AdaEva: Accelerating LLM-Driven Algorithm Design with Adaptive Partial Evaluation

2026-10-02

AI Summary

This study addresses the prohibitive computational cost of candidate evaluation in large language model (LLM)-driven algorithm design by proposing AdaEva, an adaptive partial evaluation framework. Requiring no prior knowledge, AdaEva integrates successive halving, statistical racing, and LLM techniques to implement progressive subset evaluation and early elimination mechanisms. This approach dynamically optimizes evaluation budget allocation while preserving the underlying pipeline. Experimental results demonstrate that AdaEva significantly enhances algorithm search efficiency, anytime performance, and cross-task generalization capabilities, thereby establishing a new paradigm for efficient automated algorithm design.

Abstract

Large Language Models (LLMs) are increasingly used for automated algorithm design. However the computational cost of evaluating the generated algorithms can be excessive. We consider the common LLM-driven automated algorithm design (LLM4AD) setting in which a candidate algorithm is evaluated by aggregating its performance over a shared set of training instances. This instance-wise structure raises a natural question: must every candidate be evaluated on the entire instance set before deciding whether it remains competitive? Taking inspiration from algorithm configuration, we introduce AdaEva, a drop-in adaptive partial-evaluation framework that progressively evaluates candidates on larger subsets of the same instance pool and eliminates unpromising candidates as evidence accumulates. Importantly, AdaEva leaves the underlying LLM4AD procedure and per-instance evaluator unchanged and requires no prior knowledge about instance difficulty. We instantiate this idea using successive halving (AdaEva-S) and statistical racing (AdaEva-R), and evaluate both mechanisms across three representative LLM4AD frameworks, multiple LLM backbones, and optimization domains spanning combinatorial and continuous black-box optimization. Under matched evaluation budgets, AdaEva more reliably balances evaluation effort across candidates than fixed partial-evaluation strategies, yielding strong search efficiency and anytime performance together with improved held-out generalization across the evaluated settings.

Problem

• Large Language Models
• Automated Algorithm Design
• Computational Cost
• Partial Evaluation
• Algorithm Configuration

Innovation

• Adaptive Partial Evaluation
• LLM-driven Algorithm Design
• Successive Halving
• Statistical Racing
• Search Efficiency