Executable area calculating device

Abstract

To retrieve an executable area while suppressing deterioration in calculation speed.SOLUTION: A reading section 12 reads a retrieval condition file which determines a retrieval condition of an executable area and a calculation model file which determines a calculation model for retrieving an executable area. A setting section 14 sets a retrieval condition based on the retrieval condition file. A calculation section 16 calculates one or more establishment solutions by executing optimization calculation. An output section 18 retrieves an executable area according to the set retrieval condition, and outputs it as a retrieval result file containing the calculated establishment solution.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to an apparatus for calculating a feasible region. [Background technology]

[0002] In development, one of the important issues is to establish a design methodology that can optimize each of multiple conflicting performances. For example, a set-based design method is known in which the range of possible solutions (i.e., the feasible region) is calculated to reduce rework in development, and development proceeds based on that range (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2022-012511 Summary of the Invention [Problem to be solved by the invention]

[0004] In conventional methods, the calculation speed may decrease as the feasible region is explored, so there is room for improvement in the technique for exploring the feasible region.

[0005] An object of the present disclosure is to explore a feasible region while suppressing a decrease in calculation speed. [Means for solving the problem]

[0006] One aspect of the present disclosure is a feasible range calculation device characterized by having a reading unit that reads a search condition file that defines search conditions for a feasible range and a computational model file that defines a computational model for searching the feasible range, a setting unit that sets the search conditions based on the search condition file, a calculation unit that calculates one or more valid solutions by performing an optimization calculation, and an output unit that searches the feasible range in accordance with the set search conditions and outputs a search result file including a group of calculated valid solutions.

[0007] The functions of the computing device may be realized by artificial intelligence (AI). For example, machine learning such as deep learning may be used. Specific techniques include neural networks (for example, convolutional neural networks (CNNs) and recurrent neural networks (RNNs)), autoencoding, backpropagation, ensemble learning, and the like. Of course, the present invention is not limited to these, and any technique related to artificial intelligence may be applied to the above configuration. Effect of the Invention

[0008] According to the present disclosure, it is possible to search for a feasible region while suppressing a decrease in calculation speed. [Brief description of the drawings]

[0009] [Figure 1] FIG. 1 is a block diagram showing a configuration of a calculation device according to an embodiment. [Diagram 2] FIG. 11 is a flowchart showing a flow of processing by the calculation device. [Diagram 3] FIG. 11 is a diagram for explaining a process of searching a feasible region. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] 1 is a block diagram showing the configuration of a calculation device 10 according to an embodiment. The calculation device 10 is an example of a feasible region calculation device, and includes a reading unit 12, a setting unit 14, a calculation unit 16, and an output unit 18.

[0011] The reading unit 12 reads a search condition file and a calculation model file. The search condition file is a file that defines conditions (hereinafter referred to as "search conditions") for searching a feasible region. The feasible region is a region in which each of a plurality of performances (multiple performances) is feasible. For example, when the development target is a structure, the performances are strength, rigidity, weight, vibration, cost, etc. The calculation model file is a file that defines a calculation model for searching a feasible region.

[0012] The setting unit 14 sets search conditions based on the search condition file. The calculation unit 16 executes an optimization calculation to calculate one or more valid solutions. The output unit 18 searches a feasible region according to the set search conditions, and outputs a search result file including the calculated valid solutions.

[0013] Hereinafter, the flow of processing by the calculation device 10 will be described with reference to Fig. 2 and Fig. 3. Fig. 2 shows a flowchart illustrating the flow of the processing. Fig. 3 is a diagram for explaining the processing of searching for a feasible region.

[0014] First, the reading unit 12 reads the search condition file and the calculation model file (S1). The setting unit 14 sets search conditions based on the set condition file (S2). The calculation unit 16 executes an optimization calculation to calculate one or more valid solutions (S3). Next, the output unit 18 searches a feasible region according to the set search conditions, and outputs a search result including the calculated valid solutions (S4 to S13).

[0015] The process by the output unit 18 will be described below with reference to Fig. 2 and Fig. 3. As an example, a feasible region is searched for two design variables. Variables X1 and X2 shown in Fig. 3 are design variables used to calculate evaluation indexes such as performance. A two-dimensional map is formed by the variables X1 and X2.

[0016] 3(a), the output unit 18 sets an initial value S, a range A centered on the initial value S, a reduction rate of the range A, and a loop count N (upper limit) (S4). The output unit 18 randomly samples the range A to extract valid solutions that satisfy the constraint conditions (S5).

[0017] If a new valid solution is extracted (S6, Yes), the processes from step S7 onwards are executed. For example, if a new valid solution group F is extracted, as shown in FIG. 3(b), the output unit 18 calculates the center of gravity G of the new valid solution group F (S7). The output unit 18 also calculates the distance between each point of the valid solution group F and the center of gravity G (S8). Furthermore, the output unit 18 extracts the valid solution that is the farthest from the center of gravity G (hereinafter referred to as the "maximum distance point f") from among the valid solutions included in the valid solution group F (S9).

[0018] If the number of processes (loop count) after step S4 is less than the upper limit N (S10, No), the process returns to step S4. In this case, as shown in FIG. 3(c), the output unit 18 sets the maximum distance point f as the initial value S of the next process (next loop) and executes the processes after step S4.

[0019] If the number of loops is equal to or greater than the upper limit N (S10, Yes), the output unit 18 outputs a search result file including the extracted established solutions (S13).

[0020] In step S5, if a valid solution that satisfies the constraint conditions is not extracted (S6, No), and if the number of loops is less than N (S11, No), the output unit 18 reduces the range A (S12). The process proceeds to step S4, and the output unit 18 executes the processes from step S4 onwards.

[0021] If the loop count is N or more (S11, Yes), the output unit 18 outputs a search result file including the extracted established solution group (S13).

[0022] As described above, the initial value S in the n+1th loop is set using only the valid solutions extracted in the nth loop. This makes it possible to suppress a decrease in the calculation speed even in the later stages of the search, regardless of the number of valid solutions accumulated as the search progresses. In other words, when performing distance calculations on all points included in the feasible region extracted in the past, the number of points included in the feasible region increases as the search progresses, so the number of calculations increases by the amount of the increase, and the calculation speed decreases. In this embodiment, the initial value S in the n+1th loop is set using only the valid solutions extracted in the nth loop, so that a decrease in the calculation speed can be suppressed even when the search progresses.

[0023] The calculation device 10 can be realized by using hardware resources such as a processor and an electronic circuit, and a device such as a memory may be used as necessary in the realization. The calculation device 10 may also be realized by, for example, a computer. That is, the calculation device 10 may be realized in whole or in part by cooperation between hardware resources such as a CPU (Central Processing Unit) and a memory provided in a computer and software (program) that specifies the operation of the CPU and the like. The program is stored in the storage device of the calculation device 10 via a recording medium such as a CD or a DVD, or via a communication path such as a network. As another example, the calculation device 10 may be realized by a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), or an FPGA (Field Programmable Gate Array), or the like. Of course, a GPU (Graphics Processing Unit), or the like, may be used. The functions of the calculation device 10 may be realized by a single device or by multiple devices. [Explanation of symbols]

[0024] 10 calculation device, 12 reading unit, 14 setting unit, 16 calculation unit, 18 output unit.

Claims

[Claim 1] a reading unit that reads a search condition file that defines search conditions for a feasible region and a computation model file that defines a computation model for searching the feasible region; A setting unit that sets search conditions based on the search condition file; A calculation unit that calculates one or more valid solutions by executing an optimization calculation; an output unit that searches a feasible region according to the set search conditions and outputs a search result file including a calculated established solution group; A feasible region calculation device comprising:

Images