aiqbench analysis
MuEvo represents a significant advancement in automated heuristic design by leveraging large language models to evolve entire ensembles rather than isolated components. This approach addresses critical limitations in current methods where early selection biases or ignored inter-component dependencies hinder optimization performance across complex combinatorial domains. The framework demonstrates that coordinating multiple interacting algorithmic parts via LLM-driven co-evolution yields consistent improvements over both human-designed systems and existing single-heuristic extensions.
MuEvo: LLM-Driven Evolution of Multi-Heuristic Ensemble
arXiv:2608.03636v1 Announce Type: cross Abstract: Large language model-based automated heuristic design (LLM-AHD) has shown strong potential in discovering effective heuristics for combinatorial optimization problems. However, existing methods primarily optimize a single heuristic, whereas practical optimization frameworks often rely on multiple interacting components. Directly extending single-heuristic methods is challenging because early component selection can overlook components with late potential, while independent evolution ignores inter-component dependencies. We propose MuEvo, an LLM-driven framework for evolving heuristic ensembles under ensemble-level feedback. MuEvo combines Dynamic Component Management, which uses short-budget probing and a reversible lifecycle to revise component priorities throughout the search, with LLM-Driven Co-Evolution, which coordinates component populations through Multi-Ensemble Evaluation, Cross-Component Information Sharing, Relation-Guided Pair Evolution, and Adaptive Budget Allocation. We evaluate MuEvo on selection hyper-heuristics and componentized ant colony optimization across four combinatorial optimization domains. Results show that MuEvo consistently improves human-designed frameworks and outperforms representative multi-component extensions of state-of-the-art LLM-AHD methods, demonstrating its effectiveness across both controller-mediated heuristic pools and functionally differentiated algorithmic components.
Abstract:Large language model-based automated heuristic design (LLM-AHD) has shown strong potential in discovering effective heuristics for combinatorial optimization problems. However, existing methods primarily optimize a single heuristic, whereas practical optimization frameworks often rely on multiple interacting components. Directly extending single-heuristic methods is challenging because early component selection can overlook components with late potential, while independent evolution ignores inter-component dependencies. We propose MuEvo, an LLM-driven framework for evolving heuristic ensembles under ensemble-level feedback. MuEvo combines Dynamic Component Management, which uses short-budget probing and a reversible lifecycle to revise component priorities throughout the search, with LLM-Driven Co-Evolution, which coordinates component populations through Multi-Ensemble Evaluation, Cross-Component Information Sharing, Relation-Guided Pair Evolution, and Adaptive Budget Allocation. We evaluate MuEvo on selection hyper-heuristics and componentized ant colony optimization across four combinatorial optimization domains. Results show that MuEvo consistently improves human-designed frameworks and outperforms representative multi-component extensions of state-of-the-art LLM-AHD methods, demonstrating its effectiveness across both controller-mediated heuristic pools and functionally differentiated algorithmic components.
Neural and Evolutionary Computing (cs.NE); Artificial Intelligence (cs.AI)
arXiv:2608.03636 [cs.NE]
(or arXiv:2608.03636v1 [cs.NE] for this version)
https://doi.org/10.48550/arXiv.2608.03636 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 257 words
Related coverage
ReVEL: Multi-Turn Reflective LLM-Guided Heuristic Evolution via Structured Performance Feedback
arXiv · Sep 1, 2026
Neuro-Evolved Heuristics for Variable Gapped Common Subsequence Identification
arXiv · Aug 5, 2026
Medical Heuristic Learning: An LLM-Driven Framework for Interpretable and Auditable Clinical Decision Rules
arXiv · Jun 17, 2026
Medical Heuristic Learning: An LLM-Driven Framework for Interpretable and Auditable Clinical Decision Rules
arXiv · Jun 16, 2026