Construction method of production scheduling scheme of heterogeneous workpiece system and production scheduling method

Abstract

The present application belongs to the field of workpiece processing production scheduling, and particularly relates to a construction method of a heterogeneous workpiece system production scheduling scheme, a production scheduling method, a computer system and a computer readable storage medium. The construction method comprises: 1) selecting a workpiece to be processed currently, determining a current processing procedure of the workpiece according to a serial logical order of the processing procedure of the workpiece, and determining a decision time of the processing procedure according to a time line of the processing procedure; 2) inputting a state space corresponding to the decision time, which comprises the current processing procedure, an occupation state of each processing device and an occupation state of each processing resource, into a decision model corresponding to the workpiece to obtain a decision result for determining a processing device used for the processing procedure, and updating a time line of the processing procedure and a locking time period of the processing device and the processing resource; repeating 1)-2) until all workpieces are processed to complete the construction of the production scheduling scheme.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of workpiece processing production scheduling, and particularly relates to a heterogeneous workpiece system production scheduling scheme construction method, a production scheduling method, a computer system and a computer readable storage medium. BACKGROUND

[0002] Production scheduling is essentially a scheduling problem under complex space-time constraints, and essentially requires providing a feasible and highly complex combinatorial optimization scheme. Researchers widely use optimization algorithms and intelligent optimization algorithms for solving. However, if a quick and accurate decision is to be made in a high dynamic and strong real-time scenario, the traditional off-line static scheduling scheme lacks environmental adaptability and real-time feedback mechanism, and it is difficult to meet the demand of quick response, and is often limited by real-time requirements due to computational complexity.

[0003] In the prior art, a Markov model is often constructed according to time steps, and a state space corresponding to the time step is set on this basis, and a decision model of the workpiece is trained accordingly; when constructing a production scheduling scheme, each time step of each workpiece in its corresponding time line determines whether the current processing is completed, if completed, the state space corresponding to the current time step is input into the decision model of the workpiece, and the processing equipment for executing the workpiece processing in the next step is determined, so as to realize real-time scheduling; but this way is completely serial between each time step, that is, it is assumed that a processing equipment can only process a single process of a workpiece in each time, and the case that a processing equipment can simultaneously process two or more processes of a workpiece cannot be considered, so that the constructed production scheduling scheme cannot fully utilize the processing capacity of the processing equipment, resulting in that the total length of the constructed production scheduling scheme is longer and the overall efficiency is lower. SUMMARY

[0004] The present application aims to provide a heterogeneous workpiece system production scheduling scheme construction method, a production scheduling method, a computer system and a computer readable storage medium, which are used to solve the problem that the method of constructing a production scheduling scheme in the prior art cannot consider the case that a processing equipment can simultaneously process two or more processes of a workpiece, so that the constructed production scheduling scheme cannot fully utilize the processing capacity of the processing equipment, resulting in that the total length of the constructed production scheduling scheme is longer and the overall efficiency is lower.

[0005] In order to achieve the above-mentioned purpose, the present application provides a heterogeneous workpiece system production scheduling scheme construction method, comprising:

[0006] 1) selecting a workpiece to be processed, determining the current processing process of the workpiece according to the serial logic sequence of the processing process of the workpiece, and determining the decision time of the process according to the time line of the process;

[0007] 2) inputting the state space corresponding to the decision moment, which contains the current machining process, the occupation state of each machining device and the occupation state of each machining resource, into the decision model trained for the heterogeneous workpiece category to which the workpiece belongs, to obtain a decision result for determining the machining device used for the machining process, and update the timeline of the machining process and the locking time period of the machining device and the machining resource; repeat 1)-2) until all workpieces are machined to complete the construction of the production scheduling scheme;

[0008] The determination manner of the state space corresponding to the decision moment comprises: obtaining the occupation state of each machining device and the occupation state of each machining resource in the state space corresponding to the decision moment of the machining process by using the decision moment, the timeline of the machining process and the locking time period of the machining device and the machining resource;

[0009] The determination manner of the serial logical order of the machining process of the workpiece comprises: setting the logical order of the preceding machining process before the subsequent machining process; and setting the serial logical order between the machining processes without the preceding and subsequent relationship according to the set order;

[0010] The setting manner of the timeline of the machining process comprises: setting a corresponding parallel timeline for each machining process set of a single workpiece which are parallel in time sequence; setting a corresponding serial timeline for each machining process set of the workpiece which are serial in time sequence; each machining process set contains at least one machining process, and different machining processes in the set are serial in time sequence;

[0011] The updating manner of the timeline of the machining process comprises: if the decision result of executing the machining process does not need to change the machining device of the workpiece, directly adding the execution time of the machining process to the corresponding timeline; otherwise, setting the end time of the timeline with the longest total time length in the timelines of all machining processes of the workpiece to be machined as a reference, unifying the end time of the timelines of all machining processes of the workpiece, and then adding the execution time of the machining process to the corresponding timeline.

[0012] Further, the manner of selecting the workpiece to be machined is to select the workpiece with the highest current priority from the workpieces to be machined;

[0013] The determination manner of the current priority of the workpiece comprises: for the workpiece whose machining processes are not all completed, the priority of the workpiece with a longer total time length required for executing the remaining machining processes of the workpiece is higher; if the total time lengths are equal, the priorities of the workpieces with the equal total time lengths are set according to a preset workpiece number order.

[0014] Further, the current priority of the workpiece is updated according to the situation after the execution of the decision result for determining the machining device used for the machining process each time the decision result is obtained.

[0015] Further, the setting sequence is determined according to the time consumption of the machining process and a preset process number; the shorter the time consumption of the machining process, the earlier the setting sequence, and if the time consumption of the machining process is the same, the sequence of the machining processes with the same time consumption is set according to the preset process number.

[0016] Further, the determination manner of the execution time of the process comprises:

[0017] a) intersecting the idle time period of the machining equipment used by the process determined according to the decision result and the non-locking time period of the machining resource required for executing the process, and determining the machining start time corresponding to the decision time of the process; the idle time period comprises the non-locking time period of the machining equipment and the locking time period of the machining equipment satisfying a set condition; the set condition is that the workpiece corresponding to the machining of the locking time period is the same workpiece as the workpiece corresponding to the process;

[0018] b) determining the machining end time of the process in combination with the machining start time of the process and the time required for executing the process itself, and checking whether the machining time period from the machining start time to the machining end time of the process is contained in the intersection, if contained, determining the execution time of the process according to the machining time period; otherwise, delaying the machining start time of the process, and repeating b) until the execution time of the process is determined.

[0019] Further, the state space further comprises the execution state of each machining process of the workpiece to be machined, the distance of the workpiece to be machined relative to each machining equipment, and the machining equipment where the workpiece to be machined is located;

[0020] In the training process of the decision model, the reward function corresponding to the output decision result is used to update the parameters of the decision model; the reward function comprises a first function item corresponding to the completion of the machining process corresponding to the decision result, a second function item corresponding to the waiting time consumption of the machining process corresponding to the decision result, a third function item corresponding to the transfer time consumption of the machining process corresponding to the decision result, and a fourth function item corresponding to the load balancing of the machining process corresponding to the decision result.

[0021] The value of the first function item in the case that the machining process corresponding to the decision result is successfully completed is greater than the value of the first function item in the case that the machining process corresponding to the decision result is not successfully completed; the shorter the waiting time consumption of the machining process corresponding to the decision result, the greater the value of the second function item; the shorter the transfer time consumption of the machining process corresponding to the decision result, the greater the value of the third function item; the smaller the difference between the load rate of each machining resource corresponding to the machining process corresponding to the decision result and the load rate of all machining resources, the greater the value of the fourth function item.

[0022] Further, each decision model is trained by a DQN algorithm.

[0023] The technical solution of the present application provides a new construction method of a heterogeneous workpiece system production scheduling scheme, which has the following beneficial effects:

[0024] By determining whether there is a predecessor-successor relationship and a set ordering mode, each processing procedure of a single workpiece that may exist in parallel in terms of time is uniformly converted into an overall serial logic that conforms to the actual processing order. Thus, the time sequence that needs to be executed in series when constructing the production scheduling scheme can be modified into a logic sequence that needs to be executed in series. In combination with each decision, the state space corresponding to the decision time of the current processing procedure of the current workpiece is input into the decision model corresponding to the workpiece to achieve the effect of no longer using time steps as the standard for dividing the front and back state spaces of the Markov model, but rather using the execution of each decision step corresponding to each procedure to divide the front and back state spaces. At the same time, a parallel timeline is set for the set of procedures that are in parallel in terms of time. When the execution of the workpiece procedures does not change the processing equipment, the procedure execution time is directly added to the timeline corresponding to the procedure. Thus, for the processing procedures of a workpiece that do not change the processing equipment and are adjacent in the serial logic sequence but are in parallel in terms of time, their respective times are only added to the respective corresponding parallel timelines, rather than being added to the serial timeline, thereby achieving the effect of simultaneously processing different procedures of the workpiece on the same processing equipment. In addition, based on the execution of each decision step corresponding to each procedure to divide the front and back state spaces, since the procedures of different types of workpieces in the heterogeneous workpiece system are different, the decision model corresponding to each type of workpiece needs to be obtained through a respective training process to ensure that the decision model is adapted to the procedure of the workpiece itself.

[0025] The present application also provides a heterogeneous workpiece system production scheduling method. In the construction process of the heterogeneous workpiece system production scheduling scheme, the actual production scheduling of the heterogeneous workpiece system is controlled through the information obtained from the construction process. The construction of the heterogeneous workpiece system production scheduling scheme is achieved according to the construction method of the heterogeneous workpiece system production scheduling scheme described above.

[0026] The heterogeneous workpiece system production scheduling method described above can achieve the same beneficial effects as the construction method of the heterogeneous workpiece system production scheduling scheme described above.

[0027] The present application also provides a computer system. The processor is used to execute executable program instructions for being executed to implement the heterogeneous workpiece system production scheduling method described above.

[0028] The computer system described above can achieve the same beneficial effects as the heterogeneous workpiece system production scheduling method described above.

[0029] The application further provides a computer readable storage medium, wherein computer program instructions are stored in the storage medium, and the computer program instructions are used to realize the heterogeneous workpiece system production scheduling method.

[0030] The computer readable storage medium of the application has the same beneficial effects as the heterogeneous workpiece system production scheduling method. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 A principle block diagram of the construction method of the heterogeneous workpiece system production scheduling scheme in the construction method embodiment of the heterogeneous workpiece system production scheduling scheme of the application;

[0032] Figure 2 A training principle schematic diagram of the decision model trained by the DQN algorithm in the construction method embodiment of the heterogeneous workpiece system production scheduling scheme of the application;

[0033] Figure 3 A Gantt chart of the scheduling decision made by the heterogeneous workpiece system production scheduling scheme corresponding to the two kinds of heterogeneous workpieces constructed according to the construction method in the embodiment under the guidance of the decision model in the construction method embodiment of the heterogeneous workpiece system production scheduling scheme of the application;

[0034] Figure 4 A timing relationship schematic diagram of the work processes required by the workpiece category A in the construction method embodiment of the heterogeneous workpiece system production scheduling scheme of the application;

[0035] Figure 5 A timing relationship schematic diagram of the work processes required by the workpiece category B in the construction method embodiment of the heterogeneous workpiece system production scheduling scheme of the application;

[0036] Figure 6 An example diagram of the total processing time of the heterogeneous workpiece system production scheduling scheme constructed by the construction method of the embodiment and the genetic algorithm of the classical combination optimization algorithm under the condition of different workpiece quantities in the construction method embodiment of the heterogeneous workpiece system production scheduling scheme of the application. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical scheme and advantages of the application more clear, the application is further described in detail below with reference to the drawings and embodiments.

[0038] Construction method embodiment of the heterogeneous workpiece system production scheduling scheme

[0039] The embodiment gives a technical scheme of a construction method of a heterogeneous workpiece system production scheduling scheme, which uses the decision steps of processes instead of time steps to realize the construction of the production scheduling scheme, and sets parallel timelines to represent processes that can be executed in parallel in logic, so as to fully consider the case that more than two machining processes are executed in parallel on the same machine.

[0040] The scheme refers to Figure 1 , and includes

[0041] 1) selecting a workpiece to be machined at present, determining the machining process of the workpiece at present according to the serial logical order of the machining process of the workpiece, and determining the decision time of the process according to the timeline of the machining process;

[0042] 2) inputting the state space corresponding to the decision time, which contains the current machining process, the occupation state of each machining device and the occupation state of each machining resource, into the decision model trained for the heterogeneous workpiece category to which the workpiece belongs, to obtain the decision result for determining the machining device used for the machining process, and updating the timeline of the machining process and the locking time period of the machining device and the machining resource; repeating 1)-2) until all workpieces are machined to complete the construction of the production scheduling scheme;

[0043] The determination method of the state space corresponding to the decision time includes: using the decision time, the timeline of the machining process and the locking time period of the machining device and the machining resource to obtain the occupation state of each machining device and the occupation state of each machining resource in the state space corresponding to the decision time of the process; the updating of the locking time period of the machining device and the machining resource can directly add the time period in which the machining device and the machining resource are occupied by the process for machining as new locking time periods of the machining device and the machining resource, and the workpieces corresponding to the machining of these locking time periods are the workpieces corresponding to the process; in other embodiments, other locking time period updating methods that can reflect the occupation of the machining device and the machining resource by the machining process can also be used.

[0044] The determination method of the serial logical order of the workpiece machining process includes: setting the logical order of the preceding machining process before the subsequent machining process; setting the serial logical order between the machining processes without the predecessor-successor relationship according to the set order;

[0045] The setting method of the timeline of the process (i.e. the timeline allocation rule in Figure 1 ) includes: setting corresponding parallel timelines for each process set of a single workpiece that are parallel in time sequence; setting corresponding serial timelines for each process set of the workpiece that are serial in time sequence; each process set contains at least one machining process, and different machining processes in the set are serial in time sequence;

[0046] The way of updating the timeline of the processing procedure includes: if the decision result of executing the procedure does not need to change the processing equipment where the workpiece is located (for example, according to the serial logical order of the processing procedures of a workpiece, the previous processing procedure X of the current processing procedure Y of the workpiece is processed by the processing equipment A, and the decision result of executing the current processing procedure Y is also processed by the processing equipment A, which indicates that the decision result of executing the procedure Y does not need to change the processing equipment where the workpiece is located), the execution time of the procedure is directly added to the corresponding timeline of the procedure; otherwise, the end time of the timeline with the longest total time in the timelines of all the procedures of the workpiece to be processed is taken as the reference, and the end times of the timelines of all the procedures of the workpiece are unified, and then the execution time of the procedure is added to the corresponding timeline of the procedure.

[0047] In summary, the construction method of the heterogeneous workpiece system production scheduling scheme in the embodiment unifies the processing procedures of a single workpiece that may exist in parallel in time sequence into an overall serial logic that conforms to the actual processing order according to whether there is a predecessor-successor relationship and the set order, realizes the effect of dividing the front and rear state spaces by the execution of each decision step of the procedure by combining the setting of the decision model using the state space input corresponding to the decision time of the current processing procedure, sets the parallel timelines and the corresponding updating methods for the parallel procedure set in time sequence, realizes the effect of simultaneously processing different procedures of the workpiece on the same processing equipment, and respectively obtains the decision model corresponding to each kind of workpiece through the training process of each kind of workpiece to ensure that the decision model is adapted to the procedure conditions of the heterogeneous workpiece itself.

[0048] Specifically, the set order in the serial logical order is determined according to the processing time of the processing procedure and the preset procedure number; the shorter the processing time of the processing procedure, the earlier the set order, and if the processing times of the processing procedures are the same, the order of the processing procedures with the same processing time is set according to the preset procedure number order. That is, the serial logical order is set according to the predecessor procedure priority rule and the short procedure priority rule in the Figure 1

[0049] For example, the serial logical order is set according to the predecessor procedure priority rule and the short procedure priority rule in the Figure 1 ​As shown, in the machining topological sequence of a certain workpiece (specifically, the topological sequence in time), there are two groups of machining processes in parallel in time, which can be divided into two sets of machining processes in parallel in time, one of which has 2 machining processes in series in time, and the other has 3 machining processes in series in time; In the machining topological sequence, there are also two machining processes in series in time and having a series in time relationship with the two groups of machining processes in parallel in time, which can be divided into two sets of machining processes in series in time or into the same set of machining processes; The machining processes in series in time are divided into the same set of machining processes, which can reduce the number of time lines set and the number of variables that need to be concerned in the construction process of the heterogeneous workpiece system production scheduling scheme, so the two machining processes are divided into the same set of machining processes in this embodiment; However, in principle, since the serial time lines are set for the sets of machining processes in series in time, the execution order of the processes corresponding to each serial time line in time is consistent with the order relationship between the serial time lines itself when real-time scheduling and statistical total processing time, and the length of each serial time line is also directly superimposed, Therefore, whether the machining processes in series in time are divided into two sets of machining processes in series in time and set serial time lines respectively, or divided into the same set of machining processes and share the same time line, it does not affect the final presentation effect of the time line in real-time scheduling and statistical total processing time. According to the time line distribution rule, the corresponding parallel time lines are set for the above two sets of machining processes in parallel in time, that is, the time line 1 and the time line 2 in the formula (1) are set for the other set of machining processes. Time line 3; Combined with the serial logic order of the machining processes in the three sets of machining processes set according to the predecessor process priority rule and the short process priority rule, that is, the equivalent serial serial logic in the formula (2) is obtained. Figure 1 Figure 1

[0050] ​​In the equivalent serial serial logic, each processing procedure has a corresponding time line, and the parallel time line 1 and the time line 2 are selected when the total time is counted. The time length of the longest time line in the time line 1 and the time line 2 is added to the total time, so that the parallel processing operation in time sequence is realized in the total time counting; in the scheduling process, if the decision result of executing the second procedure in the time line 1 does not need to change the machining equipment where the workpiece is located, for example, according to the serial logic sequence of the machining procedures of the workpiece, the next procedure of the first procedure in the time line 1 is exactly the second procedure in the time line 1, and the decision result of the second procedure in the time line 1 corresponds to the same machining equipment as the decision result of the first procedure in the time line 1, then the execution time of the second procedure in the time line 1 is directly added to the time line 1; if the decision result of executing the second procedure in the time line 2 does not need to change the machining equipment where the workpiece is located, for example, according to the serial logic sequence of the machining procedures of the workpiece, the next procedure of the first procedure in the time line 2 is exactly the second procedure in the time line 2, and the decision result of the second procedure in the time line 2 corresponds to the same machining equipment as the decision result of the first procedure in the time line 2, then the execution time of the second procedure in the time line 2 is directly added to the time line 2; since the time line 1 and the time line 2 have the same starting time and are in parallel relationship, such time line updating method can ensure that the first procedure and the second procedure in the time line 1 and the first procedure and the second procedure in the time line 2 are executed in parallel in time sequence; and on this basis, if the decision result of executing the third procedure in the time line 2 needs to change the machining equipment where the workpiece is located, for example, according to the serial logic sequence of the machining procedures of the workpiece, the next procedure of the second procedure in the time line 1 is exactly the third procedure in the time line 1, and the decision result of the third procedure in the time line 2 corresponds to different machining equipment from the decision result of the second procedure in the time line 2, then the end time of the time line 1 and the time line 2 (i.e. the end time of the time line of all procedures of the workpiece currently to be machined) with the longest total time is taken as the reference, the end times of the time line 1 and the time line 2 are unified, and then the execution time of the third procedure in the time line 2 is added to the time line 2; for example, the first procedure and the second procedure in the time line 1 consume 1 min, the first procedure and the second procedure in the time line 2 consume 40 s, the end time of the time line 1 is taken as the reference to unify the end times of the time line 1 and the time line 2, which means that the time consumed after the end of the first procedure and the second procedure in the time line 2 is regarded as 1 min, and then the execution time of the third procedure in the time line 2 is added to the unified time line 2, so that the final total time of the time line 2 is the time length corresponding to the total time of 1 min plus the execution time of the third procedure.In actual scheduling, if the preparation time before the processing equipment performs the processing procedure is not considered, the scheduling according to the timeline 1 and the timeline 2 is that the processing equipment simultaneously performs the procedures in the timeline 1 and the timeline 2, after 40s, the first procedure and the second procedure in the timeline 2 are processed, at this time, since the third procedure in the timeline 2 needs to be transported to another processing equipment to perform processing, the end time of the timeline 1 (in which the first procedure and the second procedure are processed) is taken as a reference, the end time of the timeline 1 and the end time of the timeline 2 (in which the first procedure and the second procedure are processed) are unified, which is equivalent to controlling the third procedure in the timeline 2 to wait until the first procedure and the second procedure in the timeline 1 are both processed in the current processing equipment, and then start to be transported to another processing equipment to perform processing, which conforms to the normal logic of actual scheduling. Moreover, the determination manner of the execution time of the procedure includes:

[0051] a) intersecting the idle time period of the processing equipment used by the procedure determined according to the decision result and the non-locking time period of the processing resource required for processing the procedure, determining the processing start time corresponding to the decision time of the procedure, the idle time period contains both the non-locking time period of the processing equipment and the locking time period of the processing equipment that satisfies the set condition; the set condition is that the workpiece corresponding to the processing of the locking time period is the same workpiece as the workpiece corresponding to the procedure;

[0052] b) determining the processing end time of the procedure in combination with the processing start time of the procedure and the time required for processing the procedure itself, checking whether the processing time period from the processing start time of the procedure to the processing end time of the procedure is contained in the intersection, if contained, determining the execution time of the procedure according to the processing time period; otherwise, the processing start time of the procedure is delayed, and b) is repeated until the execution time of the procedure is determined.

[0053] Therefore, it can be seen that the execution duration of the process needs to include the duration of waiting for the machining equipment or machining resource to end the occupation before the process execution and the duration required for the execution of the process itself; wherein, the step a) first determines the time period in which both the machining equipment and the machining resource used by the process are idle by taking the intersection of the idle time period of the machining equipment and the non-locking time period of the machining resource of the process; it should be noted that, considering the ability of the same machining equipment to have different processes of machining the same workpiece, the idle time period of the machining equipment of the process not only includes the time period in which the machining equipment does not perform any machining task (i.e. the non-locking time period of the machining equipment), but also includes the time period in which the machining equipment is occupied by the workpiece corresponding to the current process (i.e. the locking time period of the machining equipment that meets the set condition in the locking time period of the machining equipment); for example, the current process is numbered as 1, which is referred to as process 1; the workpiece corresponding to the process 1 is numbered as B, which is referred to as workpiece B; the machining equipment determined to process the process 1 according to the decision result is m, and the idle time period of the machining equipment used by the process 1 includes both the non-locking time period of the machining equipment m and the locking time period of the machining equipment m occupied by the workpiece B. Thus, it can be ensured that the determined idle time period is compatible with the case of simultaneously machining different processes of the same workpiece by the same equipment.

[0054] The step b) represents that whether the machining start time corresponding to the workpiece decision time (i.e. the earliest machining start time in the possible machining start time of the process) is valid is judged from the machining start time; if the machining start time can ensure that the process is completed in the time period in which both the equipment and the resource are idle, it is indicated that the machining start time is valid; if the machining start time is invalid, the machining start time is continuously pushed back, and whether the new machining start time is valid is continuously judged, until the valid machining start time is found, which can be used to more reasonably determine the execution duration of the process.

[0055] Specifically, in combination with the above content, the pseudo code of the determination of the serial logical sequence and the updating mode of the time line of the machining process in the embodiment is shown in Table 1 as follows:

[0056] Table 1

[0057]

[0058] The state space of the input decision model above is the state space of the Markov process model corresponding to the workpiece machining; the decision result output by the decision model for determining the machining equipment used by the machining process is the action space of the Markov process model; the Markov process model is usually represented as a five-tuple ; wherein is the state space, which is represented as , the state space before the n-th decision step is executed (each decision step gets a decision result), and the state space after the n-th decision step is executed is , that is, the state space before the n+1-th decision step is executed; is the action space, which is usually represented as , that is, the decision result obtained by the n-th decision step; is the state transition probability, which describes the probability of the agent transitioning from the current state space to another state space after performing a certain action. Since the change in the state space after executing a process in this embodiment is unique, the state transition probability is set to 1; is the reward function, which in this embodiment can also be represented as is the immediate feedback provided by the workpiece after taking a certain action in each state. That is, after executing action in state and transitioning to state , the reward obtained is the reward function, which in this embodiment corresponds to the output decision result. During the process of training the decision model, the parameters of the decision model are updated each time the iteration is updated; is the discount factor, which is a constant during the iteration process of training the decision model. During the process of training the decision model, the parameters of the decision model are updated each time the iteration is updated. It is usually set to a real number between 0 and 1 according to the specific situation, which is used to determine the importance of future rewards relative to immediate rewards; when is close to 1, more emphasis is placed on future rewards; when is close to 0, more attention is paid to immediate rewards; in this embodiment, the discount factor is determined by comparing the experimental results of the average reward function curve obtained by training with different discount factors under the same environment and workpiece set. In other embodiments, the discount factor can also be set according to actual requirements or empirical values.

[0059] Specifically, since the production scheduling of a heterogeneous workpiece system is usually performed in time waves, a wave needs to go through the process of machining-transportation-re-machining, and the Markov process model of this embodiment focuses on the optimization of the machining process of all workpieces in a single wave and the optimal allocation of machining equipment and its machining resources. The Markov process model has the following constraints:

[0060] (1) A workpiece can only choose one machine in the same process stage;

[0061] (2) The total number of processes allocated to the machine for execution in the entire process is equal to the total number of processes that all workpieces need to execute in the entire process;

[0062] (3) The completion time of a workpiece in the corresponding process is equal to the sum of its start time and the processing time of the process, and a process cannot be interrupted after it has started processing;

[0063] (4) Each machine can only provide services for one workpiece at a time.

[0064] In this embodiment, in addition to the current processing step, the occupancy status of each processing device, and the occupancy status of each processing resource mentioned above, the state space also includes the execution status of each processing step of the workpiece to be processed, the distance of the workpiece to be processed relative to each processing device, and the processing device where the workpiece to be processed is located; specifically expressed by the following formula:

[0065]

[0066] In the formula, Let n be the state space corresponding to the workpiece that needs to be processed before the nth decision step is executed; n is the decision step number. This represents the occupancy status of processing equipment m in step n. This indicates that processing equipment m is not occupied in step n, otherwise... ; This represents the occupancy status of processing resource k in step n, and its value has the same meaning as that of processing equipment. Let i be the execution status of the i-th operation of workpiece j in step n. This indicates that the i-th operation of workpiece j in step n is executed, otherwise... ; Let be the distance of workpiece j relative to processing equipment m in step n; Let j be the location of workpiece j in step n. For ease of representation, let be... This indicates that workpiece j is located at processing equipment m; The number of the processing steps currently required for workpiece j, for example This indicates that the current workpiece j is about to undergo the i-th processing step.

[0067] Action space The specific formula is as follows:

[0068]

[0069] In the formula, m represents the serial number of the processing equipment, and M is the maximum serial number of the processing equipment; specifically, The situation (i.e.) The situation indicates that the current processing step of the workpiece continues on the original processing equipment (i.e., the processing equipment where the workpiece is currently located, which is also the processing equipment of the previous processing step of the workpiece).

[0070] The decision model corresponding to each different kind of workpiece in the heterogeneous workpiece is obtained through a respective training process, that is, the decision model corresponding to the kind of workpiece to which the workpiece belongs is obtained through training; in the training process of the decision model, the reward function corresponding to the output decision result is used to update the parameters of the decision model; the reward function includes a first function item corresponding to the completion of the machining process corresponding to the current decision result, a second function item corresponding to the waiting time of the machining process corresponding to the current decision result, a third function item corresponding to the transfer time of the machining process corresponding to the current decision result, and a fourth function item corresponding to the load balancing of the machining process corresponding to the current decision result;

[0071] The value of the first function item in the case where the machining process corresponding to the current decision result is successfully completed is greater than the value of the first function item in the case where the machining process corresponding to the current decision result is not successfully completed; the shorter the waiting time of the machining process corresponding to the current decision result, the greater the value of the second function item; the shorter the transfer time of the machining process corresponding to the current decision result, the greater the value of the third function item; the smaller the difference between the load rate of each machining resource corresponding to the machining process corresponding to the current decision result and the load rate of all machining resources, the greater the value of the fourth function item. Specifically, the determination mode of the reward function is expressed by the following formula:

[0072]

[0073] In the formula, is a hyperparameter for controlling the weight of each type of reward, and ; is the first function item corresponding to the completion of the machining process corresponding to the current decision result; is the second function item corresponding to the waiting time of the machining process corresponding to the current decision result; is the third function item corresponding to the transfer time of the machining process corresponding to the current decision result; is the fourth function item corresponding to the load balancing of the machining process corresponding to the current decision result.

[0074] In this embodiment, the first function item corresponding to the completion of the machining process in each iteration is artificially judged according to the output decision result of each iteration, and if the decision result of this decision can make the process successfully machined, the value is a first set constant, and the first set constant is greater than 0; if the machining of the process is not successful, the value is a second set constant less than the first set constant, and the second set constant is usually 0. 、 The waiting time and the transfer time of the machining process corresponding to the decision result of this decision are determined respectively; The calculation can be performed according to the following formula:

[0075]

[0076] In the formula, represents the number of all processing resources, represents the load rate of the kth processing resource; represents the average of the load rates of all processing resources.

[0077] Specifically, each decision model is trained by a DQN algorithm, and the training principle is referred to Figure 2 . Since the way of training the decision model of the output Markov process model based on the state space of the input Markov process model by the DQN algorithm belongs to the prior art, it is usually realized by an existing DQN algorithm encapsulation module, and thus will not be described here.

[0078] Considering that the deadlock phenomenon may occur when a workpiece waits for a processing device for a long time, causing the processing device to be idle for a long time or even permanently, thereby significantly prolonging the completion time of the entire processing process, it is crucial to ensure the balance of the processing process and avoid potential deadlock phenomena in the optimization solution of workpiece processing. Therefore, in the embodiment, to prevent the balance of the processing process and avoid potential deadlock phenomena, a deadlock prevention mechanism based on processing priority is designed, which is specifically embodied in the way of selecting the workpiece to be processed currently, that is, selecting the workpiece with the highest current priority from the workpieces to be processed. The determination method of the current priority of the workpiece includes: for the workpiece whose processing procedure is not completely executed, the total time length required for the execution of the remaining processing procedure of the workpiece is determined, and the longer the total time length, the higher the priority of the workpiece; if the total time lengths are equal, the priorities of the workpieces with equal total time lengths are set according to a preset workpiece number order.

[0079] Therefore, the above-mentioned selection method of selecting the workpiece with the highest current priority from the workpieces to be processed and the determination method of the current priority of the workpiece are equivalent to dividing the processing priorities of the workpieces in the batch according to the actual production plan and the processing progress of the workpieces, and setting the priority of the workpiece with a relatively longer total time length required for the execution of the remaining processing procedure to be higher, so as to ensure that the workpiece with a processing progress behind is processed preferentially and will not continue to be idle for a long time to cause a deadlock phenomenon, and the processing progress of each workpiece is balanced. In the embodiment, the way of setting the priorities of the workpieces with equal total time lengths according to the preset workpiece number order is taken as an example of ascending arrangement according to the preset workpiece number order, and the pseudo code of the process of selecting the workpiece to be processed currently is shown in Table 2.

[0080] Table 2

[0081]

[0082] In other embodiments, other methods can be used to set the priority of these workpieces with equal total time according to a preset workpiece number order; if there is no need to consider the balance of processing progress and the effect of preventing deadlock between workpieces, other methods of determining priority can also be used, or even other methods of selecting the workpiece to be processed at present can be used, which are not limited in this invention.

[0083] In addition, to ensure that the workpiece selected for processing based on priority can adapt to the changes in the state space after the decision result is executed, the current priority of the workpiece is updated according to the situation after the decision result is executed each time the decision result for determining the processing equipment used in the processing step is obtained.

[0084] In summary, the pseudocode for the overall process of constructing the production scheduling plan can be represented by the following Table 3:

[0085] Table 3

[0086]

[0087] Figure 3 The diagram shows the Gantt chart of the production scheduling scheme of the heterogeneous workpiece system constructed according to the construction method of this embodiment, under the guidance of the decision model, where the horizontal axis represents time and the vertical axis represents different processes. Figure 3 The description describes the actual support process for two types of workpieces with a total of four heterogeneous workpieces (0, 1, 2, and 3 represent four heterogeneous workpieces, of which 0, 1, and 2 represent the same type of heterogeneous workpiece, referred to as workpiece type A; 3 represents another type of heterogeneous workpiece, referred to as workpiece type B). The decision model allocates actual processing equipment to the workpieces at each decision moment, instructs the decision behavior of the workpieces, and thus completes the entire process production scheduling plan to minimize the total processing time.

[0088] The timing relationship diagrams for the required processes of the two heterogeneous workpieces (i.e., workpiece type A and workpiece type B) set in this embodiment are shown below. Figure 4 and Figure 5 As shown, Figure 4 and Figure 5 Dark-colored process boxes indicate processes requiring processing resources, while light-colored process boxes indicate processes not requiring processing resources, i.e., the required processing resources are empty. To visually demonstrate the superiority of the heterogeneous workpiece system production scheduling scheme construction method in this embodiment, this construction method is compared with the classic combinatorial optimization algorithm and genetic algorithm. The comparison of the two methods under the same environment and resource settings, resulting in different heterogeneous workpiece system production scheduling schemes, is evaluated, highlighting their different construction speeds and final construction quality. The final results are shown in Table 4 below. Figure 6 As shown:

[0089] Table 4

[0090] Workpiece quantity Construction method of the present embodiment Genetic algorithm 1 0.3620s 92.2158s 2 0.3980s 148.3864s 3 0.4869s 196.6379s 4 0.5694s 208.4142s 5 0.7539s 300.6371s

[0091] Table 4 shows the time required for the construction method of the embodiment to construct a heterogeneous job system production scheduling scheme and the genetic algorithm of the classical combinatorial optimization algorithm to construct a heterogeneous job system production scheduling scheme under different job quantities; Figure 6 Table 4 shows the time required for the construction method of the embodiment to construct a heterogeneous job system production scheduling scheme and the genetic algorithm of the classical combinatorial optimization algorithm to construct a heterogeneous job system production scheduling scheme under different job quantities; Figure 6 Table 4 shows the time required for the construction method of the embodiment to construct a heterogeneous job system production scheduling scheme and the genetic algorithm of the classical combinatorial optimization algorithm to construct a heterogeneous job system production scheduling scheme under different job quantities; Figure 6 Table 4 shows the time required for the construction method of the embodiment to construct a heterogeneous job system production scheduling scheme and the genetic algorithm of the classical combinatorial optimization algorithm to construct a heterogeneous job system production scheduling scheme under different job quantities;

[0092] Heterogeneous job system production scheduling method embodiment

[0093] The embodiment provides a heterogeneous job system production scheduling method. In the construction process of a heterogeneous job system production scheduling scheme, the actual production scheduling of a heterogeneous job system is controlled through information obtained from the construction process. The construction of the heterogeneous job system production scheduling scheme is achieved according to the construction method of the heterogeneous job system production scheduling scheme in the above-mentioned heterogeneous job system production scheduling scheme construction method embodiment.

[0094] Since the specific principles and beneficial effects of the heterogeneous job system production scheduling method in the embodiment have been described in detail in the above-mentioned heterogeneous job system production scheduling scheme construction method embodiment, they will not be described here again.

[0095] Computer system embodiment

[0096] The embodiment provides a technical scheme of a computer system. The computer system includes a processor, and the processor has executable program instructions stored therein. The executable program instructions are used to be executed to implement the heterogeneous job system production scheduling method in the above-mentioned heterogeneous job system production scheduling method embodiment.

[0097] Since the specific working principles and effects of the computer system in the embodiment have been described in detail in the above-mentioned heterogeneous job system production scheduling method embodiment, they will not be described here again.

[0098] Computer readable storage medium embodiment

[0099] The embodiment provides a technical scheme of a computer readable storage medium, and the storage medium stores computer program instructions, which are used to implement the heterogeneous workpiece system production scheduling method in the above-mentioned embodiment of the heterogeneous workpiece system production scheduling method when being executed.

[0100] Since the specific working principle and effects of the computer readable storage medium in the embodiment have been described in detail in the above-mentioned embodiment of the heterogeneous workpiece system production scheduling method, the description will not be repeated here.

[0101] It should be understood that the above specific embodiments of the present application are only used for illustrative or explanatory purposes of the principles of the present application, and do not constitute a limitation on the present application.

Claims

1. A method for constructing a production scheduling scheme for a heterogeneous workpiece system, characterized in that, include: 1) Select the workpiece to be processed, determine the current processing step of the workpiece according to the serial logic sequence of the processing steps, and determine the decision time of the process according to the timeline of the processing steps; 2) Input the state space corresponding to the decision moment, which includes the current processing procedure, the occupancy status of each processing equipment, and the occupancy status of each processing resource, into the decision model trained according to the heterogeneous workpiece type to which the workpiece belongs. Obtain the decision result used to determine the processing equipment used for the processing procedure, and update the timeline of the processing procedure and the locking time period of the processing equipment and processing resources. Repeat 1)-2) until all workpieces are processed to complete the construction of the production scheduling plan. The methods for determining the state space corresponding to the decision moment include: using the decision moment, the timeline of the processing procedure, and the locking time period of the processing equipment and processing resources, to obtain the occupancy status of each processing equipment and the occupancy status of each processing resource in the state space corresponding to the decision moment of the procedure. The methods for determining the sequential logical order of workpiece processing steps include: setting the logical order of the preceding processing steps before the subsequent processing steps; and setting the sequential logical order between processing steps that do not have a preceding-successor relationship according to a set order. The timeline settings for the processes include: setting corresponding parallel timelines for each set of processes that are parallel in time for a single workpiece; setting corresponding serial timelines for each set of processes that are serial in time for the remaining processes of the workpiece; each set of processes contains at least one processing process and the different processing processes in the set are serial in time; The methods for updating the timeline of a processing step include: if the decision to execute the step does not require changing the processing equipment where the workpiece is located, then the execution time of the step is directly superimposed on its corresponding timeline; otherwise, the end time of the timeline with the longest total duration among all the timelines of the workpiece to be processed is used as the benchmark, and after unifying the end times of all the timelines of the workpiece, the execution time of the step is superimposed on its corresponding timeline.

2. The method for constructing a production scheduling scheme for a heterogeneous workpiece system according to claim 1, characterized in that, The method for selecting the workpiece to be processed is to select the workpiece with the highest priority from the workpieces to be processed. The current priority of a workpiece is determined by the following method: for workpieces whose processing steps have not been fully completed, the priority is determined by the total time required for the remaining processing steps of the workpiece itself. The longer the total time, the higher the priority of the workpiece. If the total durations are equal, then the priority of these workpieces with equal total durations is set according to the preset workpiece number order.

3. The method for constructing a production scheduling scheme for a heterogeneous workpiece system according to claim 2, characterized in that, Each time a decision is made regarding the processing equipment used to determine the processing step, the current priority of the workpiece is updated based on the outcome of that decision.

4. The method for constructing a production scheduling scheme for a heterogeneous workpiece system according to claim 1 or 2, characterized in that, The setting order is determined based on the processing time of the processing steps and the preset process number; the shorter the processing time, the earlier the setting order; if the processing times are the same, the order of these processing steps with the same time is set according to the preset process number order.

5. The method for constructing a production scheduling scheme for a heterogeneous workpiece system according to claim 1 or 2, characterized in that, The methods for determining the execution time of a process include: a) Take the intersection of the idle time period of the processing equipment used in the process determined according to the decision result and the non-locked time period of the processing resources required to perform the process; determine the processing start time corresponding to the decision time of the process; the idle time period includes both the non-locked time period of the processing equipment and the locked time period of the processing equipment that meets the set conditions; the set conditions are that the workpiece processed corresponding to the locked time period is the same workpiece as the workpiece corresponding to the process. b) Determine the end time of the process by combining the start time of the process and the time required for the process to be executed; check whether the processing time period from the start time of the process to the end time of the process is included in the intersection. If it is included, determine the execution time of the process based on the processing time period; otherwise, postpone the start time of the process and repeat b) until the execution time of the process is determined.

6. The method for constructing a production scheduling scheme for a heterogeneous workpiece system according to claim 1 or 2, characterized in that, The state space also includes the execution status of each processing step of the workpiece to be processed, the distance of the workpiece to be processed relative to each processing device, and the processing device where the workpiece to be processed is located. During the training of the decision model, the parameters of the decision model are updated using the reward function corresponding to the output decision result. The reward function includes a first function term corresponding to the completion status of the processing procedure corresponding to the current decision result, a second function term corresponding to the waiting time of the processing procedure corresponding to the current decision result, a third function term corresponding to the transfer time of the processing procedure corresponding to the current decision result, and a fourth function term corresponding to the load balancing status of the processing procedure corresponding to the current decision result. If the processing step corresponding to this decision result is successfully completed, the value of the first function term is greater than the value of the first function term if the processing step corresponding to this decision result is not successfully completed; the shorter the waiting time for the processing step corresponding to this decision result, the larger the value of the second function term; the shorter the transfer time for the processing step corresponding to this decision result, the larger the value of the third function term; the smaller the difference between the load rate of each processing resource corresponding to the processing step corresponding to this decision result and the load rate of all processing resources, the larger the value of the fourth function term.

7. The method for constructing a production scheduling scheme for a heterogeneous workpiece system according to claim 1 or 2, characterized in that, All decision models were trained using the DQN algorithm.

8. A production scheduling method for a heterogeneous workpiece system, characterized in that, In the process of constructing a production scheduling scheme for a heterogeneous workpiece system, the actual production scheduling of the heterogeneous workpiece system is controlled by information obtained from the construction process; the construction of the production scheduling scheme for the heterogeneous workpiece system is implemented according to the construction method of the production scheduling scheme for the heterogeneous workpiece system as described in any one of claims 1-7.

9. A computer system comprising a processor, the processor being configured to execute executable program instructions, characterized in that, The executable program instructions are executed to implement the heterogeneous workpiece system production scheduling method according to claim 8.

10. A computer-readable storage medium storing computer program instructions, characterized in that, The computer program instructions are used to implement the heterogeneous workpiece system production scheduling method as described in claim 8 when executed.

Images