JP7288190B2 - Information processing device, information processing method and information processing program - Google Patents

Description

The present invention relates to an information processing device, an information processing method, and an information processing program.

Algorithms, such as genetic algorithms, have conventionally been proposed that seek an optimal solution while evolving a plurality of solution candidates (hereinafter also referred to as individuals). In such an algorithm, it is common to give constraint conditions to each design variable, and to calculate an objective function by simulating (hereinafter also referred to as evaluation) an individual based on the constraint conditions.

On the other hand, in the case of a problem in which the incompatibility of each individual with respect to the constraint conditions is found after the evaluation, the objective function obtained by the evaluation inevitably includes the incompatibility. In this case, the designer, for example, adopts a method of generating and evaluating a large number of offspring individuals in advance, calculating each objective function, and extracting those that satisfy the constraint conditions (constraint conditions that are clarified after evaluation). (see Patent Documents 1 and 2, for example).

JP-A-11-085720 Japanese Patent Application Laid-Open No. 2003-248810 JP-A-2002-7999

However, in the algorithm for obtaining the optimum solution as described above, when the number of design variables increases, the number of dimensions and the size of the search space become enormous, and local solutions are likely to occur.

In one aspect, an object of the present invention is to provide an information processing device, an information processing method, and an information processing program capable of stably searching for an optimal solution.

In one aspect, an information processing apparatus is an information processing apparatus that executes optimization calculations using an algorithm for finding an optimum solution while changing a plurality of design variables for each generation, wherein the change in the design variables is a constraint condition. a classifying unit for calculating a degree of contribution indicating the degree of influence exerted on each of the design variables, and classifying the design variables into a plurality of groups based on the magnitude of the degree of contribution; performing optimization calculation using a first group of the design variables, performing a first optimization calculation process in which the first group is changed first, and the remaining groups of the groups of the design variables a processing unit for repeatedly executing a second optimization calculation process for adding and performing optimization calculations until there are no more groups to be added; an output unit for outputting the optimum solution obtained as a result of the processing by the processing unit; It is an information processing device comprising

The optimal solution can be stably searched.

It is a figure showing roughly composition of an information processing system concerning one embodiment. 1 is a perspective view showing a laminated inductor; FIG. 3A and 3B are diagrams (schematic diagrams) for explaining the dimensions of each part of the laminated inductor. 4 is a table showing lower limit values and upper limit values of design variables; FIG. 5(a) is a diagram for explaining the dimension optimization process of the laminated inductor, and FIG. 5(b) is a diagram for explaining the genetic algorithm. FIG. 6A is a diagram showing the hardware configuration of the information processing apparatus, and FIG. 6B is a diagram showing the hardware configuration of the operation terminal. 3 is a functional block diagram of an information processing device and an operation terminal; FIG. 4 is a flowchart showing processing of an information processing device; FIG. 9 is a diagram for explaining a model generated in step S10 of FIG. 8; FIG. FIG. 11 is a diagram (scatter diagram) for explaining step S14; FIG. FIG. 10 is a diagram showing absolute values of assumed change ranges of design variables obtained as a result of step S14; 12(a) to 12(c) are graphs (part 1) showing results of simulation verification. 13(a) to 13(c) are graphs (part 2) showing the results of simulation verification.

An information processing system 100 according to an embodiment will be described in detail below with reference to FIGS. 1 to 13. FIG. FIG. 1 schematically shows the configuration of an information processing system 100 according to one embodiment.

The information processing system 100 includes an information processing device 10 and an operation terminal 20, as shown in FIG. The operation terminal 20 can access the information processing device 10 via the network 80 .

The operation terminal 20 is, for example, one or more PCs (Personal Computers), and is a terminal for the designer to operate the information processing device 10 and for displaying the processing results of the information processing device 10 .

The information processing device 10 is, for example, configured from one or more physical machines, and is a device that performs processing (hereinafter also referred to as optimization calculation processing) for calculating a solution to a predetermined problem using a genetic algorithm. Specifically, the information processing apparatus 10 obtains a solution (optimal solution) to a problem for which non-conformance of each individual (a plurality of design variables) to the constraint conditions is found after evaluation, while evolving each individual.

(Specific example of problem)
A specific example of a problem whose solution is calculated by the optimization calculation process will be described below. The information processing apparatus 10 of this embodiment solves the problem of optimizing various dimensions of a laminated inductor, which is one of the magnetic devices shown in FIG.

FIG. 2 shows a perspective view of a laminated inductor. As shown in FIG. 2, the laminated inductor has a coil with 2.5 coil turns, and magnetic flux is generated by passing current through the coil, producing the inductance (L value) required in the electric circuit. .

A laminated inductor is used as a part of a circuit. Therefore, the laminated inductor is required to satisfy a desired inductance value, to be within a predetermined dimension, and to have as little loss as possible inside the magnetic body. The information processing apparatus 10, for example, solves a design optimization problem of a laminated inductor, and finds the minimum value of loss while using the upper and lower limits of each dimension and the lower limit of inductance as constraints. So in this example loss is the objective function.

3A and 3B are diagrams (schematic diagrams) for explaining the dimensions of each part of the laminated inductor. FIG. 3(a) is a schematic diagram showing the state of the laminated inductor viewed from above (+Z direction) in FIG. 2, and FIG. 3(b) is a schematic diagram showing the state viewed from the side (−Y direction). It is a diagram.

As shown in FIG. 3A, in the laminated inductor, there are Xa, Xb, Xc, Xd, Ya, Yb, Yc, and Yd as parameters of the coil position and the coil width on the XY plane. Furthermore, as shown in FIG. 3(b), there are Z1 to Z11 as dimensional parameters in the Z-axis direction (thickness direction). These 19 parameters are design variables. Assume that upper and lower limits are set for these design variables, as shown in FIG. It is assumed that the overall dimensions of the laminated inductor and the dimensions such as the coil width are determined in advance.

Next, the optimization calculation processing (dimension optimization processing) of the laminated inductor will be described.

First, an example of simultaneous optimization of all design variables of a laminated inductor will be described. FIG. 5(a) is a diagram for explaining the dimension optimization processing of the laminated inductor. Specifically, FIG. 5A is a diagram for explaining the processing performed in the i-th generation. In the optimum design, for example, a magnetic field simulation based on a three-dimensional model is used.

In the optimization, as shown in FIG. 5(a), first, (1) the ith generation individuals are generated, and (2) the dimensions of each individual, Xa to Xd, Ya to Yd, and Z1, are used as design variables. gives ~Z11. Then, (3) if the dimensions of each individual do not satisfy the constraint conditions, the dimension parameters are corrected according to the constraint conditions by imposing limits on the upper and lower limits of the dimensions. As a result, (4) the design variables are changed to Xa'-Xd', Ya'-Yd', and Z1'-Z11'.

Subsequently, (5) the magnetic field simulation is performed to calculate (6) the inductance (L), which is one of the constraints, and the loss (P), which is the objective function.

After that, (7) for the loss (P), which is the objective function, application determination is performed based on the inductance (L), which is one of the constraints found by executing the magnetic field simulation. Specifically, for each i-th generation individual, it is determined whether or not the inductance value calculated by the simulation is equal to or greater than a preset constraint condition value Li.

As a result, when it is determined that the inductance value calculated by the simulation is equal to or greater than the constraint condition value Li, the individual is determined to be applicable. On the other hand, when it is determined that the inductance value calculated by the simulation is less than the constraint condition value Li, the individual is determined to be non-applicable.

In the case of the method described in FIG. 5(a), the designer generates and evaluates a large number of offspring individuals in advance because it is necessary to reach the required number of individuals that satisfy the constraint (inductance value is less than Li). There is a need. As a result, the designer sometimes evaluates an excessive number of individuals. In addition, this method relies heavily on iterative work by trial and error and on the experience of the designer. Therefore, designers may not be able to adopt this method from the viewpoint of work efficiency and reproducibility.

Therefore, as shown in FIG. 5(b), the information processing apparatus 10 of the present embodiment employs a genetic algorithm that seeks an optimal solution while evolving a plurality of individuals for each generation. Individuals of the current generation (child individuals) are generated using the individual as a parent individual. Then, the information processing apparatus 10 evaluates each individual (child individual) of the current generation by using a predetermined evaluation function (objective function).

In other words, in the genetic algorithm, a population (for example, a population consisting of tens to hundreds of individuals) is generated in each generation, and crossover and mutation are performed from among the relatively excellent ones, Generate the next generation of individuals.

Here, in the optimization problem, if the number of design variables is large as in this embodiment (19 in this embodiment), the number of dimensions and the size of the search space become enormous. Local minima are also more likely to occur. Therefore, searching with a genetic algorithm requires an extremely large number of individuals and generations, and the calculation takes a lot of time, making it difficult to reach the optimal solution. For this reason, the information processing apparatus 10 of the present embodiment divides each design variable into two groups based on the magnitude of the influence (referred to as the degree of contribution) of the change in the value of the design variable to the constraint conditions. Then, in the search using the genetic algorithm, the information processing apparatus 10 first performs an optimization calculation using the design variables belonging to the group with the greater degree of contribution, and then performs an optimization calculation using all the design variables. We are planning to perform conversion calculations.

(Regarding the information processing device 10)
Next, the information processing device 10 will be described. FIG. 6A shows the hardware configuration of the information processing device 10. As shown in FIG. As shown in FIG. 6A, the information processing apparatus 10 includes a CPU (Central Processing Unit) 90, a ROM (Read Only Memory) 92, a RAM (Random Access Memory) 94, a storage unit (here, a HDD (Hard Disk Drive )) 96, a network interface 97, a portable storage medium drive 99, and the like. Each component of the information processing apparatus 10 is connected to the bus 98 . In the information processing apparatus 10 , the program (including the information processing program) stored in the ROM 92 or the HDD 96 or the program (including the information processing program) read from the portable storage medium 91 by the portable storage medium drive 99 is executed by the CPU 90 . , the functions of the units shown in FIG. 7 are realized. 7 may be realized by an integrated circuit such as ASIC (Application Specific Integrated Circuit) or FPGA (Field Programmable Gate Array).

In the information processing apparatus 10, the functions shown in FIG. 7 are implemented by the CPU 90 executing a program. As shown in FIG. 7, the information processing apparatus 10 has functions as an information reception unit 30, a model generation unit 32, a magnetic field simulation unit 34, a group classification unit 36, and an optimization calculation unit .

The information reception unit 30 receives information transmitted from the operation terminal 20 (information such as dimensional constraints (lower limit and upper limit) of design variables, predetermined design values of laminated inductors, and physical property values).

The model generator 32 uniformly and randomly sets the values of the design variables within the range of the dimensional constraints, and generates a combined model.

The magnetic field simulation unit 34 executes a magnetic field simulation for each model generated by the model generation unit 32 and calculates a constraint (inductance in this embodiment).

The grouping unit 36 identifies the magnitude (contribution) of the influence of changes in each design variable on the constraint (inductance) based on the value of the design variable of each model and the inductance of each model, Divide each design variable into two groups based on the identified contribution.

The optimization calculation unit 38 performs optimization calculation using the group having the greater degree of contribution of the two groups divided by the group classification unit 36 . The optimization calculation unit 38 then adds the remaining groups and performs optimization calculations. The optimization calculation unit 38 transmits the result of optimization calculation to the operation terminal 20 .

(Regarding operation terminal 20)
Next, the operation terminal 20 will be explained. FIG. 6B shows the hardware configuration of the operation terminal 20. As shown in FIG. As shown in FIG. 6B, the operation terminal 20 includes a CPU 190, a ROM 192, a RAM 194, a storage unit (HDD) 196, a network interface 197, a display unit 193, an input unit 195, a portable storage medium drive 199, and the like. ing. Each component of the operation terminal 20 is connected to the bus 198 . In the operation terminal 20, the CPU 190 executes the program stored in the ROM 192 or HDD 196, or the program read from the portable storage medium 191 by the portable storage medium drive 199, thereby realizing the functions of the units shown in FIG. be done.

In the operation terminal 20, the functions shown in FIG. 7 are implemented by the CPU 190 executing a program. As shown in FIG. 7 , the operation terminal 20 has functions as an information input section 60 , a calculation result receiving section 62 and a display processing section 64 .

The information input unit 60 transmits information used for optimization calculation (information such as dimensional constraints (lower limit and upper limit) of design variables, predetermined design values of laminated inductors, physical property values, etc.) to the information processing apparatus 10 . do.

The calculation result reception unit 62 receives the calculation result by the optimization calculation unit 38 of the information processing device 10 .

The display processing unit 64 displays the calculation result of the optimization calculation unit 38 on the display unit 193 and notifies the designer.

(Processing of information processing device 10)
FIG. 8 shows the processing of the information processing apparatus 10 according to this embodiment in the form of a flowchart. 8 is started when the information reception unit 30 receives information from the information input unit 60 of the operation terminal 20. FIG. The processing shown in the flowchart of FIG. 8 can be performed without manpower by presetting various parameters (for example, the number of individuals to be randomly generated, the number of groups to be classified, the index used for classification, etc.) described later. The processing device 10 can do it automatically.

When the process of FIG. 8 is started, first, in step S10, the model generator 32 uniformly and randomly sets the values of each design variable within the dimension constraint range and generates a combined model. Here, it is thought that the larger the number of models to be created, the more accurately the tendency of the degree of influence (contribution) that changes in each design variable have on the inductance value (constraint). The number shall be 300. In this embodiment, if 50 individuals are evolved 150 generations in the optimization calculation process using a genetic algorithm, which will be described later, the calculation will be performed about 7500 times. It was decided to create only the number of models.

The left side of FIG. 9 shows a nominal model (average model) with model number = 0, and the right side of FIG. 9 shows N models (model numbers = 1 to N) created based on the nominal model. It is shown. Since the design variables (Xa to Z11) of each model are indicated with the model number, if the model number is k, the design variables are written as Xa_k to Z11_k.

Design variables for each of these models are randomly generated within the dimensional constraints shown in FIG. In this embodiment, the design variables are uniformly generated at random, but the design variables may be generated using a method such as an orthogonal array or a Latin superlattice.

Next, in step S12, the magnetic field simulation unit 34 executes a magnetic field simulation using each model (1 to N) and calculates the inductance (L) of each model (1 to N). By executing the magnetic field simulation, the inductance (L) and loss (P) of each model are calculated. and

Next, in step S14, the grouping unit 36 obtains the assumed change range of the inductance for each design variable. Here, the group classification unit 36 focuses on one design variable (for example, Yd) among the design variables and plots it on a scatter diagram with the design variable as the horizontal axis and the absolute value of the inductance as the vertical axis. An example of the results of this processing is shown in FIG. The number of points plotted in the scatter diagram is N ((N+1) if the nominal model is included). Next, the group classification unit 36 fits the results of the plots in FIG. 10 using a first-order approximation formula. This linear approximation is represented by y=α×x+β. Note that, in FIG. 10, α=-4.57×10 ^-12 and β=3.97×10 ^-9 .

Further, the grouping unit 36 obtains the assumed change range (zyd) of the inductance from the following formula (1) from the slope α of this first-order approximation formula and the change amount Dyd of the design variable Yd.
zyd=α×Dyd (1)

For example, if Dyd is 55 mm as shown in FIG. 10 and the slope α is -4.57×10 ^-12 , zyd is -2.51×10 ^-10 .

The grouping unit 36 also performs the same processing as described above for other design variables (other than Yd), and calculates the assumed change range of inductance for each design variable. FIG. 11 shows the absolute value |z| of the assumed change range of each design variable obtained as a result of step S14. It can be said that the magnitude of the absolute value of the assumed change range is a measure of how much the inductance, which is a constraint condition, changes depending on the amplitude of each design variable. That is, the larger the absolute value of the assumed change range, the greater the impact (contribution) of the design variable change on the constraint, and conversely, the smaller the absolute value of the assumed change range, the greater the effect of the design variable change on the constraint. It can be said that the influence (contribution) is small.

Next, in step S16, the group classification unit 36 calculates the average value of the absolute values of the deemed change ranges. Here, it is assumed that the average value of absolute values of assumed change ranges of 19 design variables is 1.32×10 ⁻¹⁰ .

Next, in step S18, the grouping unit 36 classifies each design variable into two groups using the average value calculated in step S16 as a threshold. In this case, in FIG. 11, the group classification unit 36 defines a group including eight design variables whose absolute values of the assumed change ranges are equal to or greater than the average value: 1.32×10 ⁻¹⁰ as group A (first group), Group B is the group containing the 11 design variables below the average.

In the above description, a case has been described in which a plurality of design variables are grouped based on the average value of absolute values of assumed change ranges, but the present invention is not limited to this. For example, the design variables whose magnitude of the absolute value of the deemed change range is up to a predetermined upper rank may be group A, and the design variables below the predetermined rank may be group B. Also, the design variables whose magnitude of the absolute value of the deemed change range is up to the upper predetermined percentage may be group A, and the other design variables may be group B. Also, a value other than the average value of the absolute values of the deemed change range may be used as the threshold.

Next, in step S20, the optimization calculation unit 38 evolves the group of design variables (here, group A) with a larger degree of contribution (absolute value of the assumed change range) by a predetermined number of generations (eg, 50 generations) to optimize the (first optimization calculation process). In this process, the design variables of group B are set to fixed values (for example, the median value of dimensional constraints).

At this time, the optimization calculation unit 38 uses the genetic algorithm shown in FIG. 5(b) to generate the individuals of the current generation using the individuals selected in the previous generation as parent individuals. That is, the optimization calculation unit 38 generates a new individual by genetic manipulation (crossover or mutation) in a genetic algorithm for each generation.

Specifically, the optimization calculation unit 38 randomly selects an individual (initial individual) from among the individuals generated within the dimensional constraints of the design variables, and performs genetic manipulation on the selected individual. Thus, a predetermined number of first generation individuals are generated. In addition, the optimization calculation unit 38 selects parent individuals to be genetically manipulated from the individuals selected in the previous generation in the second and subsequent generations. The optimization calculation unit 38 selects parent individuals from the individuals selected in the previous generation, for example, by performing binary tournament selection. In the binary tournament, a certain number of individuals are randomly selected from the individuals selected in the previous generation, and among them, the one that satisfies the constraint condition (inductance value is less than a predetermined value) and has the smallest objective function (loss) is taken as the parent individual. This is the process of selection. Then, the optimization calculation unit 38 performs crossover and mutation, which are genetic operations, on the selected parent individuals to generate offspring individuals.

After the 50th generation offspring individual is generated through the above processing, the process proceeds to the next step S22.

In step S22, the optimization calculation unit 38 adds a group of design variables with a small contribution (group B) to the design variable group, further evolves a predetermined number of generations (for example, 100 generations), and optimizes (second optimization calculation process). This optimization calculation is similar to the optimization calculation in step S20. Then, in the next step S<b>24 , the optimization calculation unit 38 transmits the optimization calculation result to the calculation result reception unit 62 of the operation terminal 20 . The results of this optimization calculation include the loss and inductance value of the laminated inductor after optimization, and the value of each design variable (each dimensional value of the laminated inductor) when optimized.

The calculation result reception unit 62 of the operation terminal 20 passes the received optimization calculation result to the display processing unit 64 . The display processing unit 64 displays the result of the optimization calculation on the display unit 193 to provide information required by the designer.

All the processing of FIG. 8 is completed by the above.

(About simulation results)
Here, simulation verification performed to confirm the effects of the present embodiment will be described. Here, three methods of setting the design variables used in the optimization calculation using the genetic algorithm were verified. In the verification, the inductance constraint was set to the value of the nominal model (4.3×10 ^-3 [μH]) or more.

The method of setting the design variables to be used is as follows (1) to (3).
(1) All design variables are used simultaneously for optimization calculations.
(2) In the optimization calculations up to the first 50 generations, 11 design variables Z1 to Z11 included in group B with small contribution are used, and in the subsequent optimization calculations up to 150 generations, all design variables are used. use.
(3) As in this embodiment, in the optimization calculations up to the first 50 generations, eight design variables Xa to Xd and Ya to Yd that are included in group A and have a large contribution are used, and up to the subsequent 150 generations All design variables are used in the optimization calculation of

FIG. 12(a) shows a graph in which the number of generations is plotted on the horizontal axis and the loss [W], which is the objective function, is plotted on the vertical axis in case (1). In optimization using a genetic algorithm, loss is set as an objective function, and a combination of design variables that minimizes this is sought. 12(b) and 12(c) show graphs corresponding to FIG. 12(a) in cases (2) and (3).

Referring to FIGS. 12(a) to 12(c), it can be seen that the loss is the smallest in case (3) (FIG. 12(c)). On the other hand, in case (1) (FIG. 12(a)), the loss value obtained in case (3) cannot be reached because it is trapped in a local solution. Also in case (2), as in case (1), it is not possible to reach the loss value obtained in case (3). From these results, as in case (3) (this embodiment), only the group of design variables that contribute greatly to the constraint conditions is optimized first, and then the remaining design variable group is added to perform optimization. It was shown that the method of

In addition, in this simulation, a loss reduction effect of about 27% (5.02×10 ^-2 to 3.65×10 ^-2 [W]) can be confirmed by comparing case (3) and case (1). did it.

As a reference, FIG. 13(a) shows a graph in which the number of generations is plotted on the horizontal axis and the value of inductance (μH) is plotted on the vertical axis in case (1). 13(b) and 13(c) show graphs corresponding to FIG. 13(a) in cases (2) and (3).

In the case of (3), it can be seen that the first 50 generations (optimization of only the group of design variables with large contribution) are optimized near the lower limit of the constraint. Then, it can be read that the optimization proceeds while adjusting the design variables so as to satisfy the constraint conditions in the subsequent optimization (optimization by adding the remaining design variables).

As can be seen from the above description, the model generation unit 32, the magnetic field simulation unit 34, and the group classification unit 36 calculate the contribution of each design variable, and classify the design variables into a plurality of groups based on the degree of contribution. A function as a classifier for classifying into groups is realized. Also, the optimization calculation unit 38 realizes the functions of a processing unit that executes optimization calculation processing and an output unit that outputs the optimum solution obtained as a result of the processing to the operation terminal 20 .

As described in detail above, according to the present embodiment, the grouping unit 36 calculates the magnitude (contribution) of the influence of changes in the design variables on the constraint conditions for each design variable (S14 ), and classify the design variables into two groups A and B based on the degree of contribution (S18). In addition, the optimization calculation unit 38 performs optimization calculation using group A, first changes the design variables included in group A (S20), and adds group B to further perform optimization calculation. Execute (S22). The optimization calculator 38 then outputs the information obtained as a result of the processing (S24). As a result, in this embodiment, after roughly changing the design variables belonging to group A, which has a large degree of contribution to the constraint, all design variables are optimized. can be stably and quickly converted. In particular, as in this embodiment, in optimization calculations using a genetic algorithm that seeks an optimal solution while evolving a plurality of individuals, when it is found that the constraints are met after individual evaluation, the design variables are Even with a large number, stable optimum solution search can be realized.

In addition, in the present embodiment, the dimension values at a plurality of locations of the laminated inductor are used as design variables, and the constraint conditions and objective functions are the inductance and loss. can be provided to the designer.

In this embodiment, the grouping unit 36 generates a predetermined number of models in which random values are set for a plurality of design variables, and the relationship between the calculation result of the constraint conditions of each model and the value of each design variable. Based on this, the degree of contribution is calculated. As a result, it is possible to appropriately calculate, as the degree of contribution of each design variable, a value that indicates the influence of changes in the design variables on the constraint conditions.

In addition, in the present embodiment, the absolute value of the assumed change range of inductance is used as the degree of contribution, so the degree of contribution can be obtained in consideration of the amplitude (dimensional constraint) of each design variable. In this case, the designer does not have to predict the degree of contribution to the constraint conditions of each dimension from the structure of the laminated inductor, etc., and input it, so that the designer can save time and effort.

In the above embodiment, the case where the group classification unit 36 classifies the design variables into two groups has been described, but the invention is not limited to this. For example, the group classification unit 36 may classify into three or more groups. In this case, the optimization calculation unit 38 performs the optimization process using the group with the highest degree of contribution among the three or more groups, then adds groups in descending order of the degree of contribution, and repeats the optimization process. You should do it like this. Further, in the above-described embodiment and modified example, the case where the optimization calculation unit 38 adds groups one by one has been described. However, the number of groups to be added may be changed as appropriate. .

In the above embodiment, the case of using a genetic algorithm in the optimization calculation has been described, but the present invention is not limited to this, and other optimization calculations (for example, simulated annealing, particle method, etc.) may be used. .

Note that the operation terminal 20 may be omitted in the above embodiment. That is, the designer may directly operate the information processing device 10 and the processing result of the information processing device 10 may be displayed on the display section of the information processing device 10 .

In the above embodiment, the case where the information processing device 10 designs the laminated inductor has been described, but the present invention is not limited to this. The information processing apparatus 10 may design other electronic components and the like.

Note that the processing functions described above can be realized by a computer. In that case, a program is provided that describes the processing contents of the functions that the processing device should have. By executing the program on a computer, the above processing functions are realized on the computer. A program describing the processing content can be recorded in a computer-readable storage medium (excluding carrier waves).

When a program is distributed, it is sold in the form of a portable storage medium such as a DVD (Digital Versatile Disc) or a CD-ROM (Compact Disc Read Only Memory) on which the program is recorded. It is also possible to store the program in the storage device of the server computer and transfer the program from the server computer to another computer via the network.

A computer that executes a program stores, for example, a program recorded on a portable storage medium or a program transferred from a server computer in its own storage device. The computer then reads the program from its own storage device and executes processing according to the program. The computer can also read the program directly from the portable storage medium and execute processing according to the program. In addition, the computer can also execute processing in accordance with the received program each time the program is transferred from the server computer.

The embodiments described above are examples of preferred implementations of the present invention. However, the present invention is not limited to this, and various modifications can be made without departing from the gist of the present invention.

In addition, the following additional remarks will be disclosed with respect to the above description of the embodiment.
(Appendix 1) An information processing device that executes optimization calculations using an algorithm for finding an optimum solution while changing a plurality of design variables for each generation,
Classification for calculating, for each of the design variables, a degree of contribution indicating the degree of influence of the change in the design variable on the constraint, and classifying the design variables into a plurality of groups based on the magnitude of the degree of contribution. Department and
performing an optimization calculation using a first group of the group of design variables, and performing a first optimization calculation process in which the first group is changed first; a processing unit for repeatedly executing a second optimization calculation process for adding the remaining groups and executing the optimization calculation until there are no more groups to be added;
An information processing apparatus comprising: an output unit that outputs an optimum solution obtained as a result of processing by the processing unit.
(Appendix 2) The classification unit generates a predetermined number of models in which random values are set for the plurality of design variables, and based on the relationship between the calculation result of the constraint conditions of each model and the value of each design variable, The information processing apparatus according to appendix 1, wherein the degree of contribution is calculated.
(Supplementary Note 3) The processing unit sets the group having the highest contribution as the first group, and adds groups other than the first group in descending order of contribution to the second optimal group. 3. The information processing apparatus according to appendix 1 or 2, characterized in that it executes a conversion calculation process.
(Appendix 4) The information processing apparatus according to any one of Appendices 1 to 3, wherein it is determined whether or not the constraint is satisfied after the optimization calculation.
(Appendix 5) The plurality of design variables are dimensional values at a plurality of locations of a laminated inductor having a coil,
the constraint condition is that the inductance value of the laminated inductor is within a predetermined range;
5. The information processing apparatus according to any one of appendices 1 to 4, wherein in the optimization calculation, loss generated in the laminated inductor is used as an objective function.
(Appendix 6) An information processing method for executing optimization calculations using an algorithm for finding an optimum solution while changing a plurality of design variables for each generation,
calculating, for each of the design variables, a degree of contribution indicating the degree of influence of the change in the design variable on the constraint, and classifying the design variables into a plurality of groups based on the magnitude of the degree of contribution;
performing an optimization calculation using a first group of the group of design variables, and performing a first optimization calculation process in which the first group is changed first; repeating a second optimization calculation process for adding the remaining groups and executing the optimization calculation until there are no more groups to be added;
outputting the optimum solution obtained as a result of the optimization calculation;
An information processing method characterized in that a computer executes processing.
(Appendix 7) In the classifying process, a predetermined number of models in which random values are set for the plurality of design variables are generated, and based on the relationship between the calculation result of the constraint conditions of each model and the value of each design variable , the contribution degree is calculated.
(Appendix 8) The group with the highest degree of contribution is defined as the first group, and among the groups other than the first group, the group with the highest degree of contribution is added in descending order, and the second optimization calculation process is executed. The information processing method according to appendix 6 or 7, characterized by:
(Supplementary note 9) The information processing method according to any one of Supplementary notes 6 to 8, wherein it is determined whether or not the constraint is satisfied after the optimization calculation.
(Appendix 10) The plurality of design variables are dimensional values at a plurality of locations of a laminated inductor having a coil,
the constraint condition is that the inductance value of the laminated inductor is within a predetermined range;
10. The information processing method according to any one of Appendices 6 to 9, wherein in the optimization calculation, a loss generated in the laminated inductor is used as an objective function.
(Appendix 11) An information processing program that causes a computer to execute an optimization calculation using an algorithm for finding an optimum solution while changing a plurality of design variables for each generation,
calculating, for each of the design variables, a degree of contribution indicating the degree of influence of the change in the design variable on the constraint, and classifying the design variables into a plurality of groups based on the magnitude of the degree of contribution;
performing optimization calculation using a first group of the group of design variables, and performing a first optimization calculation process in which the first group is changed first; Repeating the second optimization calculation process for adding the remaining groups and executing the optimization calculation until there are no more groups to add,
outputting the optimum solution obtained as a result of the optimization calculation;
An information processing program characterized by causing the computer to execute processing.

10 information processing device 32 model generation unit (part of classification unit)
34 Magnetic Field Simulation Section (Part of Classification Section)
36 group classifier (part of classifier)
38 optimization calculation unit (processing unit, output unit)

Claims (7)

An information processing device that performs optimization calculations using an algorithm for finding an optimum solution while changing a plurality of design variables for each generation,
Classification for calculating, for each of the design variables, a degree of contribution indicating the degree of influence of the change in the design variable on the constraint, and classifying the design variables into a plurality of groups based on the magnitude of the degree of contribution. Department and
performing an optimization calculation using a first group of the group of design variables, and performing a first optimization calculation process in which the first group is changed first; a processing unit for repeatedly executing a second optimization calculation process for adding the remaining groups and executing the optimization calculation until there are no more groups to be added;
An information processing apparatus comprising: an output unit that outputs an optimum solution obtained as a result of processing by the processing unit. The classification unit generates a predetermined number of models in which random values are set for the plurality of design variables, and calculates the contribution based on the relationship between the calculation result of the constraint conditions of each model and the value of each design variable. 2. The information processing apparatus according to claim 1, wherein the calculation is performed. The processing unit sets the group with the highest degree of contribution as the first group, and adds groups other than the first group in descending order of the degree of contribution to perform the second optimization calculation process. 3. The information processing apparatus according to claim 1, wherein: 4. The information processing apparatus according to any one of claims 1 to 3, wherein it is determined after the optimization calculation whether or not the constraint is satisfied. The plurality of design variables are dimensional values at a plurality of locations of a laminated inductor having a coil,
the constraint condition is that the inductance value of the laminated inductor is within a predetermined range;
5. The information processing apparatus according to claim 1, wherein loss generated in said laminated inductor is used as an objective function in said optimization calculation. An information processing method for executing optimization calculations using an algorithm for finding an optimum solution while changing a plurality of design variables for each generation,
calculating, for each of the design variables, a degree of contribution indicating the degree of influence of the change in the design variable on the constraint, and classifying the design variables into a plurality of groups based on the magnitude of the degree of contribution;
performing an optimization calculation using a first group of the group of design variables, and performing a first optimization calculation process in which the first group is changed first; repeating a second optimization calculation process for adding the remaining groups and executing the optimization calculation until there are no more groups to be added;
outputting the optimum solution obtained as a result of the optimization calculation;
An information processing method characterized in that a computer executes processing. An information processing program that causes a computer to execute an optimization calculation using an algorithm for finding an optimum solution while changing a plurality of design variables for each generation,
calculating, for each of the design variables, a degree of contribution indicating the degree of influence of the change in the design variable on the constraint, and classifying the design variables into a plurality of groups based on the magnitude of the degree of contribution;
performing an optimization calculation using a first group of the group of design variables, and performing a first optimization calculation process in which the first group is changed first; repeating a second optimization calculation process for adding the remaining groups and executing the optimization calculation until there are no more groups to be added;
outputting the optimum solution obtained as a result of the optimization calculation;
An information processing program characterized by causing the computer to execute processing.

Images