of machine tool failures are random, the calculated average idle time of the machine tools is also random. Therefore, we will not focus solely on the idle time of the machine tools here, but instead adopt a more reasonable method to measure the system's efficiency. To measure the system's efficiency, based on the three sets of data provided in the problem, we used random numbers to control whether a failure would occur and where it would occur for each loading operation during material processing in one process; during material processing in two processes, we randomly selected two machine tools to fail. Using this failure strategy, we conducted 100 shifts of randomized independent repeated experiments and calculated the number of finished products processed in each shift. The simulation results are as follows: Figure 5 and Figure 6 As shown, in the 100-shift experiment, the number of finished products processed fluctuated within a certain range without significant differences, indicating that the model has good stability. In other words, if one machine tool in the system fails, it will not cause significant interference to the entire processing system. Compared with the number of finished products processed in the case of no failure, it was found that the failure did not have a significant impact on the results, and the system still maintained high working efficiency.

[0167] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent scheduling of RGV guided vehicles, characterized in that, Includes the following steps: Step 1: Initialize the intelligent machining system parameters, including the RGV guide car position, CNC machine tool status, material process type, and system time threshold; The system parameters include RGV travel time parameters. CNC machine tool process time Loading and unloading time Cleaning time and the maximum working time of each shift; the CNC machine tool status is determined by the remaining task time model. Dynamic updates include processing status, idle status, and fault repair status; Step 2: Select the scheduling mode according to the material process type. For single-process material processing, a dynamic shortest path planning model is used. For dual-process material processing, the optimal quantity ratio and position distribution of machine tools for the two types of processes are first determined based on the single-parent genetic distribution optimization model, and then an improved dynamic path planning model is used. The dynamic shortest path planning model for materials in a single process includes: Based on the remaining task time model, select the k fastest CNC machine tools to complete the task to form a candidate machine tool set; The candidate machine tools are permuted to generate a set of candidate paths, and each path contains an ordered sequence of machine tool operations; Among them, the single-parent genetic distribution optimization model for materials in two processes includes: The process type allocation scheme for each CNC machine tool is represented by chromosome coding; Using the maximum finished product output within a shift as the fitness function, the machine tool distribution is iteratively optimized using rearrangement mutation operators, rotation operators, and probability swap operators. Output the optimal machine tool distribution scheme that satisfies the process time ratio constraint; The improved dynamic path planning model includes two scheduling modes: The 1-2-1 scheduling mode uses the first process machine tool as the start and end point of the path, and inserts the second process machine tool in the middle; In the 2-1-2 scheduling mode, the second process machine tool is used as the starting and ending point of the path, and the first process machine tool is inserted in the middle. The scheduling mode is automatically selected based on the ratio of the number of machine tools in the two types of processes. Specifically, during single-process machining, the dynamic shortest path planning model collects the remaining time of each CNC machine tool at time t. The candidate set is formed by selecting the k minimum machine tools, that is, the machine tools corresponding to the k minimum remaining task times. Generate all permutations of machine tools within the candidate set. Candidate path set Each path is an ordered sequence of machine tool operations, that is... ; In dual-process machining, the machine tool distribution is first determined using a single-parent genetic distribution optimization model: In chromosome coding, if the gene value belongs to... The corresponding machine tool is the first process, in which This represents the number of machine tools used in the first process; otherwise, it represents the number used in the second process. Fitness is the maximum number of units produced within a given shift. During iteration, 20% of elite individuals are retained for the next generation. The distribution is optimized through rearrangement operators (randomly rearranging the coding order), rotation operators (circularly cutting the coding to generate new solutions), and probabilistic swap operators (randomly swapping two codes with a 50% probability of further swapping). The final output satisfies these optimizations. The optimal distribution, where This is the ratio of the time taken for the first process to the time taken for the second process. Step 3: When the RGV guide vehicle is in a stopped and waiting state, collect the remaining task time of each CNC machine tool in real time, generate a candidate machine tool set and construct a candidate path set; Step 4: Evaluate the total time of each candidate path based on the path time calculation model, and select the path with the shortest time as the scheduling scheme. Step 5: If a CNC machine tool malfunction is detected, the malfunctioning machine tool is removed from the candidate set and will be reinstated into the scheduling sequence after repairs are completed. Step six: Control the RGV guide vehicle to perform movement, loading and unloading, and cleaning operations along the selected path. After completion, return to step three and repeat until the system time threshold is reached.

2. The scheduling method according to claim 1, characterized in that, In step three: The candidate machine tool set is generated using a selection model based on the top k minimum values, where k is an integer greater than 1. The candidate path set contains all permutations and combinations of candidate machine tools.

3. The scheduling method according to claim 1, characterized in that, The fault handling in step five includes: The remaining task time for the faulty machine tool was replaced with a maintenance time model. ; During maintenance, the machine tool will not participate in path planning, and will be reset to its initial idle state after maintenance is completed.

4. The scheduling method according to claim 1, characterized in that, The method also includes an efficiency evaluation step: The average idle time of machine tools within a shift is used as a system efficiency indicator. The robustness of the scheduling scheme was verified based on a random fault simulation model.

5. The scheduling method according to claim 2, characterized in that, The value of k is in the range of 3 ≤ k ≤ 5.

6. The scheduling method according to claim 1, characterized in that, The operation of the single-parent genetic distribution optimization model includes: The elite retention mechanism replicates the most fit individuals to the next generation; The chromosome coding order is randomly adjusted by rearrangement mutation operators; A new solution is generated by circular shifting the rotation operator; The operation type codes of any two machine tools can be swapped using a probability-based switching operator.

Images