DGA2D evolves LLM-generated algorithm pipelines as directed walks in an operator graph with first-order transition credit, outperforming LLM-based heuristic design baselines on 12 combinatorial optimization benchmarks.
Procedia Computer Science , volume=
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
DGA$_2$D: Directed Graph-Guided Automated Algorithm Design with Large Language Models
DGA2D evolves LLM-generated algorithm pipelines as directed walks in an operator graph with first-order transition credit, outperforming LLM-based heuristic design baselines on 12 combinatorial optimization benchmarks.