optimization
combustion

Power-mean objective: Japan 2026 symposium paper

Figures from the full paper on a power-mean objective for multi-dataset combustion mechanism optimization

13 images · 1 min read

Explore in Studio Jump to gallery

Power-mean objective: Japan 2026 symposium paper

datasheet // japan-2026-paper

images: 13
videos: 0
formats: png
text: 60 words · 1 min read
studio: /atlas?ch=S13&ex=4
topics: toy landscape ×2

optimizer comparison ×1

dominance ×3

ignition delay ×1

ignition delay full ×3

heatmaps ×3

Datasheet

Title (accepted abstract): A Reference-Normalized Power-Mean Objective (PowerMeanWMRE) for Multi-Dataset Combustion Mechanism Optimization
Authors (abstract): Krunal Panchal, Krithika Narayanaswamy (IIT Madras)
Venue: symposium in Fukuoka, 4-6 November 2026 (Fukuoka International Congress Center), as recorded in the disclosure policy
Fuels shown: n-heptane, methyl butanoate, 1,3-butadiene
Media in the intake: 15 live figures; supplementary figures, filmstrips and optimizer animations are draft

Gallery

3D toy objective landscape: WMRE, fixed p=0.05 and annealed p, with the search path — 3D toy objective landscape: WMRE, fixed p = 0.05 and annealed p, with the search path Explain Optimization

3D toy objective landscape, search from a poor starting point — 3D toy objective landscape, search from a poor starting point Explain Optimization

Gradient descent, GA and CMA-ES on the toy landscape: WMRE, fixed p and annealed p — Gradient descent, GA and CMA-ES on the toy landscape: WMRE, fixed p and annealed p Explain Optimization

Classical-WMRE dominance share across the three fuels — Classical-WMRE dominance share across the three fuels Explain Optimization

Classical-WMRE dominance, n-heptane — Classical-WMRE dominance, n-heptane Explain Optimization

Classical-WMRE dominance, methyl butanoate — Classical-WMRE dominance, methyl butanoate Explain Optimization

Representative ignition-delay fits: nominal mechanism vs best-tuned powerMeanWMRE — Representative ignition-delay fits: nominal mechanism vs best-tuned powerMeanWMRE Explain Optimization

Every n-heptane condition: nominal mechanism vs best-tuned powerMeanWMRE — Every n-heptane condition: nominal mechanism vs best-tuned powerMeanWMRE Explain Optimization

Every methyl-butanoate condition: nominal mechanism vs best-tuned powerMeanWMRE — Every methyl-butanoate condition: nominal mechanism vs best-tuned powerMeanWMRE Explain Optimization

Every 1,3-butadiene condition: nominal mechanism vs best-tuned powerMeanWMRE — Every 1,3-butadiene condition: nominal mechanism vs best-tuned powerMeanWMRE Explain Optimization

Hyperparameter sweep heatmap, n-heptane — Hyperparameter sweep heatmap, n-heptane Explain Optimization

Hyperparameter sweep heatmap, methyl butanoate — Hyperparameter sweep heatmap, methyl butanoate Explain Optimization

Hyperparameter sweep heatmap, 1,3-butadiene — Hyperparameter sweep heatmap, 1,3-butadiene Explain Optimization