JP3984201B2 - Design support method and design support apparatus - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a design support method and apparatus for visualizing design points including Pareto optimal solutions in multi-objective optimization design, and a program for design support.
[0002]
[Prior art]
In product development that requires the development of high-performance products at a low price, the primary purpose is to improve the performance of the product, and the second purpose is to keep the price low. If you improve the goal of (increase performance) in the desired direction, the second goal will be worse in the undesirable direction (higher price), and vice versa. Lowering the price) often results in the primary goal of degrading in an undesirable direction (lowering performance). In this way, when different purposes cannot be improved simultaneously in a desired direction and there is a conflicting relationship, this relationship is called a “trade-off relationship”. A multi-objective optimization method or a multi-objective programming method is known as a method for simultaneously improving a plurality of purposes in a trade-off relationship.
[0003]
A set of solutions representing the limits of the trade-off relationship is called “Pareto optimal solution”. The Pareto optimal solution has the property that at least one other objective must be altered in order to improve one objective. The Pareto optimal solution is often obtained as a plurality of points, curves, curved surfaces, and the like. One purpose of the multi-objective optimization method or multi-objective programming method is to obtain such a Pareto optimal solution (see, for example, Non-Patent Document 1 below).
[0004]
In the following description, an objective that should be optimized (maximized or minimized) in decision-making is referred to as an “objective function” in accordance with the custom in the multi-objective optimization method or the multi-objective programming method. When an integer of 2 or more is represented by n, an n-dimensional space defined by a coordinate system having n different objective functions as coordinate axes is collectively referred to as an “objective function space”. In this objective function space, there is no superiority or inferiority between each point included in the Pareto optimal solution (for example, a set of points representing a Pareto optimal solution, or arbitrary points on a curve or a curved surface).
[0005]
The multi-objective optimization design method applies a multi-objective optimization method or multi-objective programming method that simultaneously improves a plurality of conflicting objective functions to a design problem, and various methods have been proposed.
[0006]
For example, with the improvement of computer processing power and the development of new methods typified by multi-objective optimization methods that use genetic algorithms (GA), many Pareto optimal solutions can be calculated in a short time. I came.
[0007]
For example, in a multi-objective optimization method using an evolutionary (genetic) algorithm, the weights between objective functions when weighting multiple objective functions into a single objective function are systematically processed during the optimization calculation. And the device which reduces calculation amount is made | formed by changing dynamically (for example, refer the following patent document 1).
[0008]
Also, a multi-objective optimization method has been proposed that can obtain a high-quality semi-optimal solution that is closer to the Pareto optimal solution in a short time by introducing an acceptance probability function (see, for example, Patent Document 2 below).
[0009]
In the multi-objective optimization method when there are constraints, for example, the evaluation value of the objective function at each design point randomly generated in the design space according to the pseudo-Monte Carlo method is calculated, the pareto confirmation and the tolerance of the objective function A method for obtaining a Pareto optimal design solution by filtering by value has also been proposed (see, for example, Patent Document 3 below).
[0010]
Here, if the number of Pareto optimal solutions obtained as a result of multi-objective optimization is several, it is relatively easy for a designer to compare these solutions with each other or with a conventional design example. is there. However, if the number of Pareto optimal solutions becomes enormous, it will be difficult to make manual comparisons manually.
[0011]
For this, by introducing a self-organizing map and performing clustering for a large number of Pareto optimal solutions of 766 discretely obtained in a four-dimensional objective function space, the Pareto surface is converted into a two-dimensional plane. There is a method of mapping and visualizing (for example, see Non-Patent Document 2 below). Although this conventional technique has the advantage of being able to classify and display multidimensional Pareto solutions on a plane, the points in the multidimensional space are mapped onto a two-dimensional plane, so the classified design solution is the target. There is a problem that it is unclear where it is located in the function space or the design variable space.
[0012]
[Patent Document 1]
JP 2002-203228 A
[0013]
[Patent Document 2]
JP 2000-268018 A
[0014]
[Patent Document 3]
JP 2001-52041 A
[0015]
[Non-Patent Document 1]
Hirotaka Nakayama, Tetsuzo Tanino, “Theory and Application of Multi-objective Programming”, Corona, 1994, pp. 22-38 (Section 2.1)
[0016]
[Non-Patent Document 2]
Shigeru Obayashi, "Multi-objective optimization and visualization of Pareto solutions", Proceedings of the Japan Society of Mechanical Engineers Computational Mechanics Lecture, Vol. 14, pp. 699-700, 2001
[0017]
[Problems to be solved by the invention]
When a designer makes a decision in multi-objective optimization design, the criterion for evaluation is mainly the position of the design point in the objective function space. On the other hand, the specific shape, material, mechanism, etc. of the product are determined by design variables. The design point is also expressed as a point in a design variable space defined by a coordinate system having different design variables as coordinate axes. In general, the relationship between design variables and objective functions is a non-linear function with a large number of design variables as arguments, so just looking at the value of the objective function (design point position in the objective function space) The value of (design point position in the design variable space) is unknown. Further, not only the design variables but also the constraint conditions and the state variables cannot grasp these evaluation values in the objective function space.
[0018]
Therefore, the present invention provides a design support method and apparatus capable of effectively visualizing design points of multi-objective optimization design in an objective function space and thereby supporting decision making by a designer, and a program for design support. The purpose is to do.
[0019]
[Means for Solving the Problems]
A design support method according to an aspect of the present invention is a design support method that supports multi-objective optimization design by visualizing design points, and stores a value of an objective function and a design variable corresponding to the design points. A step of classifying the design point based on the value of the design variable, and the value of the objective function in a plane or space defined by a coordinate system having all or part of the objective function as a coordinate axis. A display step of displaying the design point in a display mode corresponding to the classification result of the classification step at a corresponding coordinate position.
[0020]
A design support method according to another aspect of the present invention is a design support method that supports multi-objective optimization design by visualizing design points, and is represented by an objective function value and an inequality corresponding to the design points. Defined by a storage step for storing constraint evaluation values, a classification step for classifying the design points based on the constraint evaluation values, and a coordinate system having all or part of the objective function as coordinate axes And a display step of displaying the design point at a coordinate position corresponding to the value of the objective function in a plane or space in a display mode according to the classification result of the classification step.
[0021]
In addition, a design support method according to another aspect of the present invention is a design support method for supporting multi-objective optimization design by visualizing a design point, and a value of an objective function corresponding to the design point and a reference value related to a design state Stipulated by a storage step for storing evaluation values of state variables indicating classification, a classification step for classifying the design points based on the evaluation values of the state variables, and a coordinate system having all or part of the objective function as coordinate axes A display step of displaying the design point at a coordinate position corresponding to the value of the objective function in a plane or space to be displayed in a display mode according to the classification result of the classification step.
[0022]
According to the present invention, design points can be displayed in an objective function space (plane) by a display mode that can identify classifications made for design variable values, constraint conditions, and state variable evaluation values. As a result, when the designer focuses on a design point in the objective function space, not only the position of the design point (the coordinate position consisting of the value of the objective function) but also the design variables and constraints related to the design point and State variables can be roughly grasped according to the classification. Therefore, according to the present invention, it is possible to support a designer's decision making in multi-objective optimization design.
[0023]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a schematic configuration of a multi-objective optimization design display system according to an embodiment of the present invention. This system can be realized as the same program on a general-purpose computer, for example. The main function of this system is an information display function for supporting multi-objective optimization design by the designer. As shown in FIG. 1, the system includes an input unit 11, a control unit including a storage unit 13 and an arithmetic processing unit 14, and a display unit 12.
[0024]
The input unit 11 is various input devices such as a hard disk device, a CD-ROM device, a DVD device, a memory card reader, or a keyboard, and is used for inputting various information to the system. Information input to this system includes design information (objective functions, design variables, design conditions, etc.) related to multi-objective optimization design of objects (products, etc.), including specific numerical values and formulas. . In addition, instructions from the user such as activation and termination of the multi-objective optimization design display system and function selection are also input through the input unit 11.
[0025]
The display unit 12 is an output device such as a display device or a printer, and displays an information input screen to the system and an objective function space display screen with the objective function as a coordinate axis. As will be described later, design points for multi-objective optimization design are classified and displayed on the objective function display screen.
[0026]
The storage unit 13 includes, for example, a read only memory (ROM) and a random access memory (RAM). The read-only memory stores programs such as a basic input / output system (BIOS) and an initial program loader (IPL). The random access memory stores a program of the present system given from the input unit 11 and is used in a working area of the arithmetic processing unit 14 that executes the program according to a command given from the input unit 11 or the like. Further, it is also used as means for storing design point information and design condition information input from the input unit 11. The arithmetic processing unit 14 includes a central processing unit (CPU).
[0027]
When the system is started, the IPL stored in the read-only memory (ROM) of the storage unit 13 is executed, whereby the operating system (OS) is loaded into the random access memory. The program of the multi-objective optimization design display system operates under the operating system. The program is a series of processes for causing the arithmetic processing unit 14 to calculate the design state of each design point based on the design point information and design condition information stored in the storage unit 13 and to control the output to the display unit 12. Consists of procedures.
[0028]
FIG. 2 is a diagram showing a schematic operation of the multi-objective optimization design display system according to the present embodiment. This operation is to effectively visualize the design points in the multi-objective optimization design in the objective function space. The designer can observe the design points classified and displayed on the display screen of the objective function space at the same time. The design state at each design point can also be grasped as a whole.
[0029]
First, the designer inputs design point information and design condition information through the input unit 11 to the multi-objective optimization design display system (steps S21 and S22). The input design point information and design condition information are stored in a design point information storage area and a design condition information storage area provided in the storage unit 13, respectively.
[0030]
The design point information represents, for each of a plurality of design points, an objective function value, a design variable value, an evaluation value of a constraint condition represented by an inequality, an evaluation value of a state variable, and the like. The design condition information represents an upper limit value, a lower limit value of a design variable, a domain defined by a combination of the upper limit value and the lower limit value, a range of constraint conditions, and the like.
[0031]
Next, the designer sets (inputs) the classification method and setting conditions in the present system through the input unit 11 (step S23).
[0032]
Next, when the execution of the multi-objective optimization design display is instructed by the designer, the arithmetic processing unit 14 reads the design point information and the design condition information stored in the storage unit 13 and gives them to the system in step S23. The design points are classified according to the classification method and the set conditions (step S24). The design points are classified by (1) classifying a design point based on the value of a design variable, (2) classifying the design point based on an evaluation value of a constraint condition, and (3). The classification is based on the evaluation value of the state variable. Further, the design point classification includes classification by arbitrarily combining these classifications (1) to (3). In the classifications (1) to (3), classification is performed by referring to design condition information and comparing with upper and lower limit values and discrete values set in advance.
[0033]
Next, the display unit 12 displays the objective function space on the screen, and displays the design point in a manner in which the classification result in step S24 can be identified in the display screen of the objective function space (step S25). The display unit 12 displays the three-dimensional coordinate axes corresponding to the three objective functions arbitrarily selected by the designer on the display screen of the objective function space in a pseudo three-dimensional manner. Two objective functions may be provided, and an objective function plane defined by a coordinate system having the two objective functions as axes may be displayed. The objective function is selected by inputting the classification method and setting conditions in step S23.
[0034]
Each design point of the multi-objective optimization design may be a Pareto optimal solution or an approximate solution thereof. FIG. 3 is a diagram conceptually showing the Pareto optimal solution. The Pareto optimal solution represents a limit that has the property that at least one other objective must be altered in order to improve one objective. For example, the Pareto optimal solution Prt1 represents a design point that achieves a relatively high performance and represents that it is expensive. The Pareto optimal solution Prt2 represents a design point that achieves a relatively low cost and Indicates low performance. In multi-objective optimization design, it is necessary to improve a plurality of different objective functions (cost and performance in the case of FIG. 3) at the same time, find a Pareto optimal solution, and examine these.
[0035]
As described above, when a designer makes a decision, the criterion for evaluation is mainly the position of a design point in the objective function space, that is, a set of points indicating a Pareto optimal solution or a point on a curve. . The multi-objective optimization design display system according to the present embodiment can classify and display design points corresponding to such Pareto optimal solutions according to the above-described operation. Therefore, when a designer pays attention to a certain design point in the objective function space, not only the position of the design point (that is, the value of the objective function) but also the evaluation of the value of the design variable related to the design point and the constraint condition Values and evaluation values of state variables can be roughly grasped by classification.
[0036]
FIG. 4 is a diagram illustrating an example of a method of classifying design states at design points according to design variable values. Here, for convenience of explanation, it is assumed that the number of design variables is two. Note that even if the number of design variables is three or more, the number of combinations only increases, and there is no essential difference in the classification method itself.
[0037]
Design variable x₁The upper and lower limits of x are x₁ ^UBAnd x₁ ^LBAnd the design variable x₂The upper and lower limits of x are x₂ ^UBAnd x₂ ^LBThe design variable x₁Are classified into the following three types.
[0038]
(1) Match the lower limit (x₁ = x₁ ^LB)
(2) Between lower limit and upper limit (x₁ ^LB <x₁ <x₁ ^UB)
(3) Match the upper limit (x₁ = x₁ ^UB)
Design variable x₂Similarly, the values are classified into three types. Classification considering all design variables corresponds to a combination of two classifications, and the total number of classifications is nine. Each of these classifications is given an identification number (for example, 1 to 9). Generally, when the number of design variables is n (n is a positive integer), the total number of classifications is 3ⁿIt becomes street.
[0039]
Note that the way of classification is not limited to the above, and various modifications can be made. For example, instead of all design variables, some selected design variables may be combined and classified. Further, one design variable may not be classified in three ways as described above, but may be classified only by whether or not the value of the design variable matches the lower limit value, for example. In addition, when information indicating the upper limit value or the lower limit value is not given, the maximum value of the design variable is used instead of the upper limit value, or the minimum value of the design variable is used instead of the lower limit value. May be.
[0040]
FIG. 5 is a diagram illustrating another example of a method of classifying design states at design points. In this classification method, the inequality constraint g₁The classification and state variable z₁Combined with the classification about.
[0041]
Inequality constraint g₁For classification on inequality constraints g₁Are evaluated by a finite number of inequalities. Each of these finite inequalities is an inequality constraint g₁Part of In the example of FIG. 5, the inequality constraint g₁The evaluation value of g₁ ^aLess than (case 1) and g₁Value of g₁ ^aAnd g₁ ^bIt is classified into two inequalities consisting of the following cases (Case 2). Constant g₁ ^aAnd g₁ ^bAre constants. For example, the inequality constraint g₁Is the mass of the product, Case 1 has a light mass, and Case 2 has a mass heavier than Case 1 and smaller than a certain value.
[0042]
In addition, the state variable z₁The value is a discrete value z_1A, Z_1BOr z_1CIt classifies according to which one matches.
[0043]
Needless to say, the present invention is not limited to the classification method shown in FIGS. For example, classification may be performed for one design variable, one inequality constraint, or one state variable. Further, the classifications may be combined by a method other than that shown in FIGS. For example, the classification may be a combination of design variables and inequality constraints, or a combination of design variables and state variables. Moreover, it is good also as a classification | category which made the classification | category hierarchy not only 2 layers but 3 layers or more.
[0044]
Hereinafter, how to display the design points of the multi-objective optimization design based on the result of the classification as shown in FIGS. 4 and 5 will be described. On the objective function space screen displayed on the display unit 12, the design point of the multi-objective optimization design is displayed by applying a display method (display mode) corresponding to the design state classification result. The classification result of the design state is identified by the classification identification number. In addition, there may be a case where there is no design point corresponding to the classification identification number. In this case, the display method may be assigned to all the classification identification numbers, or the design point corresponding to the classification may be assigned. The display method may be assigned by excluding the identification number of the classification that does not exist. Further, if the meaning content of the classification and the identification number are displayed simultaneously by, for example, a legend, it becomes easier to understand.
[0045]
The display method of the classification that can be distinguished from each other includes the shape of the marker that indicates the position of each design point, if the marker is closed, whether or not it is filled, the size of the marker, the type of marker color (red , Green, blue, etc.), color shading or lightness, dot density or size when the inside of a figure is filled with dots, and the like. Moreover, they can also be used in combination.
[0046]
FIG. 6 shows an example of a method for identifying and displaying the classification. In this example, identification number 1 is a black triangle mark, identification number 2 is a white triangle mark, identification number 3 is a black diamond mark, identification number 4 is a white circle mark, identification number 5 is a cross mark, identification number 6 is a white diamond mark, The identification number 7 is a black circle mark, the identification number 8 is a white square mark, and the identification number 9 is a black square mark.
[0047]
FIG. 7 shows another example of a method for identifying and displaying the classification. In this example, the size of the marker and the lightness of the fill are linked according to the size of the value of the design variable or the like. In addition, a method of changing the dot density or dot size of the filling in a pseudo and continuous manner, or changing the color in a pseudo and continuous manner (color contour display) is also included. Such a display method is suitable when classification of design states is performed for a certain design variable, a certain inequality constraint condition, or a certain state variable. Further, the display method of FIG. 7 and the display method of FIG. 6 may be combined.
[0048]
In addition, when classifying design states by design variable values, inequality constraint evaluation values, or combinations of a finite number of sets divided in stages according to state variable evaluation values, The number of colors that can be displayed on the screen by associating up to three variables or conditions with three elements of color (for example, RGB (red, green, blue), HSV (hue, saturation, brightness), etc.)) Alternatively, colors that are equal to or less than the number of identified colors in printing may be used so that the colors change in a pseudo continuous manner.
[0049]
Next, specific design examples will be described. FIG. 8 is a diagram showing a design example of a beam (beam) having an I-shaped cross section. This I-shaped cross-section beam is a cantilever type in which the root end portion is fixed and load loads Fy and Fz act on the tip portion in the downward and lateral directions, respectively. The cross-sectional shape is I-type, the length direction is unloaded, and there is no expansion and contraction.
[0050]
Here, for example, the following three objective functions J₁, J₂, J_ThreeConsider the problem of designing the shape of the beam so that the values of both are minimized (J₁= Mass of beam → minimization), (J₂= Downward displacement δy → Minimization with respect to downward load Fy, (J_Three= Lateral displacement δz with respect to lateral load Fz → minimization).
[0051]
The design variables that define the shape of the beam are the four dimensions x shown in FIG.₁, X₂, X_Three, X_FourAnd In addition, each design variable is designed within a design range defined by an upper limit value and a lower limit value. The Pareto optimal solution for this design problem is obtained using a known multi-objective optimization method. The Pareto optimal solution obtained by this is given to the multi-objective optimized design display system of this embodiment as the design point information described above. Dimension x₁, X₂, X_Three, X_FourInformation such as upper and lower limit values is given to the system as design condition information.
[0052]
FIG. 9 shows a conventional example in which the Pareto optimal solution for the beam design problem is displayed as it is on the objective function space display without being classified. Each coordinate axis J₁, J₂And J_ThreeRespectively correspond to the first, second and third objective functions. The desired design for this design problem is the objective function J₁, J₂And J_ThreeIs that the value of becomes smaller. In this conventional example, since the Pareto optimal solution is displayed without classification, the position of the design point in the objective function space can be grasped, but the design state such as the value of the design variable at each design point is completely unknown. It is.
[0053]
On the other hand, FIG. 10 shows a display example of the Pareto optimal solution by the multi-objective optimization design display system of this embodiment. In this display example, each design point is displayed in the objective function space using the display method shown in FIG. 6 in accordance with the classification result of the design state shown in FIG. From this figure, in the Pareto optimal solution distributed in a curved surface,
(1) Objective function J₁(Mass) value is small, objective function J₂Since the value of (displacement in the down (y) direction) is also small, it is a black circle (identification number 7), so the design variable x₁Matches the upper limit and the design variable x₂Matches the lower limit,
(2) Objective function J₁(Mass) value is small, objective function J₂Since the black triangle mark (identification number 1) is in the region where the value of (down (y) direction displacement) is large, the design variable x₁Matches the lower limit and the design variable x₂Matches the lower limit,
(3) Objective function J₂(Down (y) direction displacement) value is small, objective function J_ThreeSince the black square mark (identification number 9) is in the area where the value of (displacement in the lateral (z) direction) is small, the design variable x₁Design variable x that matches the upper limit value₂Matches the upper limit,
Etc. becomes obvious at a glance.
[0054]
The objective function J₁In the area where the value of (mass) is small, black triangle mark (identification number 1), white circle mark (identification number 4), and black circle mark (identification number 7), design variable x₂It can also be seen that is consistent with the lower limit.
[0055]
FIG. 11 shows another display example of the Pareto optimal solution by the multi-objective optimization design display system of this embodiment. In this example, the display method of each design point in the objective function space is as shown in FIG. 7 according to the state variable value of the minimum natural frequency (primary natural frequency) at the design point, as shown in FIG. The size of the point marker (white circle) and the lightness of the fill are changed in conjunction with each other. From this figure, the objective function J₁As the value of (mass) increases, the objective function J₂(Down (y) direction displacement) and objective function J_ThreeIt can be seen that the value of (displacement in the lateral (z) direction) decreases, the static rigidity improves, and the natural frequency increases. The objective function J₁Conversely, if the value of (mass) exceeds about 1.3 kg, the objective function J₁The designer can easily grasp a complicated design state in which the natural frequency increases as the value of (mass) increases.
[0056]
As described above, according to the embodiment of the present invention, design points can be displayed in the objective function space in a display mode that can identify classifications made on values of related design variables, constraints, state variables, and the like. it can. As a result, when the designer focuses on a design point in the objective function space, not only the position of the design point (that is, the value of the objective function) but also the value of the design variable related to the design point and the constraint condition are evaluated. Values and evaluation values of state variables can be roughly grasped by classification. Therefore, it is possible to support a designer's decision making in multi-objective optimization design.
[0057]
Note that the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the components without departing from the scope of the invention in the implementation stage. In addition, various inventions can be formed by appropriately combining a plurality of components disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, constituent elements over different embodiments may be appropriately combined.
[0058]
【The invention's effect】
ADVANTAGE OF THE INVENTION According to this invention, the design support method and apparatus which can support the designer's decision-making in multi-objective optimization design, and the program for design support can be provided.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a schematic configuration of a multi-objective optimization design display system according to an embodiment of the present invention.
FIG. 2 is a diagram showing a schematic operation of the multi-objective optimization design display system according to the embodiment.
FIG. 3 is a diagram conceptually showing the Pareto optimal solution.
FIG. 4 is a diagram showing an example of a method for classifying design states at design points according to design variable values.
FIG. 5 is a diagram showing another example of a method for classifying design states at design points.
FIG. 6 is a diagram showing an example of a method for identifying and displaying a classification
FIG. 7 is a diagram showing another example of a method for identifying and displaying a classification
FIG. 8 shows a design example of a beam (beam) with an I-shaped cross section
FIG. 9 is a diagram showing a conventional example in which a Pareto optimal solution for a beam design problem is not classified but is displayed as it is on an objective function space display.
FIG. 10 is a diagram showing a display example of a Pareto optimal solution by the multi-objective optimization design display system of the embodiment.
FIG. 11 is a view showing another display example of the Pareto optimal solution by the multi-objective optimization design display system of the embodiment.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 11 ... Input part, 12 ... Display part, 13 ... Memory | storage part, 14 ... Operation processing part, S21 ... Input step of design point information, S22 ... Input step of design condition information, S23 ... Input step of classification method and setting condition, S24: Classification execution step, S25: Classification display step

Claims (8)

In a design support method that supports multi-objective optimization design by visualizing design points,
Storing a value of an objective function having three or more dimensions corresponding to the design point ;
Storing the evaluation value of the constraint condition represented by an inequality in the storage unit ;
An input unit inputs an instruction of the design point classification method;
In accordance with the instruction of the classification method, the arithmetic processing unit classifies the design point based on the evaluation value of the constraint condition, and assigns an identification number for identifying the classification result ;
The display unit pseudo-dimensionally displays at least three coordinate axes corresponding to all or part of the objective function having three or more dimensions;
The display unit displaying the design point in a display mode corresponding to the identification number at a coordinate position corresponding to the value of the objective function in a space defined by the three coordinate axes. Method. The arithmetic processing unit compares the upper limit value of the constraint condition, the lower limit value, or each allowable range of the constraint condition represented by a combination of the upper limit value and the lower limit value with the evaluation value of the constraint condition , and the comparison result The design support method according to claim 1 , wherein the identification number is assigned in accordance with. The storage unit stores evaluation values of the first and second constraint conditions,
The arithmetic control unit, the design point and sorted based on the evaluation value of the first constraint, and classified based classification results of its evaluation value of the second constraint, the classification result of its The design support method according to claim 1 , wherein the identification number is assigned accordingly. In a design support method that supports multi-objective optimization design by visualizing design points,
Storing a value of an objective function having three or more dimensions corresponding to the design point ;
The storage unit stores an evaluation value of a state variable indicating a reference value related to a design state;
An input unit inputs an instruction of the design point classification method;
In accordance with the instruction of the classification method, the arithmetic processing unit classifies the design point based on the evaluation value of the state variable, and assigns an identification number for identifying the classification result ;
The display unit pseudo-dimensionally displays at least three coordinate axes corresponding to all or part of the objective function having three or more dimensions;
The display unit displaying the design point in a display mode corresponding to the identification number at a coordinate position corresponding to the value of the objective function in a space defined by the three coordinate axes. Method. The storage unit stores evaluation values of the first and second state variables,
The arithmetic control unit, the design point and sorted based on the evaluation value of the first state variable is classified based on the classification results of its evaluation value of the second state variable, the classification result of its 5. The design support method according to claim 4 , wherein the identification number is assigned in response. The design point, the design support method according to any one of claims 1 to 5 including Pareto optimal solutions or approximate solutions thereof in multi-objective optimization design. In a design support device that supports multi-objective optimization design by visualizing design points,
Storage means for storing a value of an objective function having three or more dimensions corresponding to the design point and an evaluation value of a constraint condition represented by an inequality;
An input means for inputting an instruction of the design point classification method;
Classifying means for classifying the design point based on the evaluation value of the constraint condition according to an instruction of the classification method, and assigning an identification number for identifying the classification result ;
Coordinates corresponding to the value of the objective function in a space defined by the three coordinate axes are displayed in a pseudo three-dimensional manner at least three coordinate axes corresponding to all or a part of the objective function having three or more dimensions. A design support apparatus comprising: display means for displaying the design point at a position in a display mode corresponding to the identification number . In a design support device that supports multi-objective optimization design by visualizing design points,
Storage means for storing an objective function value having three or more dimensions corresponding to the design point and an evaluation value of a state variable indicating a reference value related to a design state;
An input means for inputting an instruction of the design point classification method;
Classifying means for classifying the design point based on the evaluation value of the state variable according to an instruction of the classification method, and assigning an identification number for identifying the classification result ;
Coordinates corresponding to the value of the objective function in a space defined by the three coordinate axes are displayed in a pseudo three-dimensional manner at least three coordinate axes corresponding to all or a part of the objective function having three or more dimensions. A design support apparatus comprising: display means for displaying the design point at a position in a display mode corresponding to the identification number .

Images