CN111159915A - Parameter optimization method and device for device design - Google Patents
Parameter optimization method and device for device design Download PDFInfo
- Publication number
- CN111159915A CN111159915A CN202010006351.9A CN202010006351A CN111159915A CN 111159915 A CN111159915 A CN 111159915A CN 202010006351 A CN202010006351 A CN 202010006351A CN 111159915 A CN111159915 A CN 111159915A
- Authority
- CN
- China
- Prior art keywords
- parameters
- electrical performance
- neural network
- network model
- data set
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000013461 design Methods 0.000 title claims abstract description 49
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000005457 optimization Methods 0.000 title claims abstract description 33
- 238000012549 training Methods 0.000 claims abstract description 59
- 238000003062 neural network model Methods 0.000 claims abstract description 49
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 15
- 230000002068 genetic effect Effects 0.000 claims abstract description 14
- 229910002601 GaN Inorganic materials 0.000 claims description 13
- 238000011156 evaluation Methods 0.000 claims description 5
- RNQKDQAVIXDKAG-UHFFFAOYSA-N aluminum gallium Chemical compound [Al].[Ga] RNQKDQAVIXDKAG-UHFFFAOYSA-N 0.000 claims description 4
- 210000002569 neuron Anatomy 0.000 claims description 3
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 2
- 230000004913 activation Effects 0.000 claims description 2
- 229910052782 aluminium Inorganic materials 0.000 claims description 2
- XAGFODPZIPBFFR-UHFFFAOYSA-N aluminium Chemical compound [Al] XAGFODPZIPBFFR-UHFFFAOYSA-N 0.000 claims description 2
- 238000012545 processing Methods 0.000 claims description 2
- JMASRVWKEDWRBT-UHFFFAOYSA-N Gallium nitride Chemical compound [Ga]#N JMASRVWKEDWRBT-UHFFFAOYSA-N 0.000 claims 1
- 230000035772 mutation Effects 0.000 claims 1
- 238000012216 screening Methods 0.000 claims 1
- 230000008569 process Effects 0.000 abstract description 12
- 238000010801 machine learning Methods 0.000 abstract description 8
- 238000004088 simulation Methods 0.000 description 11
- 229910002704 AlGaN Inorganic materials 0.000 description 8
- 238000004364 calculation method Methods 0.000 description 8
- 230000006870 function Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- PMHQVHHXPFUNSP-UHFFFAOYSA-M copper(1+);methylsulfanylmethane;bromide Chemical compound Br[Cu].CSC PMHQVHHXPFUNSP-UHFFFAOYSA-M 0.000 description 1
- 238000012938 design process Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000013213 extrapolation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
The application discloses a parameter optimization method and device for device design, which are used for solving the problems that the time consumption is too long, the efficiency is low and the parameter optimization result is not ideal when the device design is carried out through simulation software. According to preset structural parameters of a device, adopting simulation software to create a training data set, wherein the training data set comprises the structural parameters and the electrical performance parameters of the device; training a neural network model by adopting the established training data set; and determining the optimal solution of the structural parameters of the device corresponding to the target electrical performance parameters through the genetic algorithm reverse-deducing according to the trained neural network model and the target electrical performance parameters of the device. By establishing the neural network model, an agent of simulation software can be formed in a certain structural parameter range of the device, so that the process of device design is accelerated, the efficiency of device design is improved, and parameter optimization of the device can be realized according to the corresponding relation between the structural parameters and the electrical performance parameters of the device through machine learning.
Description
Technical Field
The present application relates to the field of device design technologies, and in particular, to a method and an apparatus for optimizing parameters of a device design.
Background
In the electronics industry, the use of electronic components is essential.
In the design of electronic components, the structural parameters of the electronic components, including the materials, doping concentrations, the width of PN junctions, and the like adopted by the components, determine the response speed, reverse leakage current, current conduction capability, and the like of the components.
At present, when a device is designed, simulation software is usually adopted to simulate the device so as to obtain values of electrical performance parameters corresponding to different structural parameters of the device respectively. However, the relationship between each structural parameter of the device and its electrical property is intricate, and adjusting any one of the structural parameters may result in a change in a plurality of other structural parameters or electrical properties.
For example, in a schottky diode, the larger the PN junction area, the smaller the schottky junction area, the stronger the surge current resistance, but the larger the reverse leakage current.
Therefore, at present, when the simulation of the device design is performed by simulation software, the simulation is performed by trial and error method depending on the experience of the designer. The method is easily limited by self recognition of designers, the required calculation time is too long, and the condition that the simulation result is not converged occurs, so that the simulation efficiency is low.
Disclosure of Invention
The embodiment of the application provides a method and a device for optimizing parameters of device design, which are used for solving the problems of too long time consumption, low efficiency and unsatisfactory parameter optimization result when the device design is carried out through simulation software.
The parameter optimization method for device design provided by the embodiment of the application comprises the following steps:
creating a training data set by adopting simulation software according to preset device structure parameters, wherein the training data set comprises the structure parameters and the electrical performance parameters of a device;
training a neural network model by adopting the established training data set;
and determining the optimal solution of the structural parameters of the device corresponding to the target electrical performance parameters through a genetic algorithm according to the trained neural network model and the target electrical performance parameters of the device.
The parameter optimization device for device design provided by the embodiment of the application comprises:
the device comprises a creating module, a generating module and a processing module, wherein the creating module is used for creating a training data set by adopting simulation software according to preset device structure parameters, and the training data set comprises the structure parameters and the electrical performance parameters of the device;
the training module is used for training the neural network model by adopting the created training data set;
and the determining module is used for determining the optimal solution of the structural parameters of the device corresponding to the target electrical performance parameters through a genetic algorithm according to the trained neural network model and the target electrical performance parameters of the device.
The embodiment of the application provides a parameter optimization method and device for device design, and a training data set is created by acquiring electrical performance parameter values corresponding to structural parameters of a device through simulation software. After the neural network model is trained, the trained neural network model can be used as a proxy model of simulation software in a certain range of structural parameters of a device, and compared with a simulation process with large calculation amount and long time consumption of the simulation software, the neural network model can greatly shorten the calculation time through high-speed calculation, so that the efficiency of device design is improved.
The functional relation between the structural parameters and the electrical performance parameters of the device determined based on the neural network model can be optimized through a genetic algorithm so as to reversely deduce the corresponding structural parameters according to the preset target electrical performance parameters of the device. And optimizing the obtained structural parameters of the device to obtain the optimal structural parameters of the device under the target electrical performance parameters, namely the optimal method for designing the device. The device designed and obtained according to the determined optimal structure parameters can meet the requirements of target electrical performance parameters and meet the functional requirements of the device, and is relatively most reasonable in the aspects of structural layout and the like of device design.
By the device design method, the structural parameter values required by the device under certain target electrical parameters can be quickly determined in a short time, so that the design of the device is completed. Therefore, a large number of manual trial and error and test processes are omitted, time, energy and the like of designers can be effectively saved, and parameter optimization of device design is realized.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:
fig. 1 is a flowchart of a parameter optimization method for device design according to an embodiment of the present disclosure;
fig. 2 is a schematic device structure diagram of an aluminum nitride hemt according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a curve of an electrical property parameter of a machine learning fitting device provided by an embodiment of the present application;
FIG. 4 is a flow chart of another method for optimizing parameters of a device design according to an embodiment of the present disclosure;
fig. 5 is a schematic structural diagram of a parameter optimization apparatus for device design according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be described in detail and completely with reference to the following specific embodiments of the present application and the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
In the embodiments of the present application, the devices include various types of electronic components such as diodes and triodes applied in the electronic industry, and for convenience of description, the embodiments of the present application take an AlGaN/GaN High electron mobility transistor (AlGaN/GaN HEMT) as an example for explanation. It should be noted that, besides the AlGaN/GaN HEMT, other devices not mentioned in the embodiments of the present application may also adopt a method with the same principle to perform parameter optimization of device design.
Fig. 1 is a flowchart of a parameter optimization method for device design provided in an embodiment of the present application, which specifically includes the following steps:
s101: and creating a training data set by adopting simulation software according to the preset structural parameters of the device.
In the embodiment of the application, the server can create the training data set through simulation software according to a plurality of preset structural parameters of the device. The preset structural parameters of the device represent structural parameters which mainly affect the electrical performance of the device in the design of the device. The type, number and the like of the preset structural parameters can be set according to needs, and the application does not limit the types.
Specifically, the server may determine different value ranges corresponding to the respective structural parameters according to preset structural parameters. And determining possible values of the structural parameters in the corresponding value ranges of the structural parameters. Different values of each structural parameter are combined to form a large amount of combined data, and each combined data represents one design condition of the device.
And inputting each combined data into simulation software, and simulating by the simulation software to respectively obtain a group of simulated parameter values of the electrical performance parameters of the device aiming at each combined data. The electrical performance parameter is a parameter representing some aspect of performance of the device, and may include response speed, reverse leakage current, current conduction capability, and the like of the device. The type and number of the electrical performance parameters of the device to be obtained can be set according to the requirement, and the application does not limit the parameters.
In one embodiment, as shown in table 1, a first column in table 1 represents a preset structural parameter of a device, a second column represents a minimum value of the corresponding structural parameter, and a third column represents a maximum value of the corresponding structural parameter.
TABLE 1
| Structural parameters | Min | Max |
| Distance between grid and source (um) | 1 | 2 |
| Grid, drain electrode distance (um) | 5 | 15 |
| Gate electrode thickness (um) | 1 | 3 |
| Film thickness of AIGaN layer (nm) | 20 | 30 |
| Al concentration (atm%) of AIGaN layer | 0.2 | 0.3 |
Specifically, as shown in fig. 2, the device structure of the AlGaN/GaN HEMT is schematically illustrated, and the AlGaN/GaN HEMT includes a source electrode, a drain electrode, a gate electrode, an AlGaN layer, and a GaN layer. Thus, the preset structural parameters of the AlGaN/GaN HEMT device may include five structural parameters, namely, a distance between the gate and the source, a distance between the gate and the drain, a thickness of the gate electrode, a thickness of the aluminum gallium nitride layer, and an aluminum concentration of the aluminum gallium nitride layer.
Within the corresponding value ranges of the five given structural parameters, each structural parameter can be respectively valued, and different values can be randomly combined to determine 500 groups of data. The 500 sets of data may represent 500 different design conditions of the device, respectively.
The data are used as input data of analog simulation software, and simulation is carried out through the simulation software, so that the value of leakage current corresponding to the drain voltage of the AlGaN/GaN HEMT under the corresponding design condition can be obtained.
In the embodiment of the present application, the correspondence between the drain voltage and the drain current of the device corresponding to each design condition may be specifically represented by a form as shown in fig. 3. The abscissa in fig. 3 represents the drain voltage, the ordinate represents the drain current, and the solid-line curve in fig. 3 represents the correspondence between the drain voltage and the drain current of the device obtained by simulation of simulation software under the corresponding design conditions.
Therefore, in the embodiment of the present application, a combination of different values of the five structural parameters and corresponding drain voltage and drain current determined by simulation software may be used as a training data set for training a neural network model in a subsequent step.
S102: and training the neural network model by adopting the created training data set.
In the embodiment of the application, the server can train the neural network model by creating the good training data set.
Specifically, the server may perform model training by using the structural parameters of the device and the corresponding drain voltage as input data for neural network model training and using the corresponding drain current as output data for neural network model training. Through machine learning, the neural network model can learn the corresponding relation between drain voltage and drain current under specific device design conditions.
Further, as shown in fig. 3, the solid curve in fig. 3 represents the electrical performance curve of the device obtained by simulation of the simulation software under certain design conditions. In the training process of the neural network model, the structural parameters and the drain voltage of the device are used as input data, and the corresponding drain current of the device under the corresponding structural parameters and the corresponding drain voltage is used as output data. Therefore, specifically, during training, the neural network model may determine a corresponding relationship between the drain voltage and the drain current of each point under a certain structural parameter, and then connect the corresponding relationships of the points to form a fitting curve, i.e., a dashed curve in fig. 3. Therefore, the process of training the neural network model can also be regarded as a process of approaching the fitted curve generated by the neural network model to the electrical performance curve obtained by simulation.
In one embodiment, the server may employ a BP neural network model for training of the neural network model. In the training process, the training data set may be divided into a training set and a test set, wherein if the training data set is 500 groups, the training set may include 400 groups of training data, and the test set may include 100 groups of training data. The server can train the BP neural network model by adopting the data of the training set, and verify the accuracy and the like of the trained BP neural network model by adopting the data of the test set.
Specifically, the trained BP neural network model may include four fully-connected layers, for a total of 65 neurons. Wherein, the four fully-connected layers respectively comprise 32, 16 and 1 neurons. And when training the BP neural network model, the activation function can adopt Relu, the optimization mode can adopt Adam, and the loss function can adopt average absolute error.
S103: and determining the optimal solution of the structural parameters of the device corresponding to the target electrical performance parameters through the genetic algorithm reverse-deducing according to the trained neural network model and the target electrical performance parameters of the device.
In an embodiment of the present application, a server may determine target electrical performance parameters of a device that need to be achieved. And according to the determined target electrical performance parameters, obtaining the optimal solution of the structural parameters of the device corresponding to the target electrical performance parameters through a genetic algorithm and back-stepping according to the trained neural network model. The steps of finding the optimal solution through the genetic algorithm for the electrical property parameters as specific are as follows:
step one, initializing a population;
determining an evaluation function of the correlation coefficient; the correlation coefficient represents an error between the target electrical property parameter and an electrical property parameter obtained through the neural network model, that is, an error between the target electrical property curve and a predicted electrical property curve obtained through the neural network model. The evaluation function is
Wherein X represents a target electrical performance parameter and Y represents an electrical performance parameter obtained by the neural network model.
Step three, selecting, crossing and mutating;
and step four, repeating the step two and the step three until the value of the correlation coefficient is lower than a preset threshold value. At this time, the structural parameters of the device corresponding to the electrical performance parameters obtained through the neural network model are the optimal structural parameters of the device design under the target electrical performance parameters. The preset threshold value may be set as needed, which is not limited in the present application.
Furthermore, the server can also perform optimization simultaneously according to a plurality of target electrical performance parameters, namely a plurality of target electrical performance curves, in the process of reversely deducing the optimal solution, so as to realize multi-objective optimization and save the time for searching the optimal solution.
Further, if the target electrical performance parameter is simple or the search range is wide, there may be a plurality of optimal solutions. In this case, some preset conditions can be added, target electrical performance parameters can be added, or modes such as anti-interference capability analysis and the like can be adopted to screen a plurality of optimal solutions. The preset condition can be set as required, and the preset condition is not limited in the application. For example, the maximum value of the leakage current does not exceed a preset threshold, and so on.
Fig. 4 is a flowchart of another parameter optimization method for device design according to an embodiment of the present disclosure. As shown in fig. 3, first, for a device to be optimized, a corresponding initial design configuration (i.e., a structural parameter of the device to be optimized) may be obtained, and a value (i.e., a design variable) is assigned to each structural parameter of the device according to a value range of each structural parameter. And then, a data set generated by simulation software in advance is used as a training data set of the neural network model, and after modeling of machine learning is carried out, the functional relation between the structural parameters and the electrical performance parameters of the device can be determined. And determining whether the functional relation obtained through machine learning can complete convergence on the basis of meeting the constraint condition according to a preset target function. If convergence cannot be achieved, iteration and machine learning can be continued through optimization of the optimizer, and if convergence can be achieved, the optimization process can be determined to be finished. The parameter optimization method shown in fig. 4 is substantially the same as the parameter optimization method shown in fig. 1, and for portions not described in detail in fig. 4, reference may be specifically made to the above, and details of this application are not described herein again.
In the parameter optimization of the device design, it can be specifically considered that, in a given design space X, Nx-dimensional design variable X and Ny-dimensional optimization target y have a mapping relationship y ═ f (X), in this case, how to find the design variable X corresponding to the optimal optimization target y on the basis of satisfying each preset constraint condition, that is, a process of parameter optimization, where the process of finding the optimal solution can be expressed as y ═ f (X ∈ X ∩ R ^ opt) (f (X)), opt (·) represents a Pareto optimal solution set in the given search space, and the preset constraint condition can be expressed as g _ i (X) ≦ 0, i ═ 1,2, …, N _ g, and h _ i (X) ═ 0, i ═ 1,2, …, N _ h.
In the embodiment of the present invention, the functions realized by training the neural network model and the genetic algorithm may be realized by other similar machine learning methods and optimization algorithms, and the principle is substantially the same.
When the simulation design of the device is carried out through simulation software, the simulation software needs to calculate various parameters of the device based on a finite element method. Each calculation consumes a great deal of calculation time, resulting in low efficiency of analog simulation.
Therefore, in the embodiment of the application, the neural network model is completed through training, and an agent for simulation software can be formed within a certain structural parameter range of the device. Through the high-speed calculation of the neural network model, the electrical performance parameters corresponding to the structural parameters of the device can be rapidly determined, so that the calculation time is greatly saved, the device design efficiency is accelerated, and the device design process is shortened.
In addition, according to the method and the device, the optimal solution of the structural parameters of the device can be obtained through a genetic algorithm and through reverse estimation according to the target electrical performance parameters of the device, so that the design conditions of the device meeting the requirements can be quickly determined in a short time without manual trial and error, and the determined optimal solution can be ensured to have relatively high accuracy on the basis of machine learning.
Based on the same inventive concept, the above parameter optimization method for device design provided in the embodiments of the present application further provides a parameter optimization device for corresponding device design, as shown in fig. 5.
Fig. 5 is a schematic structural diagram of a parameter optimization apparatus for device design provided in an embodiment of the present application, which specifically includes:
a creating module 401, configured to create a training data set according to preset structural parameters of a device by using simulation software, where the training data set includes the structural parameters and the electrical performance parameters of the device;
a training module 402, configured to train a neural network model using the created training data set;
the determining module 403 is configured to determine, through a genetic algorithm, an optimal solution of the structural parameters of the device corresponding to the target electrical performance parameters by performing a reverse-extrapolation process according to the trained neural network model and the target electrical performance parameters of the device.
The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.
Claims (10)
1. A method for optimizing parameters of a device design, comprising:
creating a training data set by adopting simulation software according to preset device structure parameters, wherein the training data set comprises the structure parameters and the electrical performance parameters of a device;
training a neural network model by adopting the established training data set;
and determining the optimal solution of the structural parameters of the device corresponding to the target electrical performance parameters through a genetic algorithm according to the trained neural network model and the target electrical performance parameters of the device.
2. The method according to claim 1, wherein creating a training data set using simulation software according to preset device configuration parameters specifically comprises:
determining the corresponding electrical performance parameters of the device through simulation software according to the corresponding values of the structural parameters of the device;
and taking the determined values of the electrical performance parameters of the device and the corresponding structural parameters as a training data set.
3. The method of claim 2, wherein the device is a gallium nitride high electron mobility transistor;
the preset structural parameters of the device comprise any one or more of the following parameters: the distance between the grid and the source electrode, the distance between the grid and the drain electrode, the thickness of the grid electrode, the film thickness of the aluminum gallium nitride layer and the aluminum concentration of the aluminum gallium nitride layer.
4. The method of claim 1, wherein training the neural network model using the created training data set specifically comprises:
and training the neural network model to generate the same fitting curve by adopting the created training data set according to the electrical performance curve of the device generated in advance by the simulation software.
5. The method of claim 4, wherein the neural network model comprises 4 fully-connected layers, and wherein the 4 fully-connected layers comprise 32, 16, and 1 neurons, respectively.
6. The method according to claim 5, wherein the neural network model uses Relu as an activation function, Adam as an optimization method, and the mean absolute error as a loss function.
7. The method of claim 1, wherein the step of genetic algorithm comprises:
initializing a population;
determining an evaluation function of the correlation coefficient; wherein the correlation coefficient represents an error between the target electrical performance parameter and an electrical performance parameter obtained by a neural network model, and the evaluation function is
Wherein X represents a target electrical performance parameter and Y represents an electrical performance parameter obtained by a neural network model;
selection, crossover and mutation;
and repeatedly determining the evaluation function of the correlation coefficient and repeatedly selecting, crossing and varying until the value of the correlation coefficient is lower than a preset threshold value.
8. The method of claim 7, further comprising:
and performing multi-objective optimization on the structural parameters of the device corresponding to the corresponding target electrical performance parameters according to the number of the target electrical performance parameters.
9. The method of claim 7, further comprising:
and if a plurality of optimal solutions are determined through a genetic algorithm, screening the optimal solutions by increasing preset conditions.
10. A device design parameter optimization apparatus, comprising:
the device comprises a creating module, a generating module and a processing module, wherein the creating module is used for creating a training data set by adopting simulation software according to preset device structure parameters, and the training data set comprises the structure parameters and the electrical performance parameters of the device;
the training module is used for training the neural network model by adopting the created training data set;
and the determining module is used for determining the optimal solution of the structural parameters of the device corresponding to the target electrical performance parameters through a genetic algorithm according to the trained neural network model and the target electrical performance parameters of the device.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010006351.9A CN111159915A (en) | 2020-01-03 | 2020-01-03 | Parameter optimization method and device for device design |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010006351.9A CN111159915A (en) | 2020-01-03 | 2020-01-03 | Parameter optimization method and device for device design |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111159915A true CN111159915A (en) | 2020-05-15 |
Family
ID=70561316
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010006351.9A Pending CN111159915A (en) | 2020-01-03 | 2020-01-03 | Parameter optimization method and device for device design |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111159915A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111625997A (en) * | 2020-05-27 | 2020-09-04 | 山东天岳先进材料科技有限公司 | Crystal cutting process optimization method and device |
| CN114186442A (en) * | 2020-09-14 | 2022-03-15 | 北京理工大学 | Lattice material parameter optimization method based on neural network model and numerical simulation |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107627152A (en) * | 2017-10-19 | 2018-01-26 | 电子科技大学 | A kind of digital control processing chip control method based on BP neural network |
| CN108595785A (en) * | 2018-03-30 | 2018-09-28 | 西北核技术研究所 | A kind of HPM generation device optimization methods based on multi-objective optimization algorithm |
| CN108959728A (en) * | 2018-06-12 | 2018-12-07 | 杭州法动科技有限公司 | Radio-frequency devices parameter optimization method based on deep learning |
| CN110119558A (en) * | 2019-04-30 | 2019-08-13 | 广州大学 | A kind of circuit parameter optimization method based on difference optimization algorithm |
| CN110162847A (en) * | 2019-05-05 | 2019-08-23 | 东南大学 | Machine learning auxiliary antenna design method based on addition feature policy |
-
2020
- 2020-01-03 CN CN202010006351.9A patent/CN111159915A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107627152A (en) * | 2017-10-19 | 2018-01-26 | 电子科技大学 | A kind of digital control processing chip control method based on BP neural network |
| CN108595785A (en) * | 2018-03-30 | 2018-09-28 | 西北核技术研究所 | A kind of HPM generation device optimization methods based on multi-objective optimization algorithm |
| CN108959728A (en) * | 2018-06-12 | 2018-12-07 | 杭州法动科技有限公司 | Radio-frequency devices parameter optimization method based on deep learning |
| CN110119558A (en) * | 2019-04-30 | 2019-08-13 | 广州大学 | A kind of circuit parameter optimization method based on difference optimization algorithm |
| CN110162847A (en) * | 2019-05-05 | 2019-08-23 | 东南大学 | Machine learning auxiliary antenna design method based on addition feature policy |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111625997A (en) * | 2020-05-27 | 2020-09-04 | 山东天岳先进材料科技有限公司 | Crystal cutting process optimization method and device |
| CN111625997B (en) * | 2020-05-27 | 2024-01-19 | 山东天岳先进科技股份有限公司 | Crystal cutting process optimization method and device |
| CN114186442A (en) * | 2020-09-14 | 2022-03-15 | 北京理工大学 | Lattice material parameter optimization method based on neural network model and numerical simulation |
| CN114186442B (en) * | 2020-09-14 | 2023-06-16 | 北京理工大学 | Lattice material parameter optimization method based on neural network model and numerical simulation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | Learning to design circuits | |
| Liu et al. | A fast differential evolution algorithm using k-Nearest Neighbour predictor | |
| CN111159915A (en) | Parameter optimization method and device for device design | |
| US20220207351A1 (en) | Semiconductor design optimization using at least one neural network | |
| CN114297925A (en) | Power amplifier design method based on feasible domain contraction Bayesian optimization | |
| CN113988558B (en) | Power grid dynamic security assessment method based on blind area identification and electric coordinate system expansion | |
| Elyas et al. | A multi-objective optimization algorithm for bus type assignments in random topology power grid model | |
| CN114841106A (en) | Integrated circuit optimization method and system based on rule-guided genetic algorithm | |
| CN113919221A (en) | Fan load prediction and analysis method and device based on BP neural network and storage medium | |
| KR20220051903A (en) | Method of generating circuit model and manufacturing integrated circuit using the same | |
| Liao et al. | Parasitic-aware gm/I D-based many-objective analog/RF circuit sizing | |
| CN115879412A (en) | Layout level circuit diagram size parameter optimization method based on transfer learning | |
| Qin et al. | A kind integrated adaptive fuzzy neural network tolerance analog circuit fault diagnosis method | |
| Rocha et al. | A new metaheuristc combining gradient models with NSGA-II to enhance analog IC synthesis | |
| Mori et al. | Development of GRBFN with global structure for PV generation output forecasting | |
| CN111211559B (en) | Power grid impedance estimation method based on dynamic step length firefly algorithm | |
| CN114386309A (en) | Problem scale unification method for agent optimization in cloud computing environment | |
| CN108009565B (en) | Method for calculating design flood level of river network area under variable environment | |
| Jereminov et al. | Equivalent circuit programming for power flow analysis and optimization | |
| Khatibi et al. | An improved algorithmic method for software development effort estimation | |
| Barros et al. | GA-SVM optimization kernel applied to analog IC design automation | |
| Coşkun et al. | Equivalence Checking of System-Level and SPICE-Level Models of Static Nonlinear Circuits | |
| JPWO2006107025A1 (en) | Parameter adjustment device | |
| Jabbar et al. | Learning Rate Estimation Model in Restricted Boltizmann Machine Neural Network for Building Recommender Systems | |
| Cutello et al. | Optimising an inductor circuit and a two-stage operational transconductance amplifier using evolutionary and classical algorithms |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20230801 Address after: Huaiyin District of Ji'nan City, Shandong Province, 250118 US 1929 miles Applicant after: JINAN XINGHUO TECHNOLOGY DEVELOPMENT Co.,Ltd. Address before: 250118 Zou Zhuangnan, Meilihu Office, Huaiyin District, Jinan City, Shandong Province Applicant before: Shandong Tianyue Electronic Technology Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200515 |
|
| RJ01 | Rejection of invention patent application after publication |