JP7302938B2 - System design device and method - Google Patents

Description

The present invention relates to a system design apparatus and method for designing various systems that process an input signal and output the processed result as an output signal.

Patent Literature 1 discloses a method of creating a prediction model by multiple regression analysis. Further, Patent Document 2 discloses an input data processing circuit using learning by a neural network.

However, the prediction model creation method based on multiple regression analysis disclosed in Patent Literature 1 can only express linear models, so there is a problem that errors due to linear approximation occur when applied to nonlinear systems.

In addition, in the input data processing circuit using learning by a neural network disclosed in Patent Document 2, the neural network that constitutes the system is completely black boxed, and the process of arithmetic processing for input data is checked. Even if an obviously wrong result is output, it is difficult to investigate the cause.

Japanese Patent No. 5293739 Japanese Patent No. 6183980

The present invention has been made in view of the circumstances of the prior art described above. The purpose is to realize the design of a system that can clearly grasp and analyze it easily.

To achieve the above object, an apparatus according to the present invention is a system design apparatus for designing the configuration of a system for processing an input signal and outputting the processing result as an output signal,
The components of the system are divided into processing elements for processing input signals and parameter elements that define the possible ranges of parameters used in processing by the processing elements. a storage means for storing elements;
system individual generation means for generating a component by combining arbitrarily selected processing elements and parameter elements, and for generating a system individual by combining these components;
computing means for processing input signals for each of the plurality of types of system individuals generated by the system individual generating means;
evaluation means for evaluating an output signal output as a processing result of the calculation means;
an element optimization means for optimizing the constituent elements constituting the individual system based on the evaluation of the output signal by the evaluation means.

A method according to the present invention is a system design method for designing a system that processes an input signal and outputs the processed result as an output signal,
The components of the system are divided into processing elements for processing input signals and parameter elements that define the possible ranges of parameters used when the processing elements process, and multiple types of processing elements and parameter elements are divided set in advance,
Each arbitrarily selected processing element and parameter element are combined to create a component, and the component is combined to generate a system individual,
Generate a plurality of types of system individuals, process input signals for each of the plurality of types of system individuals, evaluate an output signal output as a result of the processing, and evaluate the system individual based on the evaluation of the output signal. It is characterized by optimizing the constituent elements that make up the

As described above, according to the present invention, the processing elements and parameter elements that make up the components of the system are optimized, respectively. It is possible to output a highly accurate optimized result for the input signal, and to design a highly reliable system. Moreover, since it is not necessary to determine the structure of the predictive model in advance, it can be applied to a system whose structure is unknown, and the degree of freedom in design is extremely wide.
Furthermore, since the designed system has clear components and the process of processing the input signal can be grasped, the analysis is easy.

In the system design method of the present invention described above, for the generated system individual,
By arranging the constituent elements that make up the individual system in a grid pattern and transmitting an input signal to each constituent element according to the function of the processing element included in each constituent element, performs signal processing by elements,
Evaluate the output signal output as the processing result,
A method of optimizing the components constituting the individual system based on the evaluation of the output signal can also be used.

By arranging the constituent elements in a grid pattern in this manner and constructing a system individual with a plurality of stages of constituent elements, a more precise system can be constructed. In addition, rapid processing can be realized by transmitting an input signal to each component arranged in a grid according to the function of the processing element and sequentially executing signal processing.

Further, in the above-described system design method of the present invention, a plurality of types of system individuals are generated, input signals are processed for each of the plurality of types of system individuals, and an output signal output as a result of the processing is evaluated. , an initial optimization step of optimizing the components that make up the system individual based on the evaluation of the output signal;
Newly created components are added to the optimized components, multiple types of system individuals are generated, input signals are processed for each of the multiple types of system individuals, and the processing results are output. and an additional optimization step of evaluating the generated output signal and optimizing newly added components included in the system individual based on the evaluation of the output signal.

By optimizing individual systems in a hierarchical manner in this way, it is possible to design step-by-step, starting with important components, and to flexibly design systems according to the required accuracy. Become.

Further, in the above-described system design method of the present invention, a component composed of a combination of a processing element and a parameter element is a gene defined by a genetic algorithm, and a system individual is an individual defined by a genetic algorithm.
A genetic algorithm can be used to evaluate the output signal of an individual system, and optimize the components that make up the individual system based on the evaluation of the output signal.

A genetic algorithm (GA) is already known as a simulation technique for optimizing a system. Using this genetic algorithm method, it is possible to evaluate the output signal of the individual system and optimize the processing elements and parameter elements that make up the individual system with high precision and flexibility.

As described above, according to the present invention, the processing elements and parameter elements of the system components are optimized respectively, so that the input signal can be optimized with high accuracy and reliability, and moreover, It is possible to design a system whose constituent elements are clear, the process of processing an input signal can be clearly grasped, and the analysis thereof is easy.

It is a block diagram showing a schematic structure of a system design device concerning a 1st embodiment of the present invention. FIG. 4 is a diagram showing a specific example of a processing element; FIG. 10 is a diagram showing another specific example of a processing element; It is a figure which shows the structural example of a system individual. 3A and 3B are diagrams showing design examples of an input terminal portion, and FIGS. 4C, 4D, and 4E are diagrams showing design examples of an output terminal portion; FIG. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram for explaining a genetic algorithm, in which (a) shows an initial population, (b) shows an example of crossover, and (c) shows an example of mutation. 1 is a block diagram showing a general configuration of a computer; FIG. It is a block diagram which shows schematic structure of the system design apparatus based on 2nd Embodiment of this invention. It is a figure for demonstrating the system-design method based on 2nd Embodiment of this invention. It is a section front view showing a schematic structure of a powder filling machine. It is a figure for demonstrating the specific example of the system design method based on 2nd Embodiment of this invention. FIG. 12 is a diagram for explaining a specific example of the system design method according to the second embodiment of the present invention following FIG. 11; FIG. 13 is a diagram for explaining a specific example of the system design method according to the second embodiment of the present invention following FIG. 12; FIG. 14 is a diagram for explaining a specific example of the system design method according to the second embodiment of the present invention following FIG. 13; FIG. 15 is a diagram for explaining a specific example of the system design method according to the second embodiment of the present invention following FIG. 14;

The system design apparatus and method according to the present invention basically use a computer to automatically design various systems that process input signals and output the processed results as output signals.
BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

[First embodiment]
FIG. 1 is a block diagram showing a schematic configuration of a system design device according to the first embodiment of the present invention.
The system design apparatus comprises storage means 100 , system initial individual creation means (system individual generation means) 101 , arithmetic means 102 , evaluation means 103 and element optimization means 104 .

The storage means 100 preliminarily stores elements 10 for constructing the system to be designed (that is, "components"). Here, the component 10 consists of options for the "processing element 11" for processing the input signal, and "parameter element 12" that defines the possible range of parameters used when this processing element 11 processes. It is stored in the storage means 100 .

For example, if any system obeys Hooke's Law, the input signal X (elongation of the spring) is multiplied by a gain k (kX) to obtain the output F (force received from the spring). kX in this case corresponds to the processing element 11 and the possible range of the parameter k corresponds to the parameter element 12 .
However, if it is unknown whether any system obeys Hooke's Law, then the processing elements used by the system are unknown. In the present invention, in order to automatically design a system from a system whose processing elements are unknown, the processing elements 11 (kX, k/X, cosX, etc.) that may be used in the system and the parameter k The possible range of (parameter element 12) is stored in the storage means 100.

In the processing elements 11 used in the present embodiment, each of the constituent elements 10 arranged in a lattice as described later is aligned in the row direction (X direction), or in the row direction (X direction) and the column direction (Y direction). ) to input signals from two directions.

2 and 3 are diagrams showing specific examples of processing elements.
In the specific examples shown in these figures, the processing elements 11 are divided into circuit elements and arithmetic elements, and various types are available. A circuit element is a path element that configures a signal transmission path. The processing element 11 shown in FIG. 2 is configured to be able to superimpose and input signals from a plurality of directions. On the other hand, the processing element 11 shown in FIG. 3 is configured to receive signals from only one direction.

Among the processing elements 11 shown in these figures, there is a mixture of those that contain parameters indicated by the coefficients "a" and "b" and those that do not. For example, the path elements, multiplier (MP), sign function (SGN), arcsine (ASIN), arccosine (ACOS), arctangent (ATAN) processing elements 11 do not contain parameters.

Returning to FIG. 1, system initial individual creation means 101 is a component having a function of configuring a system individual using processing elements 11 and parameter elements 12 stored in storage means 100 .
The system initial individual creation means 101 creates an initial system individual in the following procedure. First, the processing element 11 and the parameter element 12 stored in the storage means 100 are arbitrarily selected, and one component 10 is created by combining the selected processing element 11 and parameter element 12 . By repeating this process, a plurality of types of constituent elements 10 are created.

As noted above, some processing elements 11 do not contain parameters. For such a processing element 11, a parameter element 12 meaning no parameter may be combined.

Next, an initial “system individual 13” is generated by arbitrarily combining the plurality of types of the created components 10 . The element optimization means 104, which will be described later, performs an operation for optimizing the constituent elements 10 included in the individual system 13 as a unit.

FIG. 4 is a diagram showing a configuration example when executing signal processing for an individual system. As shown in the figure, in the present embodiment, individual constituent elements 10 constituting the individual system are arranged in a grid pattern, and signal processing is performed by computing means 102 . In this way, by arranging individual components 10 in a grid pattern and configuring the individual system 13 with a plurality of stages of components 10, a more precise system can be configured.

An input terminal 14 is connected to each component 10 arranged in the first row, and an output terminal 15 is connected to each component arranged in the last row. An input signal is input from an arbitrary input terminal 14, and the result processed by the system individual 13 is outputted from an arbitrary output terminal 15 as an output signal.
FIGS. 5(a) and 5(b) are diagrams showing design examples of the input terminal portion, and FIGS. 5(c), 5(d) and 5(e) are diagrams showing design examples of the output terminal portion.
As for the input terminals 14, as shown in FIG. 1(a), it is also possible to configure a signal to be input from one input terminal 14. Alternatively, as shown in FIG. 1(b), all input terminals 14 It is also possible to adopt a configuration in which each signal is input.
As for the output terminals 15, as shown in FIG. may be superimposed and output, or as shown in FIG. Note that these design examples shown in FIG. 5 are merely examples, and it goes without saying that other configurations may be used.

Various configurations of the input terminal portion and the output terminal portion can also be stored in advance in the storage means 100 as the processing elements 11, and can be arbitrarily selected to configure the individual system 13. FIG. This makes it possible to design an optimal system even if the types of input parameters important for the calculation of the output signal are unknown. Further, when designing the system, the user may specify the configuration of the input terminal portion and the output terminal portion, and the individual system may be configured using the configuration specified by the user.

The calculation means 102 performs calculation processing on the input signal for the system individual 13 constructed by the system initial individual creation means 101, and outputs the processing result as an output signal.
In this embodiment, as shown in FIG. 4, for a system individual 13 having a configuration in which components 10 are arranged in a grid pattern, each component is configured according to the function of the processing element 11 included in each component 10. By transmitting an input signal to 10, arithmetic processing is sequentially executed by each component 10, so that rapid processing can be realized.

The evaluation means 103 evaluates the output signal (process result) output from the individual system by the calculation means 102 based on preset evaluation criteria.
For example, the evaluation level can be set according to the magnitude of the difference between the output signal and the true value, and the evaluation result can be based on which evaluation level the output signal from the system individual 13 corresponds to. In addition, the evaluation means 103 can evaluate not only the relationship between input and output, but also evaluation based on various indexes such as disturbance response and target value follow-up characteristics.

The element optimization means 104 optimizes the components 10 (that is, each of the processing elements 11 and the parameter elements 12) that make up the system individual 13 based on the evaluation of the output signal by the evaluation means 103. FIG. In this embodiment, a genetic algorithm technique is used as this optimization technique.
Here, the component 10 (processing element 11 and parameter element 12) corresponds to the gene 20 defined by the genetic algorithm, and the system individual 13 corresponds to the individual 21 defined by the genetic algorithm.

Genetic Algorithm (GA) is a search method that attempts to obtain an optimal solution by performing evolutionary processing that mimics the evolutionary process of living organisms. Specifically, as shown in FIG. 6(a), multiple types of individuals (chromosomes) 21 are generated by combining constituent elements (genes 20) in which information is written. After that, each individual 21 is evaluated for fitness. Then, from the viewpoint of leaving those with high fitness to the next generation, each individual 21 is optimized by genetic manipulation, and the next-generation individuals 21 with further high fitness are generated. While repeating this series of operations, the individual 21 is evolved, and the evolution process ends when the individual 21 with the preset required accuracy is obtained.
Aspects of optimization by genetic engineering for the individual 21 include, for example, selection, crossover, and mutation. FIG. 6(b) shows an example of crossover, in which parts of two individuals are rearranged with each other. FIG. 6(c) shows an example of mutation, in which some genes 20 constituting an individual are changed to other genes 20. FIG.

Using such a genetic algorithm technique, the system individual 13 is evaluated by the evaluation means 103 described above, and the processing elements 11 and parameter elements 12 that make up the system individual 13 are optimized based on the evaluation. , to complete the configuration of the system with the desired accuracy requirements.

The system design apparatus according to the present embodiment, which includes each component shown in FIG. 1, can be specifically configured using a computer. FIG. 7 is a block diagram showing the general configuration of a computer. As shown in the figure, the computer includes components such as a central processing unit (CPU) 200, a RAM (random access memory) 201, a storage unit 202, an input unit 203, and a display unit 204. .

The storage unit 202 includes an external storage device such as a HDD (Hard Disk Drive) or an SSD (Solid State Drive), and a ROM (Read Only Memory) directly connected to the circuit board. The storage unit 100 of the system design apparatus according to the present embodiment is configured by this storage unit 202, and various configurations of the processing elements 11, the parameter elements 12, the input terminal part and the output terminal part, etc. are stored in this storage part 202. be. A program for designing the system according to the procedure described above is also stored in the storage unit 202 in advance.

The central processing unit 200 temporarily stores programs and various data read from the storage unit 202 in the RAM 201, and executes a series of processes according to the programs. This program and the central processing unit 200 constitute system initial individual creation means 101, calculation means 102, evaluation means 103, and element optimization means 104 of the system design apparatus according to this embodiment.

The input unit 203 is a component for inputting data, issuing commands, performing manual operations, and the like through user operations, and is composed of a keyboard, a mouse, and the like. The display unit 204 is composed of a liquid crystal display or the like, and various operation screens in the process of processing, various processing results, and the like are displayed on the display unit 204 .

[Second embodiment]
FIG. 8 is a diagram showing a system design device according to the second embodiment of the present invention, and FIG. 9 is a diagram for explaining a system design method according to the embodiment.

In this embodiment, the constituent elements 10 of the individual system 13 are hierarchized and optimized step by step. For this purpose, the system design apparatus according to this embodiment is provided with component adding means 105 as shown in FIG. The component adding means 105 has a function of creating and adding a new component 10 to the optimized system entity 13 . This component adding means 105 is constituted by the central processing section 200 of the computer shown in FIG.
In FIG. 8, the same reference numerals are given to the same or corresponding means as those in FIG.

In the system design method of this embodiment, for example, when finally designing a system individual 13 in which the constituent elements 10 are arranged in a grid pattern as shown in FIG. The system entity 13 is composed only of the components 10a of the first hierarchy 13a corresponding to the first row (see FIG. 1B).

Then, the arithmetic means 102 arithmetically processes the input signal to the system individual 13 and outputs the processing result as an output signal.

The evaluation means 103 evaluates the signal (process result) output from the system individual 13 by the calculation means 102 based on preset evaluation criteria.

The element optimization means 104 optimizes the component 10a of the first layer 13a based on the evaluation of the output signal by the evaluation means 103 (initial optimization step).
The system individual 13 composed of the components 10a of the first hierarchy 13a is designed by such a procedure.

Next, as shown in FIG. 9(c), a second hierarchy composed of a combination of the newly created components 10b is added to the first hierarchy 13a composed of the optimized components 10a by the component adding means 105. 13b (corresponding to the second row of the lattice) is added to generate the system individual 13 including the component 10a of the first hierarchy 13a and the component 10b of the second hierarchy 13b.

Then, the arithmetic means 102 arithmetically processes the input signal to the system individual 13 and outputs the processing result as an output signal.

The evaluation means 103 evaluates the signal (process result) output from the system individual 13 by the calculation means 102 based on preset evaluation criteria.

The element optimization means 104 optimizes the component 10b of the second hierarchy 13b based on the evaluation of the output signal by the evaluation means 103 (additional optimization step). The object of optimization here is only the component 10b that constitutes the newly added second hierarchy 13b, and the component 10a that constitutes the first hierarchy 13a, for which the optimization has already been completed, is left as it is. Preserve configuration. In such a procedure, the system individual 13 including the component 10a of the first hierarchy 13a and the component 10b of the second hierarchy 13b is designed.

In this way, the component adding means 105 sequentially adds lower layers composed of combinations of the newly created components 10 (see FIG. 9(d)), and optimizes the added components 10. (additional optimization step), finally designing a system individual 13 having a configuration in which the constituent elements 10 are arranged in a grid pattern as shown in FIG. 9(a).

By layering the system individuals 13 and optimizing them sequentially in this way, it is possible to design step by step from the important components 10 . By ending the design when a system that satisfies the required accuracy is completed, flexible system design becomes possible.

According to the system design device and method according to the present embodiment, when creating a lower hierarchy with a combination of newly added components 10, the combination of the components 10 is determined through dialogue between the device and the user. You can also include an operation to It is also possible to reflect the user's knowledge and intention in part of the process of system design, such as processing element preparation and the evaluation method by the evaluation means 103 .
Specifically, commands and data based on the user's knowledge and intention can be input from the input unit 203 shown in FIG. 7, and the input can be reflected in the system design.

[Specific example of the second embodiment]
Next, a specific example for designing a system for a friction compensator of a powder filling machine using the system design method according to the second embodiment described above will be described with reference to FIGS. 10 to 15. FIG.
As shown in FIG. 10 , the powder filling machine stores powder in a funnel-shaped storage portion called a hopper 301 , and rotates an auger screw 303 with the rotational driving force of a motor 302 to cause the hopper 301 to move. The powder is carried downward in the hollow portion of the funnel tube 304, is discharged from the lower end opening of the funnel tube 304, and is filled in the packaging bag arranged below. An agitating member called an agitator 305 is provided in the hopper 301, and the agitator 305 is rotated by the rotational driving force of the motor 302 to agitate the powder in the hopper 301, thereby smoothing the powder. It is designed to be guided to a funnel tube 304 .

Now, in the powder filling machine having such a configuration, an appropriate amount of powder is discharged from the lower end opening of the funnel tube 304 by driving the motor 302 . In order to achieve such quantitative filling, it is necessary to accurately detect the powder torque that the auger screw and the agitator receive from the powder to be filled. However, in the powder filling machine, for example, friction torque acts on the power transmission system of the motor due to various factors. Therefore, a friction compensator is provided for the purpose of removing the friction torque other than the powder torque received by the auger screw and agitator and feeding back the motor to the drive control.

In order to design this friction compensator using the system design method according to the second embodiment described above, for example, steps 1 to 8 shown in FIGS. 11 to 15 may be sequentially executed.
First, as shown in FIG. 11, a plurality of types of system initial individuals 13-1 to 13-N are generated by combining the constituent elements 10 (step 1). These system initial individuals 13-1 to 13-N, for example, to remove basic static friction torque and viscous friction torque that affect the rotation speed (angular speed) of the motor from the processing elements shown in FIG. The processing element 11 is focused on, the constituent elements 10 are created as appropriate, and these constituent elements 10 are combined to generate.
The process of step 1 is executed by the system initial individual creation means 101 shown in FIG.

Next, as shown in FIG. 12, each constituent element 10 is arranged in a lattice for each of the plurality of types of system initial individuals 13-1 to 13-N (step 2). For example, the constituent elements of the system initial individual 13-1 are divided into two constituent element groups 10A and 10B, one constituent element group 10A is arranged in the first row, and the other constituent element group 10B is arranged in the second row. .
Then, the input signal is arithmetically processed for each of the system initial individuals 13-1 to 13-N, and the output signal output as the processing result is evaluated (step 3). Subsequently, based on the evaluation of each system initial individual 13-1 to 13-N, the components of the system individual are optimized by a genetic algorithm (step 4).
The processes of steps 2 to 4 described above are executed by the calculation means 102, the evaluation means 103 and the element optimization means 104 shown in FIG.

After that, as shown in FIG. 13, a new component group is added to the bottom row (third row) of the component groups arranged in a grid forming the optimized system individual (step 5). Here, the group of components to be added is, for example, focusing on the processing elements for removing the friction torque depending on the rotation angle of the motor from the processing elements shown in FIG. You can combine the elements.
The process of step 5 is processed by the component adding means 105 shown in FIG.

Then, the computing means 102 performs arithmetic processing on the input signal for the system individual, the evaluation means 103 evaluates the output signal output as the processing result (step 6), and the element optimization means 104 performs the system individual Based on the evaluation of , a genetic algorithm is used to optimize the components of the individual system (step 7).

The steps 5-7 described above are repeated until a system with satisfactory accuracy is extracted, as shown in FIG.
In other words, the component adding means 105 adds another new component group to the bottom row of the group of components arranged in a lattice that constitutes the optimized individual system (step 5), and On the other hand, the arithmetic means 102 performs arithmetic processing on the input signal, the evaluation means 103 evaluates the output signal output as the processing result (step 6), and the element optimization means 104 evaluates the system individual, A genetic algorithm optimizes the components of the individual system (step 7).

For example, the component adding means 105 focuses on the processing element for removing another frictional torque that depends on the rotation angle of the motor in the fourth line of the component group that constitutes the optimized individual system. It suffices to create constituent elements as appropriate and combine those constituent elements. Furthermore, on the fifth line of the group of components that make up the optimized individual system, create appropriate components by focusing on the processing elements for removing the friction torque that depends on the angular acceleration of the motor, All you have to do is combine those components.

After the optimization of the hierarchized system individual by adding a new component group in this way is completed, the code of the system individual thus optimized is analyzed to obtain the following as shown in FIG. A system circuit can be obtained in which the function of each component can be understood.

It should be noted that the present invention is not limited to the above-described embodiments, and it goes without saying that various modifications and applications are possible as required.
For example, as the optimization method by the element optimization means 104, it is possible to use methods other than the genetic algorithm.
In addition, in the additional optimization step, it is possible to optimize by adding new components not only to the "rows" arranged in a grid pattern but also to the "columns".
In addition, it is possible to optimize by arranging not only two-dimensionally rows and columns, but also multi-dimensionally.

10: component, 11: processing element, 12: parameter element, 13: system individual, 14: input terminal, 15: output terminal,
20: gene, 21: individual,
100: storage means, 101: system initial individual creation means, 102: calculation means, 103: evaluation means, 104: element optimization means, 105: component addition means,
200: central processing unit, 201: RAM, 202: storage unit, 203: input unit, 204: display unit,
301: hopper, 302: motor, 303: auger screw, 304: funnel tube, 305: agitator

Claims (5)

A system design device for designing a system that processes an input signal and outputs the processed result as an output signal,
The components of the system are divided into processing elements for processing the input signal and parameter elements that define possible ranges of parameters used when the processing elements process. storage means for storing said parameter elements of types;
system individual generating means for generating the component by combining the arbitrarily selected processing elements and the parameter elements, and for generating a system individual by combining these components;
computing means for processing input signals for each of the plurality of types of system individuals generated by the system individual generating means;
evaluation means for evaluating an output signal output as a processing result of the calculation means;
an element optimization means for optimizing the constituent elements constituting the system individual based on the evaluation of the output signal by the evaluation means;
A system design device comprising: A system design method for designing a system that processes an input signal and outputs the processed result as an output signal,
The components of the system are divided into processing elements for processing the input signal and parameter elements that define possible ranges of parameters used in processing by the processing elements, and a plurality of types of the processing elements and the Set parameter elements in advance,
combining the arbitrarily selected processing elements and parameter elements to create the constituent elements, and combining these constituent elements to generate a system individual;
generating a plurality of types of system individuals, processing input signals for each of the plurality of types of system individuals, evaluating an output signal output as a result of the processing, and evaluating the output signal based on the evaluation of the output signal; A system design method characterized by optimizing the constituent elements that constitute an individual. For the generated system individual,
By arranging the constituent elements that make up the individual system in a grid pattern and transmitting the input signal to each constituent element according to the function of the processing element included in each constituent element, perform signal processing by each of the components;
Evaluating the output signal output as a result of the processing,
3. The system design method according to claim 2, wherein said components constituting said system individual are optimized based on the evaluation of said output signal. generating a plurality of types of system individuals, processing input signals for each of the plurality of types of system individuals, evaluating an output signal output as a result of the processing, and evaluating the output signal based on the evaluation of the output signal; an initial optimization step of optimizing the components that make up an individual;
adding the newly created component to the optimized component to generate a plurality of types of system individuals, processing an input signal for each of the plurality of types of system individuals, and processing the processing and an additional optimization step of evaluating a resulting output signal and optimizing the newly added component included in the system individual based on the evaluation of the output signal. 3. The system design method according to item 2. Assuming that the constituent element composed of the combination of the processing element and the parameter element is a gene defined by a genetic algorithm, and the system individual is an individual defined by the genetic algorithm,
5. The system according to any one of claims 2 to 4 , wherein a genetic algorithm is used to evaluate the output signal of the system individual and optimize the components constituting the system individual based on the evaluation of the output signal. or the system design method according to item 1.

Images