Parallel Processing Designing Device for High-Dimensional Optimization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for designing structural bodies, such as vehicle bodies, face challenges in optimizing multiple contradictory performances like strength, rigidity, and weight reduction, especially in high-dimensional design spaces where efficient searching becomes difficult due to exponential increases in computing costs.

Innovation Solution

A parallel processing designing device and method that eliminate design variables with low contribution to computation, using an acquisition function and a penalty function to efficiently search for executable regions, and employing Gaussian process regression to model probability distributions and limit search regions based on Lipschitz continuity, allowing for simultaneous observation point acquisition.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If the number of design variables is increased to optimize multiple performances, then the search space increases exponentially, but efficient searching becomes difficult

Engineering Contradiction:
Improveoptimization capabilityVSAvoidsearch space complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent extracts and eliminates design variables with low contribution to the objective function by calculating sensitivity coefficients. This reduces the dimensionality of the search space while maintaining the optimization capability for multiple performances, directly addressing the exponential growth problem in high-dimensional spaces

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent changes the parameter of design variable order (dimensionality) by dynamically reducing the number of variables based on sensitivity analysis. This transforms the search space from high-dimensional exponential complexity to lower-dimensional manageable complexity while preserving optimization effectiveness

Inventive Principle:
Principle #35Parameter changes

2Adaptability or versatility

If the order of design variables exceeds a predetermined order, then more design aspects can be covered, but computing costs increase exponentially

Engineering Contradiction:
Improvedesign coverageVSAvoidcomputing cost
Core Design Contradiction:
Adaptability or versatilityVSUse of energy by stationary object

Solution Approach 1:

The patent removes design variables with low sensitivity coefficients from the optimization problem. This extraction reduces the number of variables (order) while maintaining coverage of critical design aspects, thereby reducing computing costs exponentially without sacrificing essential design coverage

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies partial action by focusing computational resources only on the most influential design variables identified through sensitivity analysis. Instead of exhaustively searching all high-dimensional spaces, the method concentrates on a reduced set of critical variables, achieving efficient computing with adequate design coverage

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3961470A1Parallel processing designing device and parallel processing designing method
Publication Date: 2022.03.02 TOYOTA JIDOSHA KK
  • EP3961470A1 patent drawingFigure 1
  • EP3961470A1 patent drawingFigure 2
  • EP3961470A1 patent drawingFigure 3

AI summary

A parallel processing designing device includes: an observation point computing section that, in a case in which an order of plural design variables exceeds a predetermined order, eliminates a design variable that has a low contribution to designing, and for each of plural design variables that are less than or equal to the predetermined order, computes plural observation points for searching for a region in which a performance relating to the design variable is executable, by using an acquisition function and a penalty function; a probability distribution computing section that, for each of the plural performances, computes a probability distribution of the performance being executable at the computed plural observation points; and a multiple performance executable region outputting section that outputs, as a multiple performance executable region, an infinite product of the probability distributions that are respectively computed for the plural performances.