To path indicator To navigation To quick access To footer with other services
To content
Faculty of mechanical engineering
Studierende sitzt am Arbeitsplatz und schaut auf den Laptop. © Aliona Kardash​/​TU Dortmund

Award-winning research: IMI receives Best Paper Award in Spain

-
in
  • News
  • Forschungsnews
  • Forschungsprojekte
  • FlexTools
  • FOISM
Thomas Boemer auf der AGENTICS in Spanien © Boemer​/​ KIT
Das IMI hat ein neues Forschungspapier auf der International Conference on Agentic and Generative Techniques in Intelligent Computational Systems (AGENTICS) vorgestellt, die vom 22. bis 24. Oktober in Marbella, Spanien stattfand
Big performance from small models: At AGENTICS, IMI presented how specialized prompt augmentation improves AI-assisted optimization. The results show that energy-efficient models can often keep up with the "big ones". This groundbreaking work has now been honored with the Best Paper Award.

IMI presented a new research paper at the International Conference on Agentic and Generative Techniques in Intelligent Computational Systems (AGENTICS), which took place from October 22-24 in Marbella, Spain.

The paper, entitled "Prompt Augmentation for Evolving Heuristics for a Niche Optimization Problem" , investigates the use of Large Language Models (LLMs) for the automated development of heuristics for specialized optimization problems. The focus is on two newly developed approaches: Contextual Evolution of Heuristics (CEoH), which integrates problem-specific information into the prompt design, and Literature-Based CEoH (LitCEoH), which additionally incorporates knowledge from scientific literature. The aim is to enable LLMs to design effective and generalizable heuristics.

The results show that these strategies enable smaller, energy-efficient language models in particular to generate high-quality heuristics - in some cases with better performance than significantly larger models. The work was also awarded the Best Paper Award and thus makes an important contribution to the further development of automated, AI-supported heuristic design in optimization research.

Back