Semiconductor wafer manufacturing system scheduling method based on improved genetic algorithm

Abstract

The invention particularly relates to a semiconductor wafer manufacturing system scheduling method based on an improved genetic algorithm, and the method comprises the steps: building a time Petri net model according to a wafer manufacturing system and a processing flow, and determining an initial identifier, a target identifier and a transition emission rule of the time Petri net model; initializing parameters of a genetic algorithm, and generating an initial population by adopting a random search algorithm; the fitness function of the individuals in the population is calculated, the fitness function is updated through a fitness sharing mechanism, and the fitness sharing mechanism comprises the steps that the similarity between the individuals is calculated, and the fitness value is adjusted to maintain population diversity; and performing selection, crossover and mutation operation on the population according to the updated fitness function, and performing iteration to obtain a scheduling result of the time Petri network model. On the basis of the improved genetic algorithm, the exploration space of the Petri network model reachable graph is expanded through a shared fitness mechanism, so that a local optimal solution is jumped out to obtain a better scheduling result.

Description

Technical Field

[0001] The present invention belongs to the technical field of semiconductor wafer manufacturing, and in particular relates to a semiconductor wafer manufacturing system scheduling method based on an improved genetic algorithm. Background Art

[0002] In the context of Industry 4.0, the semiconductor industry, as the foundation and core of the electronic information industry, occupies a crucial strategic position in industrial development. Wafer manufacturing systems, as the most crucial link in the semiconductor production process, are characterized by complex process flows, high-precision equipment requirements, and a high-purity production environment. Traditional wafer manufacturing systems rely primarily on a strictly step-by-step batch processing method, which significantly limits their production efficiency and flexibility. With the introduction of Industry 4.0 technologies, intelligent scheduling of wafer manufacturing systems to improve production efficiency, optimize resource utilization, and shorten production cycles has become a research hotspot.

[0003] As a typical discrete event system, wafer manufacturing systems are modeled using a variety of methods, including finite state automata, Petri nets, machine learning, and data-driven models. Petri nets, as a mathematical modeling tool for describing concurrent systems, can effectively represent event dependencies and resource contention within wafer manufacturing systems. This helps analyze potential conflicts and deadlocks within the system, enabling real-time control and monitoring of the production process.

[0004] In wafer manufacturing systems, the Petri net models established are typically large in scale, making them difficult to solve during actual scheduling. Currently, the main solutions for Petri net scheduling include traditional scheduling algorithms, intelligent scheduling algorithms, and machine learning. Traditional scheduling algorithms primarily include linear programming and its improved methods. These algorithms transform the Petri net scheduling problem into a mathematical optimization problem, solving it by setting an objective function and constraints. These algorithms are more suitable for Petri net models with small scale and clear constraints. Intelligent scheduling algorithms primarily include genetic algorithms, ant colony algorithms, A* algorithms, and greedy algorithms. These algorithms use iterative evolution or breadth-first search based on the reachable graph of the Petri net model to solve the problem. Machine learning models primarily use large amounts of production data to continuously adjust scheduling strategies to achieve adaptive scheduling.

[0005] Genetic algorithms are evolutionary algorithms based on natural selection and heredity. They are highly versatile, require no gradient information, and utilize fitness-driven selection. This makes them highly effective in solving complex search and optimization problems, particularly discrete scheduling tasks. However, in current scheduling problems, genetic algorithms are prone to the rapid emergence of a dominant population during the early stages of the search, leading to premature convergence and a localized optimal solution.

[0006] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present invention, and therefore may include information that does not constitute prior art known to ordinary technicians in this field. Summary of the Invention

[0007] The present invention provides a semiconductor wafer manufacturing system scheduling method based on an improved genetic algorithm, which is used to solve the problem of complex wafer manufacturing system scheduling and can overcome the defects in the prior art to a certain extent.

[0008] Other features and advantages of the present invention will become apparent from the following detailed description, or may be learned in part by practice of the present invention.

[0009] According to a first aspect of the present invention, a semiconductor wafer manufacturing system scheduling method based on an improved genetic algorithm is provided, the method comprising:

[0010] Step S1: Establish a time Petri net model based on the wafer manufacturing system and processing flow, and determine the initial identifier, target identifier and transition emission rules of the time Petri net model;

[0011] Step S2: Initialize the parameters of the genetic algorithm and use the random search algorithm to generate the initial population;

[0012] Step S3: Calculating the fitness function of individuals in the population and updating the fitness function using a fitness sharing mechanism, wherein the fitness sharing mechanism includes calculating the similarity between individuals and adjusting the fitness value to maintain population diversity;

[0013] Step S4: Perform selection, crossover and mutation operations on the population according to the updated fitness function, and obtain the scheduling result of the time Petri net model through iteration.

[0014] In some exemplary embodiments, the transition emission rules of the time Petri net model include:

[0015] The number of tokens in the transition's pre-place satisfies the transition's weight;

[0016] The transition satisfies the time constraint.

[0017] In some exemplary embodiments,

[0018] In some exemplary embodiments, the parameters for initializing the genetic algorithm include: the number of individuals in the population, the maximum number of iterations, the encoding method, the number of mutations, the mutation probability, the number of mutation nodes, the initial identifier, and the termination identifier.

[0019] In some exemplary embodiments, the specific steps of generating the initial population using a random search algorithm include:

[0020] Step S221: add the initial identifier to the queue and set it as the current identifier;

[0021] Step S222: Randomly select a transition and determine whether it satisfies the transition emission rule. If so, emit the transition to the new identifier and update it to the current identifier. Otherwise, repeat step S222. If all transitions do not satisfy the transition emission rule, roll back to the previous identifier and update it to the current identifier and repeat step 222.

[0022] Step S223: Determine whether the current identifier is the termination identifier. If so, stop searching and return the result; otherwise, repeat step S222 until the current identifier is the termination identifier.

[0023] In some exemplary embodiments, the specific steps of updating the fitness function using the fitness sharing mechanism include:

[0024] Step S321: Calculate the distance between individuals;

[0025] Step S322: Calculate the similarity between individuals based on the distance;

[0026] Step S323: Calculate population count based on similarity;

[0027] Step S324: Update the fitness function according to the population count.

[0028] In some exemplary embodiments, the formula used to calculate the distance between individuals is as follows:

[0029]

[0030] Among them, Marking(i) and Marking(j) are the identification sets of individuals i and j.

[0031] In some exemplary embodiments, step 4 includes the following steps:

[0032] Step S41: The selection operation in the genetic algorithm adopts a roulette strategy, which determines the probability of being selected according to the fitness function value of the individual. The higher the fitness of the individual, the greater the probability of being selected. The probability expression is:

[0033]

[0034] Among them, f i 'is the fitness function updated using the shared fitness mechanism;

[0035] Step S42: crossover is performed using a single-point crossover strategy: a crossover operation is performed when the maternal individual and the paternal individual have the same identifier, and the paternal gene fragment and the maternal gene fragment are spliced using the same identifier as the dividing point to produce an offspring individual;

[0036] Step s43: Mutation is performed using a multi-point mutation strategy: mutation sites are selected through multiple random windows and swapped between them. Since swapping mutation sites may cause the mutated transition sequence to not satisfy the transition emission rule, it is necessary to determine whether the transition sequence within the window can be emitted. If the mutated transition sequence satisfies the transition emission rule, the mutation is successful and the mutated individual is retained; otherwise, the mutation fails and the original individual is retained.

[0037] Step s44: Iterate to determine whether the population converges or meets the number of iterations N, where the expression of the convergence judgment condition is

[0038]

[0039] Among them, minTime and maxTime are the shortest and maximum global completion times among individuals in the population;

[0040] Finally, the Petri net model scheduling results are output.

[0041] According to a second aspect of the present invention, a storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the semiconductor wafer manufacturing system scheduling method based on the improved genetic algorithm described in the first aspect is implemented.

[0042] According to a third aspect of the present invention, a computer program product is provided, on which a computer program is stored. When the computer program is executed by a processor, the semiconductor wafer manufacturing system scheduling method based on the improved genetic algorithm described in the first aspect is implemented.

[0043] According to a fourth aspect of the present invention, there is provided an electronic device, comprising:

[0044] processor; and

[0045] a memory for storing executable instructions of the processor;

[0046] Wherein, the processor is configured to implement the semiconductor wafer manufacturing system scheduling method based on the improved genetic algorithm described in the first aspect above by executing the executable instructions.

[0047] The semiconductor wafer manufacturing system scheduling method based on an improved genetic algorithm provided by the embodiments of the present invention utilizes the improved genetic algorithm to solve and obtain scheduling results in a timed Petri net model. During the solution process, a random search algorithm is used to generate the initial population to expand the search space. A fitness sharing mechanism is proposed. While retaining the advantages of the original fitness function, the similarity between individuals in the population is evaluated and applied in subsequent selection, crossover, and mutation operations. This avoids the problem of dominant individuals in the genetic algorithm quickly dominating the population, leading to premature convergence, thereby avoiding local optimal solutions and achieving better results.

[0048] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The accompanying drawings are incorporated into and constitute a part of this specification, illustrate embodiments consistent with the present invention, and together with the description, serve to explain the principles of the present invention. Obviously, the drawings described below are only some embodiments of the present invention, and it is clear that those skilled in the art can derive other drawings based on these drawings without inventive effort.

[0050] Figure 1 A flowchart of a semiconductor wafer manufacturing system optimization algorithm based on an improved genetic algorithm provided by an embodiment of the present invention;

[0051] Figure 2 A schematic diagram of a time Petri net model of a wafer manufacturing system provided by an embodiment of the present invention;

[0052] Figure 3 A schematic diagram of the improved genetic algorithm coding provided by an embodiment of the present invention.

[0053] Figure 4 A schematic diagram of the crossover operation of the improved genetic algorithm provided by an embodiment of the present invention.

[0054] Figure 5 A schematic diagram of the mutation operation of the improved genetic algorithm provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0055] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be embodied in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete and will fully convey the concepts of the example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0056] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the blocks shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0057] The embodiment of the present invention provides a semiconductor wafer manufacturing system scheduling method based on an improved genetic algorithm. Figure 1 , Figure 1 1 is a flow chart of a semiconductor wafer manufacturing system optimization method based on an improved genetic algorithm provided by an embodiment of the present invention. The semiconductor wafer manufacturing system optimization algorithm based on the improved genetic algorithm of this embodiment includes the following steps:

[0058] Step S1: Establish a time Petri net model based on the wafer manufacturing system and processing flow, and determine the initial identifier, target identifier, and transition emission rules of the time Petri net model;

[0059] A Petri net consists of places, transitions, arcs, and tokens. A place represents a system state, a transition represents an operation, and arcs connect places and transitions to represent resource flows. Tokens placed in places represent the current resource quantity. The number and distribution of tokens are called identities. Given that wafer processing time between various devices in a wafer manufacturing system must be strictly controlled, a timed Petri net model was chosen to model the wafer manufacturing system. This model introduces time constraints and can more accurately describe the wafer processing process across various devices.

[0060] The wafer processing flow mainly consists of calibration, feeding, processing, cooling and unloading. Figure 2 As shown in the figure, the wafer manufacturing system consists of an LP transmitter module, an LP receiver module, an AL calibration module, a LAB / LCD atmospheric side module, a PM processing chamber module, a LAB / LCD vacuum side module, an atmospheric side robot (ATR), and a vacuum side robot (VTR). The LP transmitter module and the LP receiver module, respectively, remove wafers to be processed from the wafer cassette and receive processed wafers. The LAB / LCD atmospheric side module and the LAB / LCD vacuum side module are used to transfer wafers between the vacuum side and atmospheric side. The PM module represents the wafer processing chamber module and is used for specific wafer processing steps. The ATR and VTR robots represent transition operations in the Petri net model and are responsible for picking, placing, and swapping wafers between the atmospheric side and vacuum side devices during the processing process.

[0061] The wafer manufacturing system's specific processing flow is as follows: The VTR robot picks up unprocessed wafer modules from the LP transmitter module and transports them to the AL calibration module for calibration. The modules are then moved to the vacuum lock LAB / LCD module for evacuation and vacuumization before entering the processing area. Within the processing area, the VTR robot picks up wafers and processes them sequentially in the PM modules. After processing, the wafers are placed in the LAB / LCD atmospheric module for cooling before entering the LP receiver module.

[0062] Furthermore, the temporal Petri net model constructed by the present invention also includes: the determination of the transition emission rules, initial identifier, and termination identifier of the temporal Petri net model. The transition emission rules of the temporal Petri net are used to clarify whether the identifier of the current state can be transferred to a new identifier. Specifically, to realize the emission of a transition in the temporal Petri net model, the number of tokens in the preceding library of this transition must meet the weight of the transition, and the transition must meet the time constraint, and a new identifier is generated according to the weight after the transition. The initial identifier and the termination identifier are used to represent the initial conditions and termination conditions of the processing process. In this embodiment, the initial identifier represents that the number of tokens in the library in the LP sending module is 13, and the termination state represents that the number of tokens in the library in the LP receiving module is 13.

[0063] Step S2: Initialize the parameters of the genetic algorithm and generate the initial population;

[0064] Step S21: In this embodiment, the parameters of the genetic algorithm include: the number of individuals in the population, the maximum number of iterations, encoding and decoding, the number of mutations, the probability of mutation, the number of mutation nodes, the initial identifier, and the end identifier. Figure 2 As shown, in this embodiment, a combination of transition sequence and identification sequence is used for encoding, and each transition transmission sequence is recorded, and an identification sequence generated by transmitting the corresponding transition sequence is recorded.

[0065] Step S22: The diversity of the initial population in the genetic algorithm affects the final scheduling results. The reachability graph of the time Petri net model consists of all identifier states, and each identifier represents the distribution of the number of tokens in the current Petri net. However, the reachability graph of the Petri net model established for the wafer manufacturing system is large, and the resulting initial population would occupy too much space. Therefore, in this embodiment, a random search algorithm is used to generate the initial population.

[0066] The random search algorithm is more suitable for complex search spaces than the standard breadth-first algorithm. The main difference is that it introduces randomness when selecting nodes to expand (selecting transitions), reducing the path dependence of breadth-first search, helping to explore larger spaces and preventing infinite loops. Because there are markers in the reachable graph where all transitions cannot be enabled, i.e., deadlock markers, the algorithm can roll back to the previous marker and select other transitions to explore the remaining markers. The random search workflow is as follows:

[0067] Step S221: Add an initial identifier to the queue and set it as the current identifier.

[0068] Step S222: Randomly select a transition and determine whether it meets the transition emission rules. If so, emit the transition to the new identifier and update it to the current identifier. Otherwise, repeat step S222. If all transitions do not meet the transition emission rules, roll back to the previous identifier and update it to the current identifier and repeat step S222.

[0069] Step S223: Determine whether the current identifier is the termination identifier. If so, stop searching and return the result. Otherwise, repeat step S222 until the current identifier is the termination identifier.

[0070] Step S3: Calculate the population fitness function and update the fitness function using the fitness sharing mechanism;

[0071] Step S31: The genetic algorithm has the characteristic of fitness-driven selection. The better the quality of the individual, the higher the fitness function calculated. Therefore, the selection of the fitness function is crucial. In this embodiment, the goal is to minimize the global completion time of wafer processing. Considering the long wafer processing time, a normalized fitness function is used, and its formula is:

[0072]

[0073] Where time(i) represents the global completion time of individual i in the population, minTime and maxTime are the shortest and longest global completion time values in the current population, respectively.

[0074] Step S32: In the genetic algorithm, individuals with high fitness within the population are more likely to be selected for subsequent selection and crossover operations, and the resulting offspring are often highly similar to the individuals of the parent and parent generations. This causes these highly similar individuals to quickly dominate the population and form a dominant group. The premature emergence of a dominant group can lead to a loss of population diversity and premature convergence to a local optimal solution. Therefore, a shared fitness mechanism is proposed to reduce the fitness of similar individuals, allowing diverse individuals to have the opportunity to participate in the next generation of reproduction. The operation steps are as follows:

[0075] Step S321: Calculate the distance between individuals. Here, the identifier is used as an indicator to determine the distance between individuals. The Jaccard similarity coefficient is used. The distance expression between individual i and individual j is:

[0076]

[0077] Where Marking(i) and Marking(j) are the identification sets of individual i and individual j.

[0078] Step S322: Calculate the similarity between individuals. The similarity expression between individual i and individual j is:

[0079]

[0080] Among them, d(i,j) is the distance between individual i and individual j, σ s It is the threshold for measuring the similarity between individuals, and a is the coefficient.

[0081] Step S323: Calculate the population count, the formula is:

[0082]

[0083] Among them, N is the number of individuals in the population, and sh(i,j) is the similarity between individuals i and j.

[0084] Step S323: Update the fitness function, whose formula is:

[0085]

[0086] Among them, f i is the individual’s original fitness, m i Count the population.

[0087] Step S4: Perform operations such as selection, crossover, and mutation on individuals in the population according to the updated fitness function, and finally obtain the scheduling result of the time Petri net model through iteration.

[0088] Step S41: The selection operation in the genetic algorithm adopts a roulette strategy, which determines the probability of being selected according to the individual fitness function value. The higher the fitness of the individual, the greater the probability of being selected. Its probability expression is:

[0089]

[0090] Among them, f i ' is the fitness function updated using the shared fitness mechanism.

[0091] Step S42: Crossover adopts the strategy of single-point crossover. Figure 4As shown, when the maternal individual and the paternal individual have the same identifier, a crossover operation is performed, and the paternal gene fragment and the maternal gene fragment are spliced with the same identifier as the dividing point to produce offspring individuals.

[0092] Step s43: Mutation adopts the strategy of multi-point mutation. Figure 4 As shown (taking two-point mutation as an example), mutation sites are selected through multiple random windows and swapped. Since swapping mutation sites may cause the mutated transition sequence to not meet the transition emission rule, it is necessary to determine whether the transition sequence within the window can be emitted. If the mutated transition sequence meets the transition emission rule, the mutation is successful and the mutated individual is retained. Otherwise, the mutation fails and the original individual is retained.

[0093] Step s44: Iterate to determine whether the population converges or meets the number of iterations N, where the expression of the convergence judgment condition is

[0094]

[0095] Where minTime and maxTime are the minimum and maximum global completion times of individuals in the population. The final output is the Petri net model scheduling result.

[0096] Table 1 is a comparison of the global completion time results of 10 runs of the common genetic algorithm and the semiconductor wafer manufacturing system scheduling method based on the improved genetic algorithm designed by the present invention.

[0097] Table 1 Comparison of results between common genetic algorithm and improved genetic algorithm

[0098] Shortest global completion time / s Average global completion time / s Ordinary genetic algorithm 1764.2 1975.32 Improved genetic algorithm 1600.8 1784.01

[0099] Table 1 shows that the improved genetic algorithm designed in this paper has a significant advantage in global makespan. When faced with complex Petri net models, conventional genetic algorithms are prone to falling into local optimal solutions, resulting in limited exploration of the Petri net model's reachability graph. However, the improved genetic algorithm, through a shared fitness mechanism, expands the exploration space of the Petri net model's reachability graph, thereby breaking away from local optimal solutions and achieving better scheduling results.

[0100] The present invention discloses a method for scheduling a semiconductor wafer manufacturing system based on an improved genetic algorithm, which solves a Petri net model using a genetic algorithm to obtain a scheduling result. During the solution process, in order to expand the search space of the Petri net model, a random search algorithm is used to generate an initial population, and the original fitness is updated through a shared fitness mechanism, so that when selecting individuals for subsequent crossover, mutation, and other operations, the population avoids performing operations only between similar individuals, which causes the population to converge prematurely, thereby maintaining population diversity. Furthermore, the initial population generated by the random search algorithm and the individuals selected by the shared fitness mechanism can ensure that individuals with high fitness and similarity within a reasonable range undergo crossover and mutation operations, thereby making it easier to produce excellent offspring individuals, thereby bringing better scheduling results.

[0101] Other embodiments of the present invention will readily occur to those skilled in the art after considering the specification and practicing the invention herein. This application is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, with the true scope and spirit of the invention being indicated by the claims.

[0102] It should be understood that the present invention is not limited to the exact construction described above and shown in the drawings and that various modifications and variations can be made without departing from the scope thereof, which is limited only by the appended claims.

Claims

1. A semiconductor wafer manufacturing system scheduling method based on an improved genetic algorithm, characterized in that: The method comprises: Step S1: Establish a time Petri net model based on the wafer manufacturing system and processing flow, and determine the initial identifier, target identifier and transition emission rules of the time Petri net model; Step S2: Initialize the parameters of the genetic algorithm and use the random search algorithm to generate the initial population; Step S3: Calculating the fitness function of individuals in the population and updating the fitness function using a fitness sharing mechanism, wherein the fitness sharing mechanism includes calculating the similarity between individuals and adjusting the fitness value to maintain population diversity; Step S4: Perform selection, crossover and mutation operations on the population according to the updated fitness function, and obtain the scheduling result of the time Petri net model through iteration.

2. The method according to claim 1, characterized in that The transition emission rules of the time Petri net model include: The number of tokens in the transition's pre-place satisfies the transition's weight; The transition satisfies the time constraint.

3. The method according to claim 1, characterized in that The parameters of the initialized genetic algorithm include: the number of individuals in the population, the maximum number of iterations, the encoding method, the number of mutations, the probability of mutation, the number of mutation nodes, the initial identifier and the termination identifier.

4. The method according to claim 1, wherein The specific steps of using the random search algorithm to generate the initial population include: Step S221: add the initial identifier to the queue and set it as the current identifier; Step S222: Randomly select a transition and determine whether it satisfies the transition emission rule. If so, emit the transition to the new identifier and update it to the current identifier. Otherwise, repeat step S222. If all transitions do not satisfy the transition emission rule, roll back to the previous identifier and update it to the current identifier and repeat step 222. Step S223: Determine whether the current identifier is the termination identifier. If so, stop searching and return the result; otherwise, repeat step S222 until the current identifier is the termination identifier.

5. The method according to claim 1, wherein The specific steps of updating the fitness function using the fitness sharing mechanism include: Step S321: Calculate the distance between individuals; Step S322: Calculate the similarity between individuals based on the distance; Step S323: Calculate population count based on similarity; Step S324: Update the fitness function according to the population count.

6. The method according to claim 1, characterized in that The formula used to calculate the distance between individuals is as follows: Among them, Marking(i) and Marking(j) are the identification sets of individuals i and j.

7. The method according to claim 1, characterized in that The step 4 includes the following steps: Step S41: The selection operation in the genetic algorithm adopts a roulette strategy, which determines the probability of being selected according to the fitness function value of the individual. The higher the fitness of the individual, the greater the probability of being selected. The probability expression is: Among them, f i 'is the fitness function updated using the shared fitness mechanism; Step S42: crossover is performed using a single-point crossover strategy: a crossover operation is performed when the maternal individual and the paternal individual have the same identifier, and the paternal gene fragment and the maternal gene fragment are spliced using the same identifier as the dividing point to produce an offspring individual; Step s43: Mutation is performed using a multi-point mutation strategy: mutation sites are selected through multiple random windows and swapped between them. Since swapping mutation sites may cause the mutated transition sequence to not satisfy the transition emission rule, it is necessary to determine whether the transition sequence within the window can be emitted. If the mutated transition sequence satisfies the transition emission rule, the mutation is successful and the mutated individual is retained; otherwise, the mutation fails and the original individual is retained. Step s44: Iterate to determine whether the population converges or meets the number of iterations N, where the expression of the convergence judgment condition is Among them, minTime and maxTime are the shortest and maximum global completion times among individuals in the population; Finally, the Petri net model scheduling results are output.

8. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the semiconductor wafer manufacturing system scheduling method based on the improved genetic algorithm according to any one of claims 1 to 7 is implemented.

9. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the semiconductor wafer manufacturing system scheduling method based on an improved genetic algorithm according to any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: include: processor; as well as a memory for storing executable instructions of the processor; The processor is configured to execute the semiconductor wafer manufacturing system scheduling method based on an improved genetic algorithm according to any one of claims 1 to 7 by executing the executable instructions.

Images