Scheduling method and system for wafer automatic material transfer system

Abstract

This invention discloses a scheduling method and system for an automated wafer material handling system, comprising: acquiring data information of a target automated wafer material handling system; constructing a scheduling model of the target automated wafer material handling system based on its scheduling cost; solving the constructed scheduling model using the Grey Wolf algorithm, a competitive swarm optimization algorithm, and a random walk algorithm; and completing the scheduling of the target automated wafer material handling system based on the solution results. This invention constructs a scheduling model of the target automated wafer material handling system based on scheduling cost and uses a Grey Wolf algorithm (improved from the competitive swarm optimization algorithm and random walk algorithm) to solve the constructed scheduling model. Therefore, this invention not only enables the scheduling of automated wafer material handling systems, but the improved Grey Wolf algorithm also ensures high reliability, high accuracy, and high efficiency in the scheduling process.

Description

Technical Field

[0001] This invention belongs to the field of production scheduling, specifically relating to a scheduling method and system for an automated wafer material transfer system. Background Technology

[0002] In modern wafer fabrication companies, the Automated Material Handling System (AWDS) is a critical system. The ASDS intelligently plans wafer cassette loading and unloading tasks for transport equipment such as Overhead Hoist Transport (OHT), Overhead Shuttle (OHS), and Automated Guided Vehicle (AGV). It also significantly meets production requirements such as workshop cleanliness, equipment uptime, and material handling efficiency, while ensuring personnel safety, improving product yield, and optimizing production capacity.

[0003] Currently, the intelligent planning schemes used in automated wafer material handling systems are primarily heuristic scheduling schemes. Heuristic scheduling schemes can formulate and design certain handling rules based on experience to achieve task scheduling, and their operational logic and coding are relatively simple and easy to implement. However, heuristic scheduling schemes can only achieve local optimal scheduling within a certain range of the factory. With the expansion of wafer fabrication enterprises, the increasing complexity of automated wafer material handling systems, and the rapid growth in the number of scheduling tasks, heuristic scheduling schemes cannot achieve globally optimal scheduling for the system. Moreover, faced with massive scheduling optimization tasks, heuristic scheduling schemes may also generate huge energy and time costs. Summary of the Invention

[0004] One of the objectives of this invention is to provide a scheduling method for an automated wafer material handling system that is highly reliable, accurate, and efficient overall.

[0005] The second objective of this invention is to provide a system for implementing the scheduling method of the automated wafer material handling system.

[0006] The scheduling method for the automated wafer material handling system provided by this invention includes the following steps:

[0007] S1. Obtain data information from the target wafer automated material handling system;

[0008] S2. Based on the scheduling cost of the automated wafer material handling system, construct a scheduling model for the target automated wafer material handling system;

[0009] S3. Solve the scheduling model constructed in step S2 based on the gray wolf algorithm, competitive swarm optimization algorithm, and random walk algorithm;

[0010] Among them, the Grey Wolf Algorithm is used to solve the scheduling model; the Competitive Population Optimization Algorithm is used to enable individuals in the Grey Wolf Algorithm to compete and learn, so as to improve the accuracy of the Grey Wolf Algorithm; and the Random Walk Algorithm is used to alleviate the problem of the Grey Wolf Algorithm getting stuck in local optima, so as to improve the reliability of the Grey Wolf Algorithm.

[0011] S4. Based on the solution obtained in step S3, complete the scheduling of the target wafer automated material transfer system.

[0012] Step S1 specifically includes the following steps:

[0013] The acquired data information of the target wafer automated material handling system includes the location information of the conveying equipment, the total number of conveying equipment, the status information of the conveying equipment, the location information of the wafer cassette, the total number of wafer cassettes, the waiting time information of the wafer cassettes, and the delivery date information of the wafer cassettes.

[0014] Step S2 includes the following steps:

[0015] Based on the scheduling cost of the automated wafer material handling system, a scheduling model for the target automated wafer material handling system is constructed with the goal of minimizing the scheduling cost and the constraints of the transportation capacity of each transport device and the transportation status of each wafer cassette.

[0016] Step S2 specifically includes the following steps:

[0017] The following formula is used as the objective function of the scheduling model:

[0018]

[0019] In the formula, n is the total number of wafer boxes; m is the total number of transport devices; The cost of allocating the j-th transport device to the i-th wafer cassette, and , Let be the distance from the j-th transport device to the i-th wafer cassette. Let be the delivery date of the i-th wafer cell. Let be the waiting time for the i-th wafer cassette. This represents the start-destination buffer state of the i-th wafer cassette. As the first weight value, As the second weight value, As the third weight value, It is the fourth weight value; The state variable is whether the i-th wafer cassette is transported by the j-th transport device;

[0020] The following formula is used as the constraint condition for the scheduling model:

[0021]

[0022]

[0023] in This is used to limit each wafer cassette to be assigned to only one transport device; This is used to limit each transport device to transport only one wafer cassette at a time.

[0024] Step S3 includes the following steps:

[0025] The scheduling model constructed in step S2 is solved using the Grey Wolf algorithm;

[0026] In the solution process, a competitive swarm optimization algorithm is used to optimize the individuals in the gray wolf algorithm, so as to improve the competition and learning ability of each individual in the gray wolf algorithm, and at the same time improve the accuracy of the gray wolf algorithm.

[0027] During the solution process, the linear vectors in the Grey Wolf algorithm are optimized based on the sigmoid curve to improve the accuracy of the Grey Wolf algorithm;

[0028] In the solution process, a random walk algorithm based on Cauchy distribution is used to optimize the target position of the wolf pack in the next iteration, and the optimized linear vector is used as the multiplier of the random walk algorithm to alleviate the problem of the gray wolf algorithm getting stuck in local optima and improve the reliability of the gray wolf algorithm.

[0029] Step S3 specifically includes the following steps:

[0030] A. Initialize the gray wolf population and use it as the current gray wolf population, initializing the first coefficient vector. And use it as the current first coefficient vector to initialize the second coefficient vector. And serve as the current second coefficient vector;

[0031] The linear vector is initialized using the following formula. And serve as the current linear vector:

[0032]

[0033] In the formula As the first parameter, and t is the current iteration number; T is the set maximum iteration number; It is the second parameter, and ;

[0034] B. Use the objective function of the scheduling model as the fitness function, and calculate the fitness function of the current gray wolf population;

[0035] C. Based on the fitness function obtained in step B, define the gray wolf algorithm... Wolf, wolves and Wolf; simultaneously set the current iteration number t to 1, and set the maximum iteration number T;

[0036] D. If Then proceed with the following steps:

[0037] The wolf pack's position is updated using the following formula:

[0038]

[0039] In the formula The updated location of the wolf pack; For use in simulating gray wolves During the hunting process, the estimation of the prey's location and the corresponding movement are crucial. , For the gray wolf Current location For the gray wolf The dynamic coefficient, For the current gray wolves and gray wolves The distance, and , Let be the first random coefficient vector. For the current Grey Wolf; Used to simulate gray wolves During the hunting process, the estimation of the prey's location and the corresponding movement are crucial. , For the gray wolf Current location For the gray wolf The dynamic coefficient, For the current gray wolves and gray wolves The distance, and , This is the second random coefficient vector; For use in simulating gray wolves During the hunting process, the estimation of the prey's location and the corresponding movement are crucial. , For the gray wolf Current location For the gray wolf The dynamic coefficient, For the current gray wolves and gray wolves The distance, and , The third random coefficient vector; The current linear vector; This is a random hunting increment generated by the random walk algorithm;

[0040] like Then return the current This serves as the final solution for the final Grey Wolf algorithm.

[0041] E. Define the current wolf pack's set. for ; Set the intermediate set U as At the same time, set It is an empty set;

[0042] Repeat steps (1) to (6) until U is an empty set:

[0043] (1) Randomly select two individuals from U. and ;

[0044] (2) To and Make a judgment: If The fitness function value is not greater than If the fitness function value is [value], then the winning individual is set. for At the same time, set up failed individuals for ;like The fitness function value is not less than If the fitness function value is [value], then the winning individual is set. for At the same time, set up failed individuals for ;

[0045] (3) Select the winning individual join in ;

[0046] (4) Use the following formula to calculate Update:

[0047]

[0048]

[0049] In the formula For the updated ; Before the update ; This represents the update amount in round t+1; Let be the update amount in round t; This is the first random variable; It is the second random quantity; For the current ; To update the weight values; It is the third random quantity; For the current winning individual;

[0050] (5) Update the join in ;

[0051] (6) and Delete from U;

[0052] F. Setting for ;

[0053] G. For the first coefficient vector Second coefficient vector and the current linear vector Update, and update at the same time , and ;

[0054] H. Increment the value of the current iteration number t by 1 and return to step D.

[0055] This invention also provides a system for implementing the scheduling method of the automated wafer material handling system, comprising a data acquisition module, a model building module, a model solving module, and a system scheduling module; the data acquisition module, model building module, model solving module, and system scheduling module are connected in series; the data acquisition module is used to acquire data information of the target automated wafer material handling system and upload the data information to the model building module; the model building module is used to construct a scheduling model of the target automated wafer material handling system based on the received data information and the scheduling cost of the automated wafer material handling system, and upload the data information to the model solving module; the model solving module... Based on the received data, the system is used to solve the scheduling model constructed in step S2 using the Grey Wolf Algorithm, the Competitive Swarm Optimization Algorithm, and the Random Walk Algorithm, and then upload the data to the system scheduling module. The Grey Wolf Algorithm is used to solve the scheduling model; the Competitive Swarm Optimization Algorithm enables the individuals in the Grey Wolf Algorithm to compete and learn, thereby improving the accuracy of the Grey Wolf Algorithm; the Random Walk Algorithm is used to mitigate the problem of the Grey Wolf Algorithm getting trapped in local optima, thereby improving the reliability of the Grey Wolf Algorithm; and the system scheduling module is used to complete the scheduling of the target wafer automated material handling system based on the received data and the obtained solution results.

[0056] The scheduling method and system for automated wafer material handling provided by this invention constructs a scheduling model for the target automated wafer material handling system based on scheduling cost, and uses the Grey Wolf algorithm, which is an improvement of the competitive swarm optimization algorithm and the random walk algorithm, to solve the constructed scheduling model. Therefore, this invention can not only realize the scheduling of automated wafer material handling systems, but also ensure high reliability, high accuracy and high efficiency of the scheduling process with the improved Grey Wolf algorithm. Attached Figure Description

[0057] Figure 1 This is a schematic diagram of the method flow of the present invention.

[0058] Figure 2 This is a schematic diagram comparing the convergence curves of the method of the present invention and the comparative scheme during long-term operation; wherein, Figure 2 (a) is a schematic diagram comparing the convergence curves of the F1 function of the method of the present invention and the comparative scheme during long-term operation. Figure 2 (b) is a schematic diagram comparing the convergence curves of the F2 function of the method of the present invention and the comparative scheme during long-term operation. Figure 2 (c) is a schematic diagram comparing the convergence curves of the F3 function of the method of the present invention and the comparative scheme during long-term operation.

[0059] Figure 3 This is a schematic diagram comparing the convergence curves of the method of the present invention and the comparative scheme during long-term operation; wherein, Figure 3 (a) is a schematic diagram comparing the convergence curves of the F4 function of the method of the present invention and the comparative scheme during long-term operation. Figure 3 (b) is a schematic diagram comparing the convergence curves of the F5 function of the method of the present invention and the comparative scheme during long-term operation. Figure 3 (c) is a schematic diagram comparing the convergence curves of the F6 function of the method of the present invention and the comparative scheme during long-term operation.

[0060] Figure 4 This is a schematic diagram comparing the convergence curves of the method of the present invention and the comparative scheme during long-term operation; wherein, Figure 4 (a) is a schematic diagram comparing the convergence curves of the F7 function of the method of the present invention and the comparative scheme during long-term operation. Figure 4 (b) is a schematic diagram comparing the convergence curves of the F9 function of the method of the present invention and the comparative scheme during long-term operation. Figure 4 (c) is a schematic diagram comparing the convergence curves of the F10 function of the method of the present invention and the comparative scheme during long-term operation.

[0061] Figure 5 This is a schematic diagram of the functional modules of the system of the present invention. Detailed Implementation

[0062] like Figure 1The diagram shown illustrates the method flow of the present invention: a scheduling method for an automated wafer material handling system, comprising the following steps:

[0063] S1. Obtain data information from the target wafer automated material handling system; specifically including the following steps:

[0064] The acquired data information of the target wafer automated material handling system includes the location information of the conveying equipment, the total number of conveying equipment, the status information of the conveying equipment, the location information of the wafer cassette, the total number of wafer cassettes, the waiting time information of the wafer cassettes, and the delivery date information of the wafer cassettes.

[0065] S2. Based on the scheduling cost of the automated wafer material handling system, construct a scheduling model for the target automated wafer material handling system; including the following steps:

[0066] Based on the scheduling cost of the automated wafer material handling system, with the minimum scheduling cost as the objective function and the transportation capacity of each transportation device and the transportation status of each wafer box as constraints, a scheduling model for the target automated wafer material handling system is constructed.

[0067] In practical implementation, the following formula is used as the objective function of the scheduling model:

[0068]

[0069] In the formula, n is the total number of wafer boxes; m is the total number of transport devices; The cost of allocating the j-th transport device to the i-th wafer cassette, and , Let be the distance from the j-th transport device to the i-th wafer cassette. Let be the delivery date of the i-th wafer cell. Let be the waiting time for the i-th wafer cassette. This represents the start-destination buffer state of the i-th wafer cassette. As the first weight value, As the second weight value, As the third weight value, It is the fourth weight value; The state variable is whether the i-th wafer cassette is transported by the j-th transport device;

[0070] The following formula is used as the constraint condition for the scheduling model:

[0071]

[0072]

[0073] in This is used to limit each wafer cassette to be assigned to only one transport device; This is used to limit each transport device to transport only one wafer cassette at a time.

[0074] S3. Solve the scheduling model constructed in step S2 based on the gray wolf algorithm, competitive swarm optimization algorithm, and random walk algorithm;

[0075] Among them, the Grey Wolf Algorithm is used to solve the scheduling model; the Competitive Population Optimization Algorithm is used to enable individuals in the Grey Wolf Algorithm to compete and learn, so as to improve the accuracy of the Grey Wolf Algorithm; and the Random Walk Algorithm is used to alleviate the problem of the Grey Wolf Algorithm getting stuck in local optima, so as to improve the reliability of the Grey Wolf Algorithm.

[0076] Includes the following steps:

[0077] The scheduling model constructed in step S2 is solved using the Grey Wolf algorithm;

[0078] In the solution process, a competitive swarm optimization algorithm is used to optimize the individuals in the gray wolf algorithm, so as to improve the competition and learning ability of each individual in the gray wolf algorithm, and at the same time improve the accuracy of the gray wolf algorithm.

[0079] During the solution process, the linear vectors in the Grey Wolf algorithm are optimized based on the sigmoid curve to improve the accuracy of the Grey Wolf algorithm;

[0080] In the solution process, a random walk algorithm based on Cauchy distribution is used to optimize the target position of the wolf pack in the next iteration, and the optimized linear vector is used as the multiplier of the random walk algorithm to alleviate the problem of the gray wolf algorithm getting stuck in local optima and improve the reliability of the gray wolf algorithm.

[0081] The Gray Wolf Algorithm is a swarm intelligence algorithm that simulates the tracking, hunting, and predation processes of a wolf pack. The organizational structure of a complete wolf pack resembles a pyramid, with the top being... The wolf is the leader of the pack, responsible for deciding important matters such as hunting and resting locations and activity times; wolves and Wolves occupy the second and third levels of the hierarchy, and they belong to... Wolves; the lowest rank in a wolf pack is... Wolves obey the command and deployment of higher-ranking wolves when carrying out hunting tasks;

[0082] In the Grey Wolf Algorithm, the optimal solution for each iteration is... The solutions for the wolf, the second and third positions are respectively wolves and Wolf, the rest are Wolf.

[0083] This invention employs an improved Grey Wolf algorithm for model solving; in specific implementation, it includes the following steps:

[0084] A. Initialize the gray wolf population and use it as the current gray wolf population, initializing the first coefficient vector. And use it as the current first coefficient vector to initialize the second coefficient vector. And serve as the current second coefficient vector;

[0085] During initialization, the first coefficient vector is initialized. for ,in for The first random vector within the range; initialize the second coefficient vector. for ,in for The second random vector within the range;

[0086] While the standard sigmoid curve possesses non-linear characteristics, it lacks the ability to translate or adjust its slope. Therefore, an improved sigmoid curve was designed to represent linear vectors. Therefore, the linear vector is initialized using the following formula. And serve as the current linear vector:

[0087]

[0088] In the formula As the first parameter, and t is the current iteration number; T is the set maximum iteration number; It is the second parameter, and ;

[0089] B. Use the objective function of the scheduling model as the fitness function, and calculate the fitness function of the current gray wolf population;

[0090] C. Based on the fitness function obtained in step B, define the gray wolf algorithm... Wolf, wolves and Wolf; simultaneously set the current iteration number t to 1, and set the maximum iteration number T;

[0091] D. If Then proceed with the following steps:

[0092] The target position of the wolf pack in the next iteration was optimized using a random walk algorithm based on Cauchy distribution. Furthermore, to reduce the impact of the random walk algorithm during the exhaustion phase, a nonlinear vector was used. As a multiplier in the random walk algorithm, the wolf pack positions are updated using the following formula:

[0093]

[0094] In the formula The updated location of the wolf pack; For use in simulating gray wolves During the hunting process, the estimation of the prey's location and the corresponding movement are crucial. , For the gray wolf Current location For the gray wolf The dynamic coefficient, For the current gray wolves and gray wolves The distance, and , Let be the first random coefficient vector. For the current Grey Wolf; Used to simulate gray wolves During the hunting process, the estimation of the prey's location and the corresponding movement are crucial. , For the gray wolf Current location For the gray wolf The dynamic coefficient, For the current gray wolves and gray wolves The distance, and , This is the second random coefficient vector; For use in simulating gray wolves During the hunting process, the estimation of the prey's location and the corresponding movement are crucial. , For the gray wolf Current location For the gray wolf The dynamic coefficient, For the current gray wolves and gray wolves The distance, and , The third random coefficient vector; The current linear vector; This is a random hunting increment generated by the random walk algorithm;

[0095] like Then return the current , and serve as the final solution of the Grey Wolf algorithm;

[0096] E. Define the current wolf pack's set. for ; Set the intermediate set U as At the same time, set It is an empty set;

[0097] Repeat steps (1) to (6) until U is an empty set:

[0098] (1) Randomly select two individuals from U. and ;

[0099] (2) To and Make a judgment: If The fitness function value is not greater than If the fitness function value is [value], then the winning individual is set. for At the same time, set up failed individuals for ;like The fitness function value is not less than If the fitness function value is [value], then the winning individual is set. for At the same time, set up failed individuals for ;

[0100] (3) Select the winning individual join in ;

[0101] (4) Use the following formula to calculate Update:

[0102]

[0103]

[0104] In the formula For the updated ; Before the update ; This represents the update amount in round t+1; Let be the update amount in round t; This is the first random variable; It is the second random quantity; For the current ; To update the weight values; It is the third random quantity; For the current winning individual;

[0105] (5) Update the join in ;

[0106] (6) and Delete from U;

[0107] F. Setting for ;

[0108] G. For the first coefficient vector Second coefficient vector and the current linear vector Update, and update at the same time , and ;

[0109] H. Increment the value of the current iteration number t by 1 and return to step D.

[0110] S4. Based on the solution obtained in step S3, complete the scheduling of the target wafer automated material transfer system.

[0111] This invention combines the Grey Wolf Algorithm with a competitive swarm optimization algorithm, offering advantages such as high efficiency, fewer hyperparameters, and competitive and learning capabilities. An improved sigmoid function is used to nonlinearly modify the linear vector, giving the Grey Wolf Algorithm flexibility and ease of adjustment. Finally, a random walk algorithm is added during the wolf pack update process, enabling the Grey Wolf Algorithm to avoid getting trapped in local optima.

[0112] The following comparative example further illustrates the effectiveness of the method of the present invention:

[0113] This comparative example uses the IEEE CEC2020 standard test set proposed by Yue, CT et al. in their 2019 paper "Problem definitions and evaluation criteria for the CEC 2020 special session and competition on single objective bound constrained numerical optimization" for comparison and verification. This test set includes 10 sets of 50-dimensional functions, where F1 and F2 are unimodal problems; F3 and F4 are unimodal and inseparable baseline functions; F5, F6, and F7 are multimodal and separable baseline functions; and F8, F9, and F10 are multimodal and inseparable baseline functions. Python was used as the programming language in the experiment, and the computer environment used was an Intel(R) Core(TM) i9-10900K CPU @ 3.7GHz, 64.0GB RAM, Windows 10. The experiment was repeated 50 times.

[0114] The comparative schemes include the scheme proposed by Mirjalili et al. in the paper "Grey Wolf Optimizer" in 2013 (hereinafter referred to as the GWO scheme), the scheme proposed by Poli et al. in the paper "Particle Swarm Optimization" in 2007 (hereinafter referred to as the PSO scheme), the scheme proposed by Ghasemi et al. in the paper "Phasor Particle Swarm Optimization: A Simple and Efficient Variant of PSO" in 2019 (hereinafter referred to as the PPSO scheme), the scheme proposed by Holland et al. in the paper "Genetic Algorithms" in 1992 (hereinafter referred to as the GA scheme), the scheme proposed by Bilal et al. in the paper "Differential Evolution: A Review of More than Two Decades of Research" in 2020 (hereinafter referred to as the DE scheme), the scheme proposed by Yang et al. in the paper "A New Metaheuristic Bat-Inspired Algorithm" in 2010 (hereinafter referred to as the BA scheme), and the scheme proposed by Mirjalili et al. in the paper "The Whale Optimization" in 2016. The scheme proposed in Algorithm (hereinafter referred to as WOA scheme).

[0115] During the experiment, the mean minimum value and the standard deviation of the minimum value were used as the criteria for judging its performance.

[0116] In addition, the ranking for each objective function is based on the average of the optimal solutions from 50 experiments, while the "overall ranking" is the final ranking of the algorithm's performance based on the "average ranking". The higher the ranking, the better the convergence effect of the algorithm.

[0117] Tables 1 and 2 are schematic tables showing the running results of the method of the present invention and other comparative schemes on the 50-dimensional CEC 2020 test function.

[0118]

[0119]

[0120] As can be seen from Table 1, except for the F8 function, the results of the proposed method are superior to other comparative methods in terms of mean and standard deviation.

[0121] As time progresses, the convergence curves of the present invention and other comparative solutions are as follows: Figures 2-4As shown; via Figures 2-4 As can be seen, the convergence effect of the present invention is significantly better than that of other comparative solutions.

[0122] Therefore, as can be seen from the above comparative examples, the overall performance of the method of the present invention is superior to that of other comparative schemes.

[0123] like Figure 5 The diagram shows the functional modules of the system of the present invention: The system for implementing the scheduling method of the automated wafer material handling system disclosed in this invention includes a data acquisition module, a model building module, a model solving module, and a system scheduling module; the data acquisition module, model building module, model solving module, and system scheduling module are connected in series; the data acquisition module is used to acquire data information of the target automated wafer material handling system and upload the data information to the model building module; the model building module is used to construct a scheduling model of the target automated wafer material handling system based on the received data information and the scheduling cost of the automated wafer material handling system, and upload the data information to the model solving module. The model solving module is used to solve the scheduling model constructed in step S2 based on the received data information, using the Grey Wolf Algorithm, the Competitive Swarm Optimization Algorithm, and the Random Walk Algorithm, and uploads the data information to the system scheduling module. Specifically, the Grey Wolf Algorithm is used to solve the scheduling model; the Competitive Swarm Optimization Algorithm is used to enable individuals in the Grey Wolf Algorithm to compete and learn, thereby improving the accuracy of the Grey Wolf Algorithm; the Random Walk Algorithm is used to alleviate the problem of the Grey Wolf Algorithm getting trapped in local optima, thereby improving the reliability of the Grey Wolf Algorithm; the system scheduling module is used to complete the scheduling of the target wafer automated material handling system based on the received data information and the obtained solution results.

Claims

1. A scheduling method for an automated wafer material handling system, characterized in that... Includes the following steps: S1. Obtain data information from the target wafer automated material handling system; S2. Based on the scheduling cost of the automated wafer material handling system, construct a scheduling model for the target automated wafer material handling system; including the following steps: Based on the scheduling cost of the automated wafer material handling system, with the minimum scheduling cost as the objective function and the transportation capacity of each transportation device and the transportation status of each wafer box as constraints, a scheduling model for the target automated wafer material handling system is constructed. S3. Solve the scheduling model constructed in step S2 based on the gray wolf algorithm, competitive swarm optimization algorithm, and random walk algorithm; including the following steps: The scheduling model constructed in step S2 is solved using the Grey Wolf algorithm; In the solution process, a competitive swarm optimization algorithm is used to optimize the individuals in the gray wolf algorithm; During the solution process, the linear vectors in the Grey Wolf algorithm are optimized based on the sigmoid curve; In the solution process, a random walk algorithm based on Cauchy distribution is used to optimize the target position of the wolf pack in the next iteration, and the optimized linear vector is used as the multiplier of the random walk algorithm. S4. Based on the solution obtained in step S3, complete the scheduling of the target wafer automated material transfer system.

2. The scheduling method of the automated wafer material handling system according to claim 1, characterized in that... Step S1 specifically includes the following steps: The acquired data information of the target wafer automated material handling system includes the location information of the conveying equipment, the total number of conveying equipment, the status information of the conveying equipment, the location information of the wafer cassette, the total number of wafer cassettes, the waiting time information of the wafer cassettes, and the delivery date information of the wafer cassettes.

3. The scheduling method of the automated wafer material handling system according to claim 2, characterized in that... Step S2 specifically includes the following steps: The following formula is used as the objective function of the scheduling model: In the formula, n is the total number of wafer boxes; m is the total number of transport devices; The cost of allocating the j-th transport device to the i-th wafer cassette, and , Let be the distance from the j-th transport device to the i-th wafer cassette. Let be the delivery date of the i-th wafer cell. Let be the waiting time for the i-th wafer cassette. This represents the start-destination buffer state of the i-th wafer cassette. As the first weight value, As the second weight value, As the third weight value, It is the fourth weight value; The state variable is whether the i-th wafer cassette is transported by the j-th transport device; The following formula is used as the constraint condition for the scheduling model: in This is used to limit each wafer cassette to be assigned to only one transport device; This is used to limit each transport device to transport only one wafer cassette at a time.

4. The scheduling method of the automated wafer material handling system according to claim 3, characterized in that... Step S3 specifically includes the following steps: A. Initialize the gray wolf population and use it as the current gray wolf population, initializing the first coefficient vector. And use it as the current first coefficient vector to initialize the second coefficient vector. And serve as the current second coefficient vector; The linear vector is initialized using the following formula. And serve as the current linear vector: In the formula As the first parameter, and ; t is the current iteration number; T is the maximum number of iterations set; It is the second parameter, and ; B. Use the objective function of the scheduling model as the fitness function, and calculate the fitness function of the current gray wolf population; C. Based on the fitness function obtained in step B, define the gray wolf algorithm... Wolf, wolves and Wolf; simultaneously set the current iteration number t to 1, and set the maximum iteration number T; D. If Then proceed with the following steps: The wolf pack's position is updated using the following formula: In the formula The updated location of the wolf pack; For use in simulating gray wolves During the hunting process, the estimation of the prey's location and the corresponding movement are crucial. , For the gray wolf Current location For the gray wolf The dynamic coefficient, For the current gray wolves and gray wolves The distance, and , Let be the first random coefficient vector. For the current Grey Wolf; Used to simulate gray wolves During the hunting process, the estimation of the prey's location and the corresponding movement are crucial. , For the gray wolf Current location For the gray wolf The dynamic coefficient, For the current gray wolves and gray wolves The distance, and , This is the second random coefficient vector; For use in simulating gray wolves During the hunting process, the estimation of the prey's location and the corresponding movement are crucial. , For the gray wolf Current location For the gray wolf The dynamic coefficient, For the current gray wolves and gray wolves The distance, and , The third random coefficient vector; The current linear vector; This is a random hunting increment generated by the random walk algorithm; like Then return the current This serves as the final solution for the final Grey Wolf algorithm. E. Define the current wolf pack's set. for ; Set the intermediate set U as At the same time, set It is an empty set; Repeat steps (1) to (6) until U is an empty set: (1) Randomly select two individuals from U. and ; (2) To and Make a judgment: If The fitness function value is not greater than If the fitness function value is [value], then the winning individual is set. for At the same time, set up failed individuals for ;like The fitness function value is not less than If the fitness function value is [value], then the winning individual is set. for At the same time, set up failed individuals for ; (3) Select the winning individual join in ; (4) Use the following formula to calculate Update: In the formula For the updated ; Before the update ; This represents the update amount in round t+1; Let be the update amount in round t; This is the first random variable; It is the second random quantity; For the current ; To update the weight values; It is the third random quantity; For the current winning individual; (5) Update the join in ; (6) and Delete from U; F. Setting for ; G. For the first coefficient vector Second coefficient vector and the current linear vector Update, and update at the same time , and ; H. Increment the value of the current iteration number t by 1 and return to step D.

5. A system for implementing the scheduling method of the automated wafer material handling system according to any one of claims 1 to 4, characterized in that... The system comprises a data acquisition module, a model building module, a model solving module, and a system scheduling module, which are connected in series. The data acquisition module acquires data information from the target wafer automated material handling system and uploads it to the model building module. The model building module constructs a scheduling model for the target wafer automated material handling system based on the received data and the scheduling cost of the system, and uploads the model solving module to the model solving module. The model solving module solves the scheduling model constructed in step S2 using the Grey Wolf algorithm, competitive swarm optimization algorithm, and random walk algorithm, and uploads the solution to the system scheduling module. The Grey Wolf algorithm is used to solve the scheduling model. The system scheduling module completes the scheduling of the target wafer automated material handling system based on the received data and the solution results.

Images