LLM-Driven Evolutionary Generation of Multi-Objective Bayesian Optimization Algorithms
We extend the LLaMEA framework to MOBO, using large language models as mutation and crossover operators within evolutionary strategies to generate complete algorithm implementations, with SMAC hyperparameter optimization integrated into the evolutionary loop. Across nine evolutionary runs we generated approximately 900 algorithms and benchmarked them on twelve synthetic problems (ZDT, DTLZ, WFG) and three real-world engineering problems (RE), using a BoFire qParEGO implementation as a state-of-the-art Bayesian-optimization baseline. On the synthetic suite the strongest generated algorithm attains the highest mean normalized hypervolume (0.971, vs.
- ▪We extend the LLaMEA framework to MOBO, using large language models as mutation and crossover operators within evolutionary strategies to generate complete algorithm implementations, with SMAC hyperparameter optimization integrated into the
- ▪Across nine evolutionary runs we generated approximately 900 algorithms and benchmarked them on twelve synthetic problems (ZDT, DTLZ, WFG) and three real-world engineering problems (RE), using a BoFire qParEGO implementation as a state-of-t
- ▪On the synthetic suite the strongest generated algorithm attains the highest mean normalized hypervolume (0.971, vs.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Neural and Evolutionary Computing arXiv:2607.08791 (cs) [Submitted on 6 Jul 2026] Title:LLM-Driven Evolutionary Generation of Multi-Objective Bayesian Optimization Algorithms Authors:Georgios Laskaris, Reuben Brasher, Niki van Stein, Elena Raponi, Thomas Bäck, Florian Neukart View a PDF of the paper titled LLM-Driven Evolutionary Generation of Multi-Objective Bayesian Optimization Algorithms, by Georgios Laskaris and 5 other authors View PDF HTML (experimental) Abstract:Designing effective multi-objective Bayesian optimization (MOBO) algorithms requires balancing many interdependent design choices whose optimal configuration is problem-dependent and typically demands deep expertise.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
