Lamination model parameter adaptive optimization method and device, equipment and storage medium

Abstract

The embodiment of the invention relates to the field of new energy battery lamination processes, and discloses a lamination model parameter adaptive optimization method, device and equipment and a storage medium, and the method comprises the steps: randomly generating a plurality of particles in a preset parameter space according to a lamination model of a battery lamination machine, and respectively representing a group of lamination model parameters; according to the lamination operation process executed by each group of lamination model parameters, the fitness value of each group of corresponding lamination model parameters is obtained according to the offset of the swing rod of the battery lamination machine; defining the distance and direction of each particle during each iteration movement, obtaining updated lamination model parameters during each iteration, and calculating an updated fitness value; and judging whether a preset convergence condition is met or not according to the updated fitness value, if so, outputting a corresponding lamination model parameter, and if not, continuing iteration. The problems that the iteration time is long and the optimization precision is low due to the fact that polling is carried out by adopting a fixed step length in an existing lamination model parameter optimization method can be solved.

Description

Technical Field

[0001] This invention relates to the field of new energy battery stacking technology, and in particular to a method, apparatus, equipment, and storage medium for adaptive optimization of stacking model parameters. Background Technology

[0002] Currently, most stacking machine models on the market use visual measurement to determine the stacking model parameters based on the shaft position, which results in poor accuracy. Another approach is an adaptive algorithm that directly inputs the stacking model parameters and performs iterative addition and subtraction. This method directly substitutes the parameters into the program to generate the dynamic diaphragm cam curve, then allows the machine to repeat one cycle of motion to obtain the optimal value for the rocker arm offset.

[0003] Because there are many mechanical parameters in the lamination process, for example, if each parameter is polled within a 1mm upper and lower limit, and if polled with a 0.2mm accuracy, each parameter would require 10 polls. Since four of the lamination mechanical parameters directly affect the curve, the traditional polling method would result in 10^4, or 10,000, polls. Since the cycle time for each movement of the laminating machine's coating axis is 20 seconds, the traditional polling method would take 55 hours, which is extremely time-consuming. This solution not only requires several days of polling to determine a more suitable solution, but also only achieves a final accuracy of 0.2mm. Summary of the Invention

[0004] The purpose of this invention is to provide an adaptive optimization method, apparatus, device, and storage medium for stacked chip model parameters, which can solve the problems of long iteration time and low optimization accuracy caused by polling with a fixed step size in existing stacked chip model parameter optimization methods.

[0005] To address the aforementioned technical problems, embodiments of the present invention provide an adaptive optimization method for stacking model parameters, comprising: randomly generating multiple particles in a preset parameter space based on the stacking model of a battery stacking machine, each particle representing a set of stacking model parameters; executing the stacking operation process of the battery stacking machine according to each set of stacking model parameters, obtaining the fitness value of each set of stacking model parameters based on the offset of the battery stacking machine's pendulum during the stacking operation; defining the distance and direction of each particle's movement in each iteration, obtaining updated stacking model parameters based on the distance and direction of movement in each iteration, and calculating the updated fitness value; determining whether a preset convergence condition is met based on the updated fitness value, and if so, outputting the stacking model parameters corresponding to the fitness value; otherwise, continuing the iteration.

[0006] In this embodiment of the invention, the fitness value is inversely proportional to the square of the offset of the pendulum.

[0007] In this embodiment of the invention, defining the distance and direction of each particle's movement in each iteration includes: obtaining the individual optimal solution and the global optimal solution among all particles based on the fitness value calculated in the current iteration; and constructing a velocity vector formula based on the individual optimal solution and the global optimal solution to represent the distance and direction of each particle's movement in each next iteration.

[0008] In this embodiment of the invention, the velocity vector formula is expressed as: ;in, i It is the particle number. d It is the particle dimension index. k It is the number of iterations. It is inertial weight. It is an individual learning factor. It is a group learning factor; , It is a random number within the interval [0-1]. It is a particle i In the k In the nth iteration d dimensional velocity vector It is a particle i In the k In the nth iteration d A dimensional position vector, It is a particle i In the k In the nth iteration d The optimal solution for an individual in dimension. The group is in the first k In the nth iteration d The group optimal solution of dimension.

[0009] In this embodiment of the invention, the inertia weight adopts a linear decreasing strategy, which decreases as the number of iterations increases.

[0010] In this embodiment of the invention, the preset convergence condition is reaching the maximum number of iterations or reaching the target fitness value.

[0011] An embodiment of the present invention also provides an adaptive optimization device for stacking model parameters, comprising: a parameter generation module, configured to randomly generate multiple particles in a preset parameter space according to the stacking model of a battery stacking machine, each particle representing a set of stacking model parameters; a fitness acquisition module, configured to execute the stacking operation process of the battery stacking machine according to each set of stacking model parameters, and obtain the fitness value of each set of stacking model parameters according to the offset of the battery stacking machine's pendulum during the stacking operation; a fitness update module, configured to define the distance and direction of each particle's movement in each iteration, obtain the updated stacking model parameters according to the distance and direction of movement in each iteration, and calculate the updated fitness value; and a convergence judgment module, configured to determine whether a preset convergence condition is met based on the updated fitness value, and if the condition is met, output the stacking model parameters corresponding to the fitness value; otherwise, continue the iteration.

[0012] In this embodiment of the invention, the fitness update module is specifically used to: obtain the individual optimal solution of each particle and the global optimal solution among all particles based on the fitness value calculated in the current iteration; and construct a velocity vector formula based on the individual optimal solution and the global optimal solution to represent the distance and direction of movement of each particle in each next iteration.

[0013] Embodiments of the present invention also provide an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the stacked model parameter adaptive optimization method as described in any of the above embodiments.

[0014] Embodiments of the present invention also provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the adaptive optimization method for stacked model parameters described in any of the above embodiments.

[0015] Compared with the prior art, the above embodiments of the present invention can have at least one or more of the following beneficial effects: The adaptive optimization method for stacked model parameters proposed in this invention utilizes the particle swarm optimization algorithm. It randomly generates multiple sets of stacked model parameters within a preset parameter space, substitutes each set of parameters into the stacking process to obtain a corresponding fitness value, and obtains updated stacked model parameters and updated fitness values ​​based on the distance and direction of movement in each iteration. Finally, it outputs the optimal stacked model parameters when the preset convergence condition is met, based on the fitness value. Compared to the iterative method of adding and subtracting parameters in existing technologies, this scheme can output the optimal solution for stacked model parameters in a very short time, significantly reducing the number of iterations and improving optimization accuracy. Attached Figure Description

[0016] One or more embodiments are illustrated by way of example with reference numerals in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0017] Figure 1 This is a flowchart of an adaptive optimization method for stacked model parameters provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the stacked model structure provided in an embodiment of the present invention; Figure 3 This is a schematic diagram illustrating the change of model parameters over time in an adaptive optimization method for stacked model parameters provided in an embodiment of the present invention. Figure 4 This is a schematic diagram of the structure of an adaptive optimization device for stacked model parameters provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention; Figure 6 This is a schematic diagram of the structure of a computer-readable storage medium provided in an embodiment of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the various embodiments of the present invention will be described in detail below with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details are presented in the various embodiments of the present invention to facilitate a better understanding of this application. However, the technical solutions claimed in this application can be implemented even without these technical details and various variations and modifications based on the following embodiments. The division of the various embodiments below is for ease of description and should not constitute any limitation on the specific implementation of the present invention. The various embodiments can be combined with and referenced by each other without contradiction.

[0019] like Figure 1As shown, the first embodiment of the present invention relates to an adaptive optimization method for stacking model parameters, comprising the following steps: Step S1, randomly generating multiple particles in a preset parameter space according to the stacking model of the battery stacking machine, each particle representing a set of stacking model parameters; Step S2, executing the stacking operation process of the battery stacking machine according to each set of stacking model parameters, and obtaining the fitness value of each set of stacking model parameters according to the offset of the battery stacking machine's pendulum during the stacking operation; Step S3, defining the distance and direction of each particle's movement in each iteration, obtaining the updated stacking model parameters according to the distance and direction of movement in each iteration, and calculating the updated fitness value; Step S4, determining whether a preset convergence condition is met based on the updated fitness value, if met, outputting the stacking model parameters corresponding to the fitness value, if not met, continuing the iteration.

[0020] The following is a detailed description of the implementation details of the adaptive optimization method for stacked model parameters in the first embodiment of the present invention. The following content is only for the convenience of understanding and is not necessary for implementing this solution.

[0021] In step S1, for example, a particle swarm is generated according to a particle swarm optimization algorithm, that is, a set number of coordinates of particles are randomly generated within a preset parameter space. The dimension of the coordinates is represented by numDimensions, and each coordinate represents a set of stacked model parameters. Figure 2 The diagram shows the stacking model of the battery stacking machine in this embodiment. The optimization objective is four stacking model parameters, so the dimension is 4. The coordinates of each particle represent the four parameters of the stacking model. In this embodiment, the number of particles is 20. The initialization process involves randomly assigning different parameters to each of the 20 particles in the parameter space, forming 20 different coordinates. The coordinates of the 20 particles are: (LowerH1, UpperH1, LowerAtoMid1, UpperAtoMid1) - (LowerH1, UpperH2 ... 20 UpperH 20 LowerAtoMid 20 UpperAtoMid 20 The algorithm then proceeds as an adaptive process of iteratively optimizing the coordinate parameters.

[0022] In this embodiment, the preset parameter space is, for example, a range of 1mm above and below the initial parameters. If the range is too large, it will take too long to find the optimal coordinate parameters through iteration. If the range is too small, it may not be able to cover the optimal coordinate parameters, resulting in low optimization accuracy. Therefore, the preset parameter space range can be set according to requirements, and this embodiment is not limited to it.

[0023] In step S2, for example, these particles are brought into the actual system for operation, and the fitness value of the particles to the system is obtained according to the evaluation equation. In this embodiment, the evaluation equation is the sum of the squares of the deviation of the pendulum rod during the cyclic motion of the laminating shaft of the stacking machine from point A to point B. This value is inversely proportional to the fitness value, that is, the smaller the value, the larger the fitness value, which represents a more stable stacking process.

[0024] In step S3, each particle iterates and updates within the parameter space, and the results of each iteration are substituted back into the stacking process to calculate the updated fitness value according to the evaluation equation.

[0025] Specifically, for example, based on all fitness values ​​calculated for each particle in the current iteration, the individual optimal solution Pbest for each particle and the global optimal solution Gbest among all particles are selected. The individual optimal solution Pbest is the optimal solution calculated by the evaluation equation for each of the 20 particle coordinates (LowerH, UpperH, LowerAtoMid, UpperAtoMid) during the iteration process within the stack parameter space. The global optimal solution Gbest is the optimal solution calculated by the evaluation equation for the 20 particle coordinates in the n iterations up to the current iteration.

[0026] The particle swarm size can range from [20, 1000]. For simple problems, a size of 20-40 is generally used, while for more difficult problems or specific types of problems, a size of 100-200 can be used. Smaller swarm sizes are more prone to getting stuck in local optima; larger swarm sizes can improve convergence and find the global optimum faster, but the computational cost per iteration will also increase accordingly; when the swarm size increases to a certain level, further increases will no longer have a significant effect.

[0027] In one implementation, for example, the distance and direction of movement of each particle in each iteration are defined based on the individual optimal solution and the global optimal solution, and expressed by the velocity update formula as follows: ; in, i It is the particle number. d It is the particle dimension index. k It is the number of iterations. It is inertial weight. It is an individual learning factor. It is a group learning factor; , It is a random number within the interval [0-1]. It is a particle i In the k In the nth iteration d 3D velocity vector It is a particle i In the kIn the nth iteration d A dimensional position vector, It is a particle i In the k In the nth iteration d The historical optimal position of a dimension, i.e., the individual optimal solution. The group is in the first k In the nth iteration d The historical optimal position of a dimension is the group's optimal solution.

[0028] The velocity update formula, expressed as velocity, is actually the distance and direction of the particle's next iteration movement, essentially a position vector. The first term is the inertial component, composed of inertial weights and the particle's own velocity, representing the particle's confidence in its previous state of motion. The second term is the cognitive component, representing the particle's own thinking, i.e., its own experience, which can be understood as the distance and direction between the particle's current position and its historical best position. The third term is the social component, representing information sharing and cooperation among particles, derived from the experience of other excellent particles in the group, which can be understood as the distance and direction between the particle's current position and the group's historical best position.

[0029] When solving practical optimization problems, it is often desirable to first employ a global search to quickly converge the search space to a certain region, and then use a local fine-grained search to obtain a high-precision solution. In one implementation, an adaptive adjustment strategy is proposed, where the value of w is linearly decreased as iterations proceed. This allows the particle swarm optimization algorithm to have strong global convergence capability in the early stages and strong local convergence capability in the later stages. The linear decreasing formula is as follows: ; in, It is the maximum inertia weight; It is the minimum inertia weight; This is the current iteration number; It represents the maximum number of iterations.

[0030] In step S4, it is determined whether the current result meets the preset convergence conditions. The preset convergence conditions are divided into two categories: ① the optimal fitness reaches the target fitness value; ② the number of iterations reaches the set maximum number of iterations. In this embodiment, for example, the target fitness is set to 0.00001mm, that is, in one cycle of the stacking machine from point A to point B, the sum of the squares of the pendulum offsets is less than 0.00001mm, which satisfies the condition and stops the iteration to output the result; the maximum number of iterations is set to 20, that is, a maximum of 20 iterations, and the coordinates of the global optimal solution Gbest after 20 iterations are output, which are the corresponding stacking model parameters.

[0031] Furthermore, if the preset convergence condition is met, the optimal stacked model parameters and the number of iterations required to reach the current result are output; if the result does not meet the convergence condition, the position vector of each particle is updated according to the velocity update formula mentioned above based on the current iteration result. The number of iterations can range from [20, 100], with preferred values ​​of 60, 70, and 100. This needs to be adjusted according to the actual situation during the optimization process. If the number of iterations is too small, the solution result will be unstable; if the number of iterations is too large, the whole process will be very time-consuming and unnecessary.

[0032] In this embodiment, the iterative process is the process by which the algorithm, based on the result of the nth iteration, calculates the next particle coordinates (i.e., the coordinates composed of the four stacked model parameters) according to the calculation formula in the algorithm, and continues to substitute them into the next iteration, thus continuously looping to find the optimal stacked model parameters.

[0033] like Figure 3 The diagram shows a program simulation performed in Sysmac Studio. The yellow line, showing a stepped downward trend, represents the global optimal solution Gbest changing over time. The skin-colored line, showing a stepped upward trend, represents the number of iterations changing over time. The other four colored lines represent the parameters of the four stacked models changing over time. In the first iteration, the global optimal value is relatively large, indicating a large pendulum deviation. As the number of iterations increases, the global optimal solution Gbest gradually converges to near 0. During this process, the dimensional coordinates also slowly change to approximate the optimal parameter coordinate values. When Gbest converges to near infinity (zero), it indicates that the pendulum has almost no deviation and runs very smoothly. The output parameter coordinate values ​​at this point are the optimal stacked mechanical parameters (LowerH, UpperH, LowerAtoMid, UpperAtoMid).

[0034] Existing optimization schemes for lamination machine parameters employ a traditional polling method. Given the large number of lamination machine parameters, each is polled within a 1mm upper and lower limit. If polling is performed with a precision of 0.2mm, each parameter requires 10 polls. Since four parameters directly affect the curve, the traditional polling method would result in 10^4 (10,000) iterations. Given that the lamination machine's coating axis travels from point A to point B in 20 seconds, the traditional polling method would take 55 hours, a very long time. This scheme not only requires several days of polling to determine a more suitable solution but also only achieves a final precision of 0.2mm.

[0035] The adaptive optimization method for stacked model parameters proposed in this embodiment is based on the PSO particle swarm optimization algorithm. For example, by iterating with 20 particles, the optimal global solution Gbest can approach 0 after 20 iterations. Therefore, it only takes 400 actions from the start of polling to the iteration to obtain the optimal solution. Assuming each iteration takes 20 seconds, the global optimal solution can be calculated in just 2 hours. Moreover, simulations show that the accuracy of the optimal solution can reach 0.0001 mm in 20 iterations. The accuracy increases with the number of iterations, thus significantly improving the efficiency and accuracy of stacked model parameter optimization.

[0036] The steps described above are for clarity only. In practice, they can be combined into one step or some steps can be broken down into multiple steps. As long as they include the same logical relationship, they are all within the scope of protection of this patent. Adding insignificant modifications or introducing insignificant designs to the algorithm or process, but without changing the core design of the algorithm and process, are also within the scope of protection of this patent.

[0037] Furthermore, the examples mentioned in the above embodiments can be freely combined, and any combination can be understood as an embodiment. The terms "embodiment" or "example" appearing in various locations in the specification do not necessarily refer to the same embodiment, nor are they independent or alternative embodiments mutually exclusive with other embodiments. Those skilled in the art will understand that the embodiments described herein can be combined with other embodiments.

[0038] like Figure 4 As shown, another embodiment of the present invention relates to an adaptive optimization device 20 for stacking model parameters, comprising: a parameter generation module 201, a fitness acquisition module 202, a fitness update module 203, and a convergence judgment module 204. The parameter generation module 201 is used to randomly generate multiple particles in a preset parameter space based on the stacking model of the battery stacking machine, each particle representing a set of stacking model parameters; the fitness acquisition module 202 is used to execute the stacking operation process of the battery stacking machine according to each set of stacking model parameters, and obtain the corresponding fitness value of each set of stacking model parameters based on the offset of the battery stacking machine's pendulum during the stacking operation; the fitness update module 203 is used to define the distance and direction of each particle's movement in each iteration, obtain the updated stacking model parameters based on the distance and direction of movement in each iteration, and calculate the updated fitness value; the convergence judgment module 204 is used to determine whether a preset convergence condition is met based on the updated fitness value. If the condition is met, the corresponding stacking model parameter is output; otherwise, the iteration continues.

[0039] In one implementation, the fitness update module 203 is specifically used to: obtain the individual optimal solution for each particle and the global optimal solution among all particles based on the fitness value calculated in the current iteration; and construct a velocity vector formula based on the individual optimal solution and the global optimal solution to represent the distance and direction of movement of each particle in each next iteration.

[0040] It is not difficult to see that this embodiment is a device embodiment corresponding to the above method embodiments, and this embodiment can be implemented in conjunction with the above method embodiments. The relevant technical details mentioned in the above method embodiments are still valid in this embodiment, and will not be repeated here to reduce repetition. Accordingly, the relevant technical details mentioned in this embodiment can also be applied to the above method embodiments.

[0041] It is worth mentioning that all modules involved in this embodiment are logical modules. In practical applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. Furthermore, to highlight the innovative aspects of this invention, this embodiment does not introduce units that are not closely related to solving the technical problem proposed by this invention; however, this does not mean that other units are absent from this embodiment.

[0042] like Figure 5 As shown, another embodiment of the present invention relates to an electronic device 30, including at least one processor 31; and a memory 32 communicatively connected to at least one processor 31; wherein the memory 32 stores instructions executable by at least one processor 31, the instructions being executed by at least one processor 31 to enable at least one processor 31 to perform the stacked model parameter adaptive optimization method as described above.

[0043] The memory 32 and processor 31 are connected via a bus, which can include any number of interconnecting buses and bridges, connecting various circuits of one or more processors and the memory. The bus can also connect various other circuits, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. A bus interface provides an interface between the bus and the transceiver. The transceiver can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by the processor is transmitted over the wireless medium via an antenna, which further receives data and transmits it to the processor 31.

[0044] Processor 31 is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory 32 can be used to store data used by processor 31 during operation.

[0045] like Figure 6 As shown, another embodiment of the present invention relates to a computer-readable storage medium 40 storing a computer program. When the computer program is executed by a processor, it implements the above-described method embodiment.

[0046] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0047] Those skilled in the art will understand that the above embodiments are specific embodiments for implementing the present invention, and in practical applications, various changes in form and detail may be made without departing from the spirit and scope of the present invention.

Claims

1. An adaptive optimization method for parameters of a stacked chip model, characterized in that, include: Multiple particles are randomly generated in a preset parameter space based on the stacking model of the battery stacking machine, and each particle represents a set of stacking model parameters; The stacking process of the battery stacking machine is executed according to the stacking model parameters of each group, and the fitness value of each group of stacking model parameters is obtained according to the offset of the battery stacking machine swing arm during the stacking process. Define the distance and direction of movement of each particle in each iteration, obtain the updated stack model parameters based on the distance and direction of movement in each iteration, and calculate the updated fitness value; Based on the updated fitness value, determine whether the preset convergence condition is met. If it is met, output the stacked model parameters corresponding to the fitness value. If it is not met, continue the iteration.

2. The adaptive optimization method for stacked model parameters according to claim 1, characterized in that, The fitness value is inversely proportional to the square of the offset of the pendulum.

3. The adaptive optimization method for stacked model parameters according to claim 1, characterized in that, The definition of the distance and direction of movement of each particle in each iteration includes: Based on the fitness value calculated in the current iteration, obtain the individual optimal solution for each particle and the global optimal solution among all particles; Based on the individual optimal solution and the global optimal solution, a velocity vector formula is constructed to represent the distance and direction of movement of each particle in each next iteration.

4. The adaptive optimization method for stacked model parameters according to claim 3, characterized in that, The velocity vector formula is expressed as follows: ； in, i It is the particle number. d It is the particle dimension index. k It is the number of iterations. It is inertial weight. It is an individual learning factor. It is a group learning factor; , It is a random number within the interval [0-1]. It is a particle i In the k In the nth iteration d dimensional velocity vector It is a particle i In the k In the nth iteration d A dimensional position vector, It is a particle i In the k In the nth iteration d The optimal solution for an individual in dimension. It is the group in the first k In the nth iteration d The group optimal solution of dimension.

5. The adaptive optimization method for stacked model parameters according to claim 4, characterized in that, The inertia weight adopts a linear decreasing strategy, which decreases as the number of iterations increases.

6. The adaptive optimization method for stacked model parameters according to claim 1, characterized in that, The preset convergence condition is reaching the maximum number of iterations or reaching the target fitness value.

7. An adaptive optimization device for stacked model parameters, characterized in that, include: The parameter generation module is used to randomly generate multiple particles in a preset parameter space based on the stacking model of the battery stacking machine, where each particle represents a set of stacking model parameters. The fitness acquisition module is used to execute the stacking operation process of the battery stacking machine according to each set of stacking model parameters, and obtain the corresponding fitness value of each set of stacking model parameters based on the offset of the battery stacking machine swing arm during the stacking operation. The fitness update module is used to define the distance and direction of each particle's movement in each iteration, obtain the updated stack model parameters based on the distance and direction of movement in each iteration, and calculate the updated fitness value. The convergence judgment module is used to determine whether the preset convergence condition is met based on the updated fitness value. If the condition is met, the corresponding stacked model parameters are output. If the condition is not met, the iteration continues.

8. The adaptive optimization device for stacked model parameters according to claim 7, characterized in that, The fitness update module is specifically used to: obtain the individual optimal solution for each particle and the global optimal solution among all particles based on the fitness value calculated in the current iteration; and construct a velocity vector formula based on the individual optimal solution and the global optimal solution to represent the distance and direction of movement of each particle in each next iteration.

9. An electronic device, characterized in that, include: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the stacked model parameter adaptive optimization method as described in any one of claims 1 to 6.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the adaptive optimization method for stacked model parameters as described in any one of claims 1 to 6.

Images