Multicriteria Optimization Visualization for Vehicle Control Units
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multicriteria optimization problems in vehicle control units, such as the NOx-soot-fuel consumption compromise, often result in non-unique solutions and require complex visualization and analysis to select the best compromise, with existing methods being limited in their ability to provide detailed and meaningful evaluations.
Innovation Solution
A method that displays the set of optimal solutions in a two- or three-dimensional model space and variation space, interactively linking the two to allow for graphical analysis of target functions and variation variables, enabling the selection of the best compromise by visualizing the Pareto front and model confidence intervals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If linearly weighted summation of target functions is used to solve multicriteria optimization, then the optimization can be performed using conventional approaches, but the meaningfulness of the optimization depends greatly on the selection of weighting factors which cannot be reliably defined in advance
Solution Approach 1:
The patent creates a visual copy or representation of the multidimensional Pareto front in a two-dimensional diagram format that preserves the essential information about target function correlations. This visual copy allows analysts to examine the optimization results without needing to interpret complex weighting factors, as the graphical representation directly shows the relationships between targets and variation variables.
Solution Approach 2:
The patent replaces the mechanical/mathematical system of weighted factor selection with a visual-information system. Instead of relying on numerical weighting and conventional optimization algorithms, the invention uses graphical display methods where the Pareto front is visually represented, allowing direct observation of target correlations and compromise solutions without mathematical transformation.
2Power
If conventional optimization methods are used to reduce computing power requirements, then optimization can be performed efficiently, but the ability to detect correlations of individual target functions and variation variables in the found solution is limited
Solution Approach 1:
The patent projects the multidimensional optimization solution into a two-dimensional graphical representation that preserves correlation information. By displaying the Pareto front in a visual format with appropriate axes and scaling, the invention enables detection of correlations between target functions and variation variables that would be difficult to perceive in raw numerical data or high-dimensional space.
Solution Approach 2:
The patent uses visual differentiation (analogous to color changes) to highlight different aspects of the optimization solution. The graphical display employs visual cues such as positioning, scaling, and potentially visual markers to distinguish between different target functions, variation variables, and their interrelationships, making correlation detection intuitive.
3Reliability
If the Pareto front is displayed in high-dimensional space, then all optimal solutions are represented, but simple yet meaningful evaluation and selection of a specific compromise becomes difficult
Solution Approach 1:
The patent extracts the essential information from the high-dimensional Pareto front and presents it in a simplified two-dimensional graphical format. By selecting and displaying the most relevant correlations and relationships in a reduced-dimensional visual representation, the invention enables meaningful evaluation and selection of compromise solutions while preserving the integrity of the underlying optimization data.
Data Source
AI summary
Solving a multidimensional multicriteria optimization problem is difficult because the correlations and dependencies between solutions, target functions, and variation variables can be detected only with difficulty. In order to facilitate this, it is proposed that a model space (1) and a variation space (2) are displayed simultaneously and in an interactively linked fashion.