JPH02162466A - Back-up method for optimum design - Google Patents

Abstract

PURPOSE:To effectively perform the optimum design by storing the contents and the result of each correction and the history of the correction contents securing the optimum design into a computer, producing automatically the correction plans for the similar memory contents of the computer first, and carrying out the simulation of the correction plans. CONSTITUTION:A 1st data memory means 11 is prepared to store the target output, an input parameter, etc., together with a 2nd data memory means 12 which stores the correction contents and the correction process history of the parameter, an extraction means 22 which extracts the parameter correction contents and correction process history, a retrieving means 23 which retrieves the similar corrections, a control means 41 which controls these means 11 - 23 and an information process means, etc. The correction processes are extracted and successively stored for each correction of the input parameter. At the same time, the record of the correction process where the similar relation is secured between the simulation result output and the target output is retrieved and taken out via the means 22. Then the input parameter is corrected by reference to the taken-out record. Thus the optimum design is effectively carried out.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a computer-based optimal design support method, particularly when input parameters are optimized by trial and error so that the output of a simulator becomes a target output. This invention relates to a suitable optimal design support method.

[Conventional technology]

Conventionally, when supporting optimal design with a general-purpose computer,
Various mathematical programming methods have been used.

In addition, methods for supporting optimization by introducing designers' know-how into trial-and-error design calculations are described in, for example, 1 Japan System Science Journal, Volume 8, Number 2゜ (1984), No. 3.
Discussed on pages 12 to 12.

In this method, a designer specifies in advance know-how such as the selection of input parameters to be changed, the priority order of changes, and the range of changes, and a computer performs optimization based on the know-how.

[Problem to be solved by the invention]

In optimization using the above mathematical programming method, the main body of the optimal design is a computer, which can save labor, but the method itself is general-purpose and difficult to incorporate information appropriate to the problem, so veteran designers generally use their know-how. This is particularly problematic when the simulator is less efficient than optimizing it, and the load on each simulator is large. Furthermore, since it is generally difficult to sufficiently reflect the designer's intentions, there is also a problem in that the optimal solution finally obtained does not match the designer's expectations. The latter problem can be solved to some extent by the known example using know-how reported in the above paper, but it cannot be said that the designer's intention is sufficiently reflected. In particular, it is difficult to extract the designer's decisions at each stage of optimization in advance and reflect them in the computer system, which has been a problem in performing efficient optimization.

SUMMARY OF THE INVENTION An object of the present invention is to provide an optimal design support method that solves the above problems, utilizes the know-how of experienced designers, and enables efficient optimal design.

[Means to solve the problem]

The purpose of the above is to have the computer memorize the details and results of each modification, as well as the history of the modifications that led to the optimal design. Either the computer automatically creates a modification plan based on similar situations, or the designer creates a modification plan, and the computer evaluates it by referring to the above memory to make it a reasonable modification plan. This is achieved by simulating the revised proposal created by the above method.

[Effect]

In the present invention, the designer supports the process of interactive optimization, and the designer sets initial values of input parameters and obtains output using a simulator. Next, the designer compares the output with the target output, modifies the input parameters to obtain the target output, obtains the output using the simulator, and repeats the above steps until the output satisfies the target output. In this flow, each time a designer makes a modification, the designer uses a graphic terminal to store the contents of the modification in a predetermined format. On the other hand, at the stage where a correction is to be made, for example, attention is paid to the output to be brought closer to the target output, and a search is made to see if corrections made under similar circumstances have already been stored by the above operation. Here, we support two types of corrections. One is to apply corrections based on the current correction status based on the contents of the successfully completed corrections among the searched corrections, and to generate a correction proposal for input parameters. Calculator suggests. The other is that a designer proposes a modification, and the computer evaluates the pros and cons of the modification by comparing it with the contents of the searched modification. If it is determined that the proposed amendment is similar to the content of the failed amendment according to the predetermined criteria, a positive evaluation will be made based on the similarity. If you decide that there is.

Make negative evaluations based on similarities. If it is determined that the proposed amendment is not similar to the content of the successful amendment, a negative evaluation will be given.If the proposed amendment is not similar to the content of the unsuccessful amendment, a negative evaluation will be given. In this way, not only can the history of the modification process of the optimal design be managed, but also the modification can be interactively computer-assisted.

Note that the modification contents to be searched are not only the modification contents stored in the ongoing optimal design, but also the modification contents extracted from another optimal design, or the modification contents input separately. .

〔Example〕

Hereinafter, one aspect of the present invention will be explained in detail.

FIG. 2 is a functional block diagram showing an example of the configuration of a device that supports optimal design. The first data storage means 11 stores target outputs, input parameters and simulation outputs.
, a second data storage means 12 for storing parameter modification details and modification process history, a simulation execution means 21,
Extraction means 22 for extracting modification details and modification process history, search means 23 for searching for similar modifications, generation means 24 for generating modification proposals. Evaluation means for evaluating proposed amendments25. N collection means 26 for adding to or changing the revised contents. Input/output means 31 for inputting and outputting data and correction contents. It consists of a control means 41 that controls the storage means and information processing means.

Figure 1 is a flowchart explaining the process flow when supporting optimal design using the apparatus shown in Figure 2.
First, in step 101 of FIG. 1, the input/output means 31 of FIG. 2 inputs the target output into the first data storage means 11. Next, in step 102, the input/output means 31 inputs the input parameters. The initial value of is input into the first data storage means 11. Next, in step 103, the simulation execution means 21 performs a simulation, and the simulation output is stored in the first data storage means 1.
Next, in step 104, the extraction means 22 extracts the modification details and the history of the modification process associated with the current modification of the input parameters, and stores them in the second data storage means 1.
Next, in step 105, the simulation output stored in the first data storage means 11 is compared with the target output, and it is determined whether all the outputs satisfy the target output. If so, finish the optimal design.

If there is an unsatisfactory output, the process moves to step 106. In step 106, the designer uses the input/output means 31 to input a judgment as to whether or not the designer inputs the next modification of the input parameter by himself or not. If you enter it yourself, step 1
Go to step 08, and if you want to automatically generate it with a computer, step 1
Move to 07. In step 107, the search means 23 determines the current situation from the data stored in the first data storage means 11, and searches for similar modification contents among the modification contents stored in the second data storage means 12. Search for.

Generating means 24 based on the similar modification contents and the current situation.
generates a proposed amendment. Note that if the designer is not satisfied with this revised proposal, the editing means 26 adds or changes the revised contents stored in the second data storage means 12. When step 107 is completed, return to step 103,
Repeat step 103 and subsequent steps.

In step 108, the designer inputs a modification plan using the input/output means 31. Further, the search means 23 determines the current situation from the data stored in the first data storage means 11 and searches for similar modification contents among the modification contents stored in the second data storage means 12. . The evaluation means 25 evaluates and gives advice on the modification proposal based on the similar modification contents and the input modification proposal.If the designer is not satisfied with this evaluation, the editing means 26 edits the second data. Additions or changes are made to the modified contents stored in the storage means 12.

Next, in step 109, the designer changes the revised plan using the input/output means 31, if necessary, based on the evaluation and advice.

When step 109 is completed, the process returns to step 103 and repeats step 103 and subsequent steps.

FIG. 3 is a diagram showing an example of input parameters and outputs stored in the first data storage means 11.

For example, when designing an optimal room air conditioner.

The simulator is a performance calculation program P, and the input parameters are the heat exchanger shape (height x width) A, air volume B,
The rotation speed of the compressor is C, and the output is cooling capacity X, power consumption Y, and noise 2. In this case, the optimal design is such that the input parameters of the heat exchanger shape (height x width ) Determine A, air volume B, and compressor rotation speed C. Note that Figure 3 is an example of graphical display on the screen using the input/output means, where the input parameters and output, which are the input/output data of the simulator, are surrounded by an ellipse, the simulation, which is the processing, is surrounded by a rectangle, and both are input as nodes. It is expressed as a directed graph by connecting output relationships with directed arcs.

FIG. 4 is an example of a method of storing input parameters, target output, and output in the first data storage means 11. In FIG. 4, the process of changing the value of each data due to optimization is stored as a set of data name, change order (#), and value. Note that the number 1 (#1) in the change order represents the initial value, and the last number in the change order represents the last situation in the optimization process or the final result of the optimization.

FIG. 5 shows an embodiment of a method for inputting input parameter modification contents. Figure 5(a) is a screen during optimization, and in addition to the screen in Figure 3, the correction process up to that point is graphically expressed.
Furthermore, the current input parameter values and output values are displayed separately. In this example, FIG. 5(b) shows input by selecting data to be corrected and data to be corrected, and displaying the results as a graph.

Focusing on the target noise value ZZ and the current output Z, the first step is to modify the rotation speed C of the compressor for the purpose of lowering the noise Z, and input the modification details in FIG. 5(C). The modifications are as follows:
Contents of Judgment Conditions for Modification and Amount of Modification9 In this example, the condition is that the target noise value ZZ is smaller than the current noise Z, and the modification can be made by lowering the rotation speed C of the compressor. , specifically, rotation speed C-(Z-ZZ)x25
is set as a new rotational speed C. Also, at this stage, the expected effects of the modifications are entered. This allows you to enter the intent of the modification. Next, the input parameters are modified based on the input modification details and the simulation is repeated. FIG. 5(d) shows the results. Enter an evaluation of whether the modifications made at this stage were successful or unsuccessful. Note that this evaluation can also be performed automatically by comparing the above-mentioned expected effect and new output.

FIG. 6 is an example of one method of storing modified contents by the second data storage means 12. The content of the correction is taken from the data name, which is the basis for judgment, the content of the judgment, the content of the correction, the name of the data to be corrected, the expected effect, the actual output, and the evaluation of the correction.

FIG. 6 shows the stored contents of the modification shown in FIG. Note that the subscripts in parentheses in the figure indicate the order in which data is changed.

FIG. 7 shows an example of a method for storing the history of the correction process by the second data storage means 12. Figure 7 (,) is an example of storing the history in chronological order of the names of the correction processes performed.
FIG. 7(b) is an example in which the history is stored by attaching the input (judgment material) and output (correction parameter) of the correction process to the time series of the name of the correction process performed.

FIG. 8 shows an example of a method for searching for similar modification contents from among the modification contents stored in the second data storage means based on the situation when a modification is to be made.
roblea+ Analysis Diagram)
It is.

First, in step 201 of FIG. 2, situation 1 to be corrected is extracted from the first data storage means 11. This situation is expressed as the difference between the current output and the target output, for example, 0.
Next, the process moves to step 202 and all correction contents stored in the second data storage means are processed in step 203.
~Repeat step 205.

In step 203, the situation to extract the situation 2 when the correction was made from the already stored correction contents is, for example, the difference between the output before the correction execution and the target output, or the difference between the output before the correction execution and the correction The difference from the output after execution is 0. Next, in step 204, the similarity between situation 1 and situation 2 is compared in a fixed format, and if they are similar, proceed to step 205, and if they are not similar, proceed to step 202. return. In this comparison of similarities, for example, if the directions (signs) of differences between all outputs and target outputs match, it is determined that they are similar, and if there are differences in direction, it is determined that they are not similar. In step 205, the modified content is registered as a similar modification, and the process returns to step 203.

Figure 9 shows an example of a method for generating a revision proposal.
cm Analysis Diagram). Situation 1 where correction is to be made in step 301 of Figure 9
is extracted from the first data storage means 11. Next, the process moves to step 302, and step 303 is repeated for successful modifications among all the similar modifications searched by the similar modification search means 23. In step 303, the similarity between situation 2 and situation 1 of each similar modification is Quantitatively calculate the degree of This degree is determined by, for example, the difference between the current output and the target output, the difference between the similarity correction output and the target output, which are normalized for each item and then summed. Find the most similar modification.Next, in step 305, a correction is made to the modification amount of the similar modification based on the difference between situation 1 and situation 2.0For example, the difference between the current output and the target output is the similarity modification. If the difference is about half of the difference when , then the amount of correction is halved.

FIG. 10 is an example of a method for displaying revision proposals. FIG. 1O shows a screen, in which the diagonal lines in the graph indicate the materials for determining the modification plan and the targets for modification, and the content of the modification plan is separately displayed to facilitate the designer's understanding.

Figure 11 shows PAD (P
In step 401 of FIG. 11, steps 402 to 408 are repeated for all similar modifications. In step 402, it is determined whether the similar modifications have been successfully completed.

If the process ends successfully, the process moves to step 403, and if it ends in failure, the process moves to step 406. In step 403, the similarity between the amendment of the proposed amendment and the amendment of the similar amendment is compared in a fixed format. If they are similar, the process proceeds to step 404; otherwise, the process proceeds to step 405. In the similarity comparison, for example, if the signs (directions) of the corrections match, it is determined that they are similar, and if they do not match, it is determined that they are not similar. In step 404, the proposed revision is evaluated positively. In step 405, the revision proposal is evaluated negatively and the basis thereof is shown. In step 406, the similarity between the revision of the revision proposal and the revision of similar revisions is compared in a certain format. If they are similar, the process proceeds to step 407. , otherwise the process moves to step 408. In the similarity comparison, for example, if the signs (directions) of the corrections match, it is determined that they are similar, and if they do not match, it is determined that they are not similar. In step 407, the revised proposal is evaluated negatively. In step 408, the revised proposal is tentatively evaluated as positive, and the basis thereof is presented.

FIG. 12 is an example of a method of displaying evaluations for revised proposals. FIG. 12(a) is an example of displaying a positive evaluation in step 404 of FIG. 11, and FIG. 12(b) is an example of displaying a negative evaluation in step 405 of FIG. 12(e) is a display example of a negative evaluation in step 407 of FIG. 11, and FIG. 12(d) is a display example of a tentatively positive evaluation in step 408 of FIG.

FIG. 13 is an example of a method for changing the content of correction. Figure 13 (a
) represents the state before change, and the case where modification process 3 is changed will be taken as an example. FIG. 13(b) is an example of adding judgment materials, which is achieved by changing the graph on the screen and changing the contents. In addition, deletion of judgment materials is done in the same way.
FIG. 3(c) is an example of adding correction parameters, which is achieved by changing the graph on the screen and changing the contents.

Note that deletion of correction parameters is also performed in the same way, as shown in Fig. 13 (
d) is an example in which only the content of the correction is changed, and is realized by only changing the content on the screen.

FIG. 14 shows an example in which the present invention is used for learning an optimal design method. In Fig. 14, the history of the modification process of the optimal design that serves as a model is stored in advance, and the modification process of the model is displayed during the student's modification process, thereby improving the learner's optimal design skills. be.

Although not shown in this embodiment, it is also possible to support optimal design in the case where a plurality of simulators exist and their input parameters and outputs are related, in the same way as in this embodiment.

〔Effect of the invention〕

As explained above, according to this embodiment, the modification process for actual optimal design is memorized using a graphic display means that can be easily changed or added, and at the same time, interactive support for modification is provided by using the graphic display means. can be realized. Therefore, there is an effect that design support such as design history management and design support such as modification advice can be performed simultaneously. It also has the effect of extracting, accumulating, utilizing, and passing on optimization knowledge, which has traditionally been difficult to pass on as know-how.

[Brief explanation of the drawing]

FIG. 1 is a flowchart of an example of a processing procedure that is a feature of the method of this embodiment, FIG. 2 is a functional block diagram showing an example of the configuration of an apparatus for realizing the method of the present invention, and FIG. Fig. 4 is an explanatory diagram of an example of a simulation of an optimal design to be supported; Fig. 4 is an explanatory diagram of input parameters, target output, and a method of storing the output; Fig. 5 is an explanatory diagram of a method of inputting correction contents of input parameters; The figure is an explanatory diagram of the method for storing input parameter modification contents, Fig. 7 is an explanatory diagram of the method for storing the history of the modification process, Fig. 8 is an explanatory diagram of the method of searching for similar modification contents, and Fig. 9 is an explanatory diagram of the method of storing the modification process history. PAD-1 explaining the generation method
Figure 0 is an explanatory diagram of how to display a revision proposal, Figure 11 is a PAD explaining how to evaluate a revision proposal, Figure 12 is an explanatory diagram of how to display an evaluation of a revision proposal, and Figure 13 is a diagram of how to change the content of revisions. Explanatory diagram: FIG. 14 is an explanatory diagram of a method used for learning an optimization technique. It is. 11... First data storage means, 12... Second data storage means, 21... Simulation execution means,
22...Extraction means, 23...Searching means, 24...
Generation means, 25... Evaluation means, 26... Editing means,
31... Input/output means, 41... Control means. Figure 1. Lame-5 with 'j'11 in the third circle.

Claims (10)

[Claims] 1. Modification of input parameters is performed in an optimal design support method to support optimal design by repeating the task of modifying input parameters to a simulator so that the output of the simulator satisfies the target output. In addition, an extraction storage means is provided to extract and sequentially store the correction process for each time, and when correcting input parameters based on the results of one simulation, the relationship between the output of the simulation result and the target output is An optimal design support method, characterized in that a record of a similar modification process is retrieved from the extraction storage means, and input parameters at that time are modified by referring to the retrieved record. 2. 2. The optimal design support method according to claim 1, further comprising the step of graphically displaying the modification contents in the input parameter modification process on a display means in a predetermined format. 3. Claim characterized in that the diagram in the predetermined format is a graph in which at least data for determining input parameter modification, modification processing, and modification parameters are nodes, and input-output relationships between the nodes are arcs. The optimal design support method described in 2. 4. 2. The optimal design support method according to claim 1, wherein the modification contents in the input parameter modification process are diagrammed in a predetermined format and input to the simulator. 5. Claim characterized in that the diagram in the predetermined format is a graph in which at least data for determining input parameter modification, modification processing, and modification parameters are nodes, and input-output relationships between the nodes are arcs. The optimal design support method described in 4. 6. 2. The optimal design support method according to claim 1, wherein the storage of the input parameter modification process includes at least a history chronologically arranging the execution order of the modification process. 7. 2. The optimal design support method according to claim 1, wherein the record of the similar modification process retrieved from the extraction storage means is referred to for automatically creating a modification plan by a computer. 8. 2. The optimal design support method according to claim 1, wherein the record of the similar modification process retrieved from the extraction storage means is referred to in order to evaluate a modification proposal input by a designer. 9. The optimal design according to claim 1, wherein the data for determining input parameter modification in the input parameter modification process is a value of the input parameter at the time of modification and/or a difference between a target output and a current output. How to help. 10. 2. The optimal design support method according to claim 1, wherein when modifying the input parameters, the modification intention of the modification proposer is expressed by storing at least an output expected to be obtained by the modification.