Skip to content
Book Open access

LLM Policy Induction for Heuristic Search Control: A Trace-Driven ALNS Framework

Jul 2026 · Annual Conference on Genetic and Evolutionary Computation · 0 citations · 30 references
Computer Science

TL;DR

Experimental results demonstrate that the proposed framework consistently outperforms classical ALNS baselines and several competitive metaheuristic methods in terms of solution quality and convergence speed.

Abstract

Nowadays, large-scale combinatorial optimization problems such as Job Shop Scheduling (JSP), Flexible Job Shop Scheduling (FJSP), and Flow Shop Scheduling (FSP) are widely encountered in manufacturing, logistics, and service systems. Despite substantial progress in metaheuristic frameworks, the design of effective search control policies, including neighborhood selection, move size, acceptance criteria, and diversification strategies, remains heavily reliant on handcrafted rules and domain expertise. In this paper, we propose a trace-driven LLM-controlled Adaptive Large Neighborhood Search (ALNS) framework for scheduling problems, in which large language models (LLMs) are employed exclusively during a policy induction phase to learn high-level search control strategies from historical optimization traces. Rather than evaluating candidate solutions or directly manipulating objective functions, the LLM infers symbolic control policies that map summarized search-state features to ALNS control actions. The induced policy is then integrated into a classical ALNS solver and executed without further LLM involvement. We evaluate the proposed approach on standard benchmark instances of JSP, FJSP, and FSP. Experimental results demonstrate that the proposed framework consistently outperforms classical ALNS baselines and several competitive metaheuristic methods in terms of solution quality and convergence speed.

Read PDF

Similar papers

2026

Feedback-Driven Population Self-Evolution Framework for Dispatching Rule Generation in Dynamic Job Shop via Knowledge Distillation

The dynamic job shop scheduling problem (DJSSP) is critical for optimizing production efficiency in intelligent manufacturing systems under dynamic constraints. Traditional approaches, including heuristic dispatching rules (HDRs) and evolutionary hyper-heuristics, often struggle to generalize across dynamic and unseen scheduling scenarios. The advent of large language models (LLMs) offers new opportunities for scheduling optimization, yet their deployment in real-world manufacturing is constrained by data privacy concerns that limit the use of online LLM services and by the restricted inference capability of local small-scale models. To address these challenges, this paper proposes a novel feedback-driven self-evolution framework employing dual-stage knowledge distillation. The framework utilizes a feedback-driven refinement mechanism to distill expert knowledge from advanced LLMs, which is then used to fine-tune the Qwen2.5-Coder-7B model into ModouGPT, a specialized model for generating HDRs. Experimental results indicate that our approach performs better than traditional methods and five state-of-the-art LLM-based evolutionary frameworks, particularly in dynamic and unseen scheduling scenarios. ModouGPT demonstrates competitive performance with online LLMs in complex dynamic environments, while ensuring data privacy. This study provides manufacturing enterprises, particularly small and medium-sized enterprises, with a deployable scheduling system that balances intelligence, privacy, and computational efficiency in dynamic job shop environments. Note to Practitioners—DJSSP involves constant disruptions such as machine breakdowns and dynamic order arrivals. Deploying artificial intelligence (AI)-based scheduling in real manufacturing environments remains difficult: cloud-based LLM services risk exposing sensitive production data, while large-scale local models demand prohibitive hardware investments, placing both options out of reach for most enterprises. This paper presents a practical solution. The proposed framework distills scheduling knowledge from a powerful online teacher model into a compact 7B local model, ModouGPT, deployable on a single consumer-grade graphics processing unit (GPU). All scheduling decisions are made entirely on-premises, with no data leaving the factory floor. Practitioners do not need domain expertise to design HDRs. The system automatically generates HDRs tailored to the specific production environment, covering job processing characteristics and dynamic disturbances. An offline HDR library can be built in advance and updated periodically without interrupting production, and new HDRs can be synthesized in under one minute during live operations. This makes the system immediately actionable for engineers seeking to upgrade scheduling capabilities without relying on external LLM services or specialized algorithm designers.

Jin Huang, Zhengqi Shi, Qihao Liu et al. · 0 citations
Preprint Jul 2026

LLM-Guided Evolutionary Search for Constraint Model Reformulation to Improve Solver Efficiency

Combinatorial problems appear in numerous industrial applications. A common approach is to formulate these problems as declarative constraint models that can subsequently be compiled to and solved by a range of back-end solvers. Recent work shows that Large Language Models (LLMs) can produce correct models from natural language, but even a correct model can be expensive to solve because performance remains sensitive to modelling choices. In this work, we investigate whether LLMs can automate performance-oriented model reformulation. Inspired by Automatic Heuristic Design (AHD), we use an evolutionary framework in which an LLM proposes candidate reformulations that are verified and benchmarked against the user-defined baseline model. We compare AHD-adapted search strategies that control which prior attempts, instructions, and measured feedback enter each prompt. Existing retention strategies prioritize recency or performance, but do not explicitly diversify the context. To cover this gap, we introduce Profile-Diverse Retention (PDR), which applies Maximal Marginal Relevance (MMR) to instance-level runtime vectors to retain behaviourally diverse attempts. We systematically evaluate the strategies on eight CSPLib problems using validation-based final model selection. The results show that: (i) iterative reformulation can produce substantial held-out speedups; (ii) strategies that keep the retained context diverse outperform those that retain only recent or the fastest attempts; and (iii) validation-based selection improves the held-out speedup of every strategy.

Kostis Michailidis, Dimos Tsouros, Nguyen Dang et al. · 0 citations

Preference-Guided Multi-Policy Optimization for Flexible Job Shop Scheduling

PGMPO is proposed, a novel learning framework consisting of a simple but effective multi-policy modeling approach that allows a single network to represent multiple decision-makers, and a preference-driven model optimization method that effectively guides policies to learn diverse and specialized problem-solving strategies without the need for explicit reward functions.

Inguk Choi, Woo-Jin Shin, Sang-Hyun Cho et al. · 0 citations
Open access Jul 2026

LLM-MetaAS: A Semantic-Statistical Policy Routing Framework for AutoML Execution Strategy Selection

LLM-MetaAS is proposed, a semantic-statistical framework for AutoML execution policy selection that improves overall AutoML performance and selects policies closer to the oracle than fixed strategies, random selection, and the native Auto-sklearn 2.0 selector.

Zhihuan Peng, Pincheng Liu, Yong Li et al. · 0 citations