Combined equipment scheduling method based on Petri network and hybrid heuristic search

Abstract

The invention discloses a combined equipment scheduling method based on a Petri network and hybrid heuristic search, which solves the problem of long program running time when a complex environment is processed in the prior art, and comprises the following steps of: establishing a time assignment Petri network model and determining an initial identifier; initializing an OPEN table and a CLOSED table, and placing the initial identifier in the OPEN table; obtaining a plurality of enabling transitions under the initial identifier according to the number of tokens in each library of the timing Petri network model, generating a current identifier under the current transition based on the search box according to the plurality of enabling transitions and an OPEN table, obtaining a current reachable graph corresponding to the search box according to the current identifier and the current transition, and obtaining the current identifier according to the current reachable graph; circularly executing the steps until the current identifier is the same as the target identifier, and outputting a transition sequence; a better transition sequence is given in a shorter time, so that the production efficiency is improved, and shared resources are reasonably arranged.

Description

Technical Field

[0001] The present invention relates to the technical field of manufacturing system modeling and simulation, and particularly to a combined equipment scheduling method based on Petri nets and hybrid heuristic search. Background Art

[0002] With the intensification of market competition, enterprises have an increasingly strong demand for maximizing production capacity and improving the utilization rate of equipment and product quality as much as possible. Therefore, planning and scheduling play an increasingly important role in modern manufacturing systems. Although the traditional large-batch and single-variety processing method improves the productivity of wafers, it is difficult to ensure processing consistency and affects production quality. Therefore, combined equipment with single-wafer processing technology is widely used in the semiconductor industry, which can shorten the manufacturing cycle, reduce costs, improve equipment utilization rate and product quality.

[0003] Combined equipment is the core equipment for semiconductor wafer manufacturing, and its scheduling and control optimization are extremely challenging topics in the field of semiconductor manufacturing. In a typical discrete event system such as a semiconductor production line, unreasonable resource allocation may lead to deadlocks, which will not only reduce productivity but also may cause the entire system to stop. Designing and using an efficient scheduling method is of great significance for improving the production efficiency of manufacturing systems, reducing costs and meeting market demands.

[0004] Petri nets are a graphical modeling tool used to describe discrete event systems, which can intuitively reflect concurrency, deadlocks, resource usage, etc. in flexible manufacturing systems, and can be used to model and analyze system scheduling problems. With the development of Petri nets, their advantages in system modeling, design and analysis have become increasingly obvious, and they have become one of the mainstream methods for analyzing and solving problems such as discrete event system modeling, control and scheduling.

[0005] In a reachability graph of a relatively large-scale Petri net, as the search progresses, the number of nodes to be searched grows exponentially, making it increasingly difficult for the search process to move one step closer to the target node. To solve this problem, a solution that combines the simulation ability of Petri nets and the A-Star heuristic search method is adopted to optimize the wafer manufacturing system. At the same time, a moving window method is added to the algorithm to limit the range of node selection for the search algorithm, and a hybrid heuristic search algorithm is proposed, which effectively reduces the search range of nodes for the algorithm when finding paths in the reachability graph, enabling the algorithm to converge quickly for relatively large-scale systems.

[0006] Combining Petri nets and hybrid heuristic search can form a new type of manufacturing system scheduling method. This method uses Petri nets to describe the dynamic behavior and resource flow of manufacturing systems, and makes scheduling decisions through hybrid heuristic strategies, thereby achieving effective scheduling and optimization of manufacturing systems with the optimization goal of minimizing the processing completion time.

[0007] In the research on the scheduling problem of manufacturing systems, Dijkstra's algorithm is an algorithm used to solve the single-source shortest path problem in a weighted directed graph or an undirected graph. It gradually expands the shortest path from the current starting point to other vertices with a greedy strategy and is implemented by maintaining a distance array and a marking array. When Petri nets and Dijkstra's algorithm are combined to solve the scheduling problem of flexible manufacturing systems, from the perspective of the reachability graph, nodes can be understood as identifiers. Only by enumerating all transition firing sequences in the reachability graph can an optimal solution be found. For systems with a large scale, there will be a problem of too long program running time.

[0008] Since the reachability graph of Petri nets grows exponentially with the system scale, the existing A-Star algorithm for the combined equipment scheduling problem needs to store a large amount of node information, resulting in large memory consumption and high computational complexity. When dealing with complex environments, it is difficult to run out the results within a tolerable time for people. Summary of the Invention

[0009] The present invention provides a combined equipment scheduling method based on Petri nets and hybrid heuristic search, which solves the problem of long program running time in the prior art when dealing with complex environments, and realizes giving a relatively optimal transition sequence in a short time, thereby improving production efficiency and reasonably arranging shared resources.

[0010] In a first aspect, the present invention provides a combined equipment scheduling method based on Petri nets and hybrid heuristic search, and the method includes:

[0011] S1, according to the processing route of the wafer, establish a Petri net model, and combine the actual processing time of each processing link in the wafer to add corresponding time delays to the places and transitions in the Petri net model to obtain a timed Petri net model, and determine the initial marking of the timed Petri net model;

[0012] S2, initialize the OPEN table and the CLOSED table, and place the initial marking in the OPEN table;

[0013] S3, according to the number of tokens in each place of the timed Petri net model, obtain multiple enabled transitions under the initial marking, and then based on the search box, generate the current marking under the current transition according to the multiple enabled transitions and the OPEN table, and obtain the current reachability graph corresponding to the search box according to the current marking and the current transition, and obtain the current marking according to the current reachability graph;

[0014] S4, update the OPEN table and the CLOSED table according to the current reachability graph;

[0015] S5. According to the downward movement condition of the search box, loop through S3 to S4 until the current identifier is the same as the target identifier, output the transition sequence from the initial identifier to the target identifier, and use the transition sequence as the optimal scheduling sequence.

[0016] In combination with the first aspect, in a possible implementation, based on the search box, generate the current identifier under the current transition according to multiple enabling transitions and the OPEN table, and obtain the current reachable graph corresponding to the search box according to the current identifier and the current transition, including:

[0017] Determine the position of the search box, and obtain the current reachable graph corresponding to the search box according to the reachable graph generation method within the search box; wherein, the reachable graph generation method includes:

[0018] Calculate the costs of each identifier in the OPEN table according to the cost calculation formula, and sort them in ascending order according to the costs of each identifier.

[0019] Judge in sequence whether there are enabling transitions for each identifier in the OPEN table.

[0020] If so, the enabling transition under this identifier fires, obtain the current identifier under the current transition, and obtain the updated current reachable graph within the search box according to the current identifier and the current transition.

[0021] If not, this identifier is a deadlock identifier, and retain this identifier.

[0022] According to the screening condition of the search box, screen the nodes in the updated current reachable graph within the search box to obtain the current reachable graph within the search box, and update the OPEN table and the CLOSED table according to the current reachable graph.

[0023] In combination with the first aspect, in a possible implementation, the "If so, the enabling transition under this identifier fires, obtain the current identifier under the current transition, and obtain the updated current reachable graph within the search box" includes:

[0024] Loop through all the first identifiers in the OPEN table, and execute the first loop body for each of the first identifiers until the loop ends to obtain the updated current reachable graph within the search box.

[0025] Wherein, the first loop body includes:

[0026] Calculate the number of enabling transitions corresponding to the first identifier, and fire the transitions in sequence until the enabling transitions corresponding to the first identifier are all fired, to obtain the updated current reachable graph within the search box.

[0027] In combination with the first aspect, in a possible implementation manner, filtering the nodes in the updated current reachable graph in the search box according to the filtering condition of the search box to obtain the current reachable graph in the search box, and updating the OPEN table and the CLOSED table according to the current reachable graph, including:

[0028] Determine a plurality of unexpanded second identifiers in the updated current reachable graph that are located in a layer corresponding to an upper boundary of the window in the search box;

[0029] Based on the cost calculation formula, respectively calculate the costs corresponding to the plurality of second identifiers;

[0030] Sort the costs corresponding to the multiple second identifiers in ascending order, and retain the first m nodes in the layer; wherein the number of nodes retained in the layer is the same as the node limit number; m is an integer greater than 0;

[0031] The updated current reachable graph is updated according to the nodes retained in the layer to obtain the current reachable graph in the search box;

[0032] The OPEN table and the CLOSED table are updated according to the current reachable graph.

[0033] In combination with the first aspect, in a possible implementation manner, the cost calculation formula is expressed as:

[0034] f(M) = CT(M) - dep(M);

[0035] Among them, f(M) represents the initial identifier M0 to the target identifier M e The cost; CT(M) represents the system clock from the initial identifier M0 to the current identifier M; -dep(M) represents the system clock from the current identifier M to the target identifier M e The estimated cost of .

[0036] In combination with the first aspect, in a possible implementation manner, the condition for moving the search box downward includes:

[0037] All nodes in the lower bound of the window in the search box are expanded; the lower bound of the window in the search box is in the opposite direction of the reachable graph generation direction; or

[0038] The number of nodes in each layer in the search box is greater than or equal to the node limit.

[0039] In a second aspect, the present invention provides a combined device scheduling device based on Petri net and hybrid heuristic search, the device comprising:

[0040] A modeling module, configured to establish a Petri net model according to the processing route of a wafer, and add corresponding time delays to places and transitions in the Petri net model in combination with the actual processing time of each processing link in the wafer, so as to obtain a timed Petri net model;

[0041] An optimization module, configured to obtain a current reachability graph by using a path optimization method based on a search box;

[0042] A scheduling module, configured to perform path optimization on the current reachability graph, output a transition sequence from an initial marking to a target marking, and use the transition sequence as an optimal scheduling sequence.

[0043] In a third aspect, the present invention provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, a combined device scheduling method based on a Petri net and a hybrid heuristic search is implemented.

[0044] In a fourth aspect, the present invention provides an electronic device, including a processor and a memory, where the memory stores a computer program, and when the processor executes the computer program, a combined device scheduling method based on a Petri net and a hybrid heuristic search is implemented.

[0045] One or more technical solutions provided in the present invention have at least the following technical effects or advantages:

[0046] The present invention models the production process of a manufacturing system through a Petri net, and at the same time restricts the scale of the reachability graph by introducing a search box, so as to be able to perform path optimization in the generated current reachability graph, avoiding the exponential growth of the number of nodes to be searched as the search progresses; finding a transition sequence from an initial marking to a target marking in a short time, and scheduling the combined device wafer processing according to the transition sequence, realizing the reasonable arrangement of shared resources and the efficient, stable and reliable operation of the manufacturing system. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a flowchart of the steps of the combined device scheduling method based on a Petri net and a hybrid heuristic search provided by an embodiment of the present invention;

[0048] Figure 2 It is a schematic diagram of a search box for node selection in a reachability graph provided by an embodiment of the present invention;

[0049] Figure 3 It is a schematic diagram of the downward movement of the search box provided by an embodiment of the present invention;

[0050] Figure 4 It is a combined device scheduling device based on a Petri net and a hybrid heuristic search provided by an embodiment of the present invention. Specific Embodiments

[0051] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0052] In a first aspect, the present invention provides a combined device scheduling method based on Petri nets and hybrid heuristic search. Refer to Figure 1 , and this method includes the following steps S1 to S5.

[0053] S1. According to the processing route of the wafer, establish a Petri net model. Combining the actual processing time of each processing link in the wafer, add corresponding time delays to the places and transitions in the Petri net model to obtain a timed Petri net model, and determine the initial marking of the timed Petri net model.

[0054] Exemplarily, first, according to the processing procedures of the processing route of the wafer and the manufacturing system resources, select a process-oriented modeling method to perform Petri net modeling on the manufacturing system, including the initial marking M0, the corresponding number of places placeNum and the number of transitions transNum during calculation, and consider the time during actual processing production to add time delays to the places and transitions to obtain a timed Petri net model, which constitutes the model basis for subsequent algorithm calculations.

[0055] S2. Initialize the OPEN table and the CLOSED table, and place the initial marking M0 in the OPEN table.

[0056] Exemplarily, the markings saved in the OPEN table are expandable markings. The OPEN table is a collection of default ascending priority queues, and the markings saved in the CLOSED table are expanded markings or non-expandable markings.

[0057] S3. According to the number of tokens in each place of the timed Petri net model, obtain multiple enabled transitions under the initial marking. Then, based on the search box, generate the current marking under the current transition according to the multiple enabled transitions and the OPEN table, and obtain the current reachable graph corresponding to the search box according to the current marking and the current transition, and obtain the current marking according to the current reachable graph.

[0058] Specifically, in step S3, based on the search box, generating the current marking under the current transition according to the multiple enabled transitions and the OPEN table, and obtaining the current reachable graph corresponding to the search box according to the current marking and the current transition includes the following steps S31 to S34.

[0059] S31. Determine the position of the search box, and generate the current reachable graph corresponding to the search box according to the reachable graph generation method within the search box. The reachable graph generation method includes:

[0060] S32. Calculate the cost of each identifier in the OPEN list according to the cost calculation formula, and sort them in ascending order according to the cost of each identifier.

[0061] Here, the cost calculation formula is expressed as:

[0062] f(M) = CT(M) - dep(M);

[0063] where f(M) represents the cost from the initial identifier M0 to the target identifier M e ; CT(M) represents the system clock from the initial identifier M0 to the current identifier M; -dep(M) represents the estimated cost from the current identifier M to the target identifier M e .

[0064] S33. Sequentially determine whether there is an enabled transition for each identifier in the OPEN list.

[0065] If so, the enabled transition under this identifier fires, obtaining the current identifier under the current transition, and obtaining the updated current reachable graph within the search box according to the current identifier and the current transition.

[0066] If not, this identifier is a deadlock identifier, and this identifier is retained.

[0067] Here, if so, the enabled transition under this identifier fires, obtaining the current identifier under the current transition, and obtaining the updated current reachable graph within the search box according to the current identifier and the current transition, including: looping through all the first identifiers in the OPEN list, and executing the first loop body for each first identifier until the loop ends to obtain the updated current reachable graph within the search box.

[0068] Among them, the first loop body includes:

[0069] Calculate the number of enabled transitions corresponding to the first identifier, and sequentially fire the transitions until the firing of the enabled transitions corresponding to the first identifier is completed, obtaining the updated current reachable graph within the search box.

[0070] S34. According to the screening conditions of the search box, screen the nodes in the updated current reachable graph within the search box to obtain the current reachable graph within the search box, and update the OPEN list and the CLOSED list according to the current reachable graph.

[0071] Here, according to the filtering conditions of the search box, the nodes in the updated current reachable graph are filtered in the search box to obtain the current reachable graph in the search box, and the OPEN table and the CLOSED table are updated according to the current reachable graph, including:

[0072] (1) determining a plurality of unexpanded second identifiers in the updated current reachable graph that are located in a layer corresponding to the upper boundary of the window in the search box;

[0073] (2) calculating the costs corresponding to the plurality of second identifiers respectively based on the cost calculation formula;

[0074] (3) sorting the costs corresponding to the multiple second identifiers in ascending order, and retaining the first m nodes in the layer, where m is an integer greater than 0; wherein the number of nodes retained in the layer is the same as the node limit number;

[0075] (4) updating the updated current reachable graph according to the nodes retained in the layer to obtain the current reachable graph in the search box;

[0076] (5) Update the OPEN table and CLOSED table according to the current reachable graph.

[0077] Here, the conditions for moving the search box downward include: all nodes in the lower bound of the window in the search box are expanded; or the number of nodes in each layer of the search box is greater than or equal to the node limit. Among them, the lower bound of the window in the search box is the opposite direction of the reachable graph generation direction; see Figure 2 , the upper limit of the window in the search box is the upper boundary of the search box.

[0078] Exemplarily, in the timed Petri net model, the initial identifier M0 is represented as:

[0079] M0=[0,13,0,0,1,0,0,1,0,0,1,1,1,1,1,0,0,1,1,0,1,0 1,0] T ;

[0080] The number of tokens in place p2 is the number of processed wafers.

[0081] Target ID M e It is expressed as:

[0082] M e =[13,0,0,0,1,0,0,1,0,0,1,1,1,1,1,0,0,1,1,0,1,0 1,0] T ;

[0083] Under this flag, the number of tokens in place p1 is 13, indicating that all wafers have been processed.

[0084] The A-Star algorithm can reduce the search space and gradually find the optimal path. Its core expression is:

[0085] f(M) = g(M) + h(M);

[0086] Among them, g(M) represents the actual cost from the initial identifier M0 to the current identifier M, and g(M) = CT(M) represents the system clock from the initial identifier M0 to the current identifier M. In the example of the present invention, the system clock from the initial identifier M0 to the current identifier M is used as the actual cost; h(M) represents the estimated cost from the current identifier M to the target identifier M e The estimated cost of, f(M) represents an approximation of the cost of the entire path, and the entire expression is called a heuristic function; h(M) = -dep(M), where dep(M) represents the depth of the current identifier M in the reachable graph, and -dep(M) represents the estimated cost from the current identifier M to the target identifier M e The estimated cost of.

[0087] In the combined device scheduling problem, the approximation f(M) of the cost of the path from the initial identifier M0 to the target identifier M e is denoted as the completion time.

[0088] Specifically, in the reachable graph, the depth of each current identifier is defined as the number of transition firings from the initial identifier to the current identifier. It can be obtained that f(M) = CT(M) - dep(M).

[0089] The reachable graph generation method proposed by the present invention controls the movement of the search box by adjusting the upper bound of the window of the search box and the lower bound of the window of the search box, and at the same time limits the number of nodes in the search box.

[0090] Specifically, the search box has two parameters: the lower bound of the window and the upper bound of the window, which respectively represent the minimum and maximum depths of the nodes included in the reachable graph.

[0091] The search box only accommodates those nodes whose depths are between the lower bound and the upper bound of the window. The number of nodes contained in the layer with depth l in the search box is size(l_depth), and l_depth represents that it cannot exceed the node limit number maxSize at most. When all the nodes at the bottom layer are expanded, the search box must move forward.

[0092] See Figure 2 which is a schematic diagram of the search box for node selection in the reachable graph provided by the embodiment of the present invention. As Figure 2 shown, the window of the search box in the reachable graph is calibrated by the upper bound and the lower bound of the window. The search box downward movement method does not consider those nodes that are too shallow or too deep in the Petri net reachable graph when selecting nodes for expansion. Only the generated nodes whose depths in the reachable graph fall between the upper bound and the lower bound of the window are eligible to be selected for expansion to generate new child nodes.

[0093] The restrictions brought by the search box on the control of the node selection range by the search algorithm, and the window movement rules include:

[0094] (1) When all the underlying nodes are expanded, the window must move forward; or

[0095] (2) When the number of top-level nodes ≥ the node limit number, the window moves down; or

[0096] (3) To avoid exponential growth of nodes, the number of nodes in each layer is restricted.

[0097] It should be noted that the node limit number is dynamically adjusted according to the depth of nodes in each layer. A smaller node limit number is set in the shallow layer to accelerate the search, and the node limit number is appropriately increased in the deep layer to retain more candidate nodes; corresponding to the window movement rule (2).

[0098] Specifically, in each layer, the nodes are sorted in ascending order of the value of f(M). If the layer is not full, the new node is directly added to the layer. If the layer is full and the f value of the new node is less than the maximum value of f(M) in the layer, the new node can be added to the layer.

[0099] See Figure 3 , which is a schematic diagram of Rule 2 of the moving window method provided by the embodiment of the present invention. When the number of nodes at the upper bound of the window is greater than or equal to the pre-given maximum number of nodes, the identifier at the lower bound of the window is discarded. Figure 3 This situation is described, where the node limit number is 5, and at this time, the number of nodes at the top layer is also exactly 5, and the window moves down.

[0100] S4. Update the OPEN table and the CLOSED table according to the current reachable graph;

[0101] S5. According to the downward movement condition of the search box, loop to execute S3 to S4 until the current identifier is the same as the target identifier, output the transition sequence from the initial identifier to the target identifier, and use the transition sequence as the optimal scheduling sequence.

[0102] Exemplarily, use the Petri net model of wafer processing by a combined device to verify the feasibility of the transition sequence;

[0103] Solve the scheduling problem of the manufacturing system by using the A-Star algorithm and the hybrid heuristic search algorithm respectively. By comparing the main performance indicators such as the completion time, program running time, and the number of expanded identifiers generated from the initial identifier to the target identifier, it can be seen that the hybrid heuristic search reduces the memory consumption and greatly reduces the program running time.

[0104] The present invention models the production process of combined equipment wafer manufacturing based on Petri nets, and considering the time constraint conditions during actual processing and production, adds time delays to places and transitions to obtain a timed Petri net model, which can clearly express the technological sequence of wafer manufacturing in the combined equipment.

[0105] The present invention provides a scheduling method for combined equipment based on Petri nets and hybrid heuristic search. On the one hand, it can achieve searching on a partial reachability graph, improving the search efficiency and reducing the computational complexity; on the other hand, through the concept of the search box entrance, it restricts the scale of the search tree, controls the window movement by adjusting the upper and lower bounds of the window, and at the same time restricts the number of nodes in the window, which can avoid, as much as possible, the exponential growth of the number of nodes to be searched as the search progresses; and by modeling the production process of the manufacturing system using Petri nets and utilizing the hybrid heuristic search strategy, a processing transition sequence can be obtained, and a relatively optimal transition sequence can be given in a relatively short time, thereby improving production efficiency and reasonably arranging shared resources.

[0106] In a second aspect, the present invention provides a scheduling device for combined equipment based on Petri nets and hybrid heuristic search. Refer to Figure 4 , the device includes:

[0107] A modeling module, configured to establish a Petri net model according to the processing route of the wafer, and combine the actual processing time of each processing link in the wafer to add corresponding time delays to places and transitions in the Petri net model to obtain a timed Petri net model;

[0108] An optimization module, configured to obtain the current reachability graph by using a path optimization method based on a search box;

[0109] A scheduling module, configured to perform path optimization on the current reachability graph, output a transition sequence from the initial marking to the target marking, and use the transition sequence as the optimal scheduling sequence.

[0110] In a third aspect, the present invention provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, it implements a scheduling method for combined equipment based on Petri nets and hybrid heuristic search.

[0111] The above storage medium includes, but is not limited to, random access memory (English: Random Access Memory; abbreviation: RAM), read-only memory (English: Read-Only Memory; abbreviation: ROM), cache (English: Cache), hard disk (English: Hard Disk Drive; abbreviation: HDD), or memory card (English: Memory Card). The memory can be used to store computer program instructions.

[0112] In a fourth aspect, the present invention provides an electronic device, including a processor and a memory. The memory stores a computer program, and when the processor executes the computer program, a combined device scheduling method based on Petri nets and hybrid heuristic search is implemented.

[0113] Although the present invention provides method operation steps as described in the embodiments or flowcharts, more or fewer operation steps may be included based on routine or non-creative labor. The step sequence listed in this embodiment is only one way among many step execution sequences and does not represent the only execution sequence. When the actual device or client product executes, it may be executed in the order of the method shown in this embodiment or the drawings, or executed in parallel (for example, in an environment of parallel processors or multi-threaded processing).

[0114] The devices or modules illustrated in the above embodiments may be specifically implemented by computer chips or entities, or by products with certain functions. For the convenience of description, when describing the above devices, they are divided into various modules according to functions and described separately. When implementing the present invention, the functions of each module may be implemented in the same or multiple software and / or hardware. Of course, the module implementing a certain function may also be implemented by combining multiple sub-modules or sub-units.

[0115] Some modules in the device of the present invention may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, classes, etc. that perform specific tasks or implement specific abstract data types. The present invention may also be practiced in a distributed computing environment where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media including storage devices.

[0116] From the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus necessary hardware. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, may be embodied in the form of a software product, or may also be embodied in the implementation process of data migration. This computer software product may be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which may be a personal computer, a mobile terminal, a server, or a network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present invention.

[0117] The various embodiments in this specification are described in a progressive manner. For the same or similar parts among the various embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. All or part of the present invention can be used in many general-purpose or special-purpose computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, mobile communication terminals, multi-processor systems, microprocessor-based systems, programmable electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on.

[0118] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the present invention; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the present invention.

Claims

1. A combined equipment scheduling method based on Petri net and hybrid heuristic search, characterized in that: include: S1, establishing a Petri net model according to the processing route of the wafer, adding corresponding delays to the library and transition in the Petri net model in combination with the actual processing time of each processing link in the wafer, obtaining a timed Petri net model, and determining the initial identification of the timed Petri net model; S2, initializing the OPEN table and the CLOSED table, and placing the initial identifier in the OPEN table; S3, according to the number of tokens in each library of the timed Petri net model, multiple enabling transitions under the initial identifier are obtained, and then based on the search box, a current identifier under the current transition is generated according to the multiple enabling transitions and the OPEN table, and a current reachable graph corresponding to the search box is obtained according to the current identifier and the current transition, and a current identifier is obtained according to the current reachable graph; S4, updating the OPEN table and the CLOSED table according to the current reachable graph; S5, according to the downward movement condition of the search box, loop through S3 to S4 until the current identifier is the same as the target identifier, output a transition sequence from the initial identifier to the target identifier, and use the transition sequence as the optimal scheduling sequence.

2. The combined device scheduling method based on Petri net and hybrid heuristic search according to claim 1 is characterized in that: The step of generating a current identifier under a current transition based on the search box according to a plurality of enabled transitions and the OPEN table, and obtaining a current reachable graph corresponding to the search box according to the current identifier and the current transition includes: Determine the position of the search box, and obtain a current reachable graph corresponding to the search box according to a reachable graph generation method in the search box; wherein the reachable graph generation method includes: Calculate the cost of each identifier in the OPEN table according to the cost calculation formula, and sort them in ascending order according to the cost of each identifier; Determine in order whether each flag in the OPEN table has an enabling transition; If yes, the enabled transition under the identifier is emitted to obtain the current identifier under the current transition, and according to the current identifier and the current transition, the updated current reachable graph in the search box is obtained; If not, the flag is a deadlock flag and is retained; According to the filtering conditions of the search box, the nodes in the updated current reachable graph are filtered in the search box to obtain the current reachable graph in the search box, and the OPEN table and the CLOSED table are updated according to the current reachable graph.

3. The combined device scheduling method based on Petri net and hybrid heuristic search according to claim 2 is characterized in that: If so, the enabled transition under the identifier is emitted, the current identifier under the current transition is obtained, and the updated current reachable graph in the search box is obtained according to the current identifier and the current transition, including: Looping all the first identifiers in the OPEN table, and executing the first loop body for each of the first identifiers until the loop ends to obtain an updated current reachable graph in the search box; Wherein, the first loop body includes: The number of enabled transitions corresponding to the first identifier is calculated, and transitions are emitted in sequence until the emission of the enabled transitions corresponding to the first identifier is completed, thereby obtaining an updated current reachable graph in the search box.

4. The combined device scheduling method based on Petri net and hybrid heuristic search according to claim 2, characterized in that: The step of filtering the nodes in the updated current reachable graph in the search box according to the filtering condition of the search box to obtain the current reachable graph in the search box, and updating the OPEN table and the CLOSED table according to the current reachable graph includes: Determine a plurality of unexpanded second identifiers in the updated current reachable graph that are located in a layer corresponding to an upper boundary of the window in the search box; Based on the cost calculation formula, respectively calculate the costs corresponding to the plurality of second identifiers; Sort the costs corresponding to the multiple second identifiers in ascending order, and retain the first m nodes in the layer; wherein the number of nodes retained in the layer is the same as the node limit number; m is an integer greater than 0; update the updated current reachable graph according to the nodes retained in the layer to obtain the current reachable graph in the search box; The OPEN table and the CLOSED table are updated according to the current reachable graph.

5. The combined device scheduling method based on Petri net and hybrid heuristic search according to claim 2, characterized in that: The cost calculation formula is expressed as: f(M) = CT(M) - dep(M); Among them, f(M) represents the initial identifier M0 to the target identifier M e The cost; CT(M) represents the system clock from the initial identifier M0 to the current identifier M; -dep(M) represents the system clock from the current identifier M to the target identifier M e The estimated cost of .

6. The combined device scheduling method based on Petri net and hybrid heuristic search according to claim 1, characterized in that: The downward movement conditions of the search box include: All nodes in the lower bound of the window in the search box are expanded; the lower bound of the window in the search box is in the opposite direction of the reachable graph generation direction; or The number of nodes in each layer in the search box is greater than or equal to the node limit.

7. A combined equipment scheduling device based on Petri net and hybrid heuristic search, characterized in that: include: A modeling module is used to establish a Petri net model according to the processing route of the wafer, and add corresponding delays to the library and transition in the Petri net model in combination with the actual processing time of each processing link in the wafer to obtain a timed Petri net model; The optimization module is used to obtain the current reachable graph by adopting a path optimization method based on a search box; The scheduling module is used to perform path optimization in the current reachable graph, output a transition sequence from an initial identifier to a target identifier, and use the transition sequence as an optimal scheduling sequence.

8. A computer-readable storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a processor, the combined device scheduling method based on Petri net and hybrid heuristic search according to any one of claims 1 to 6 is implemented.

9. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores a computer program, and when the processor executes the computer program, the combined device scheduling method based on Petri net and hybrid heuristic search according to any one of claims 1 to 6 is implemented.

Images