New Adaptive Mechanism for Large Neighborhood Search using Dual Actor-Critic

2026-01-16

AI Summary

This work addresses the limitation of traditional Adaptive Large Neighborhood Search (ALNS) algorithms, which neglect the synergistic interaction between destroy and repair operators, thereby constraining performance improvements. To overcome this, the authors propose DAC-ALNS, a novel algorithm that introduces, for the first time in ALNS, a dual Actor-Critic reinforcement learning framework. This framework explicitly models the destroy and repair processes as separate Markov decision processes to optimize their coordination strategy. Furthermore, graph neural networks (GNNs) are integrated to perform feature aggregation and normalization, enhancing the model’s generalization across varying problem scales. Experimental results demonstrate that DAC-ALNS significantly improves both solution quality and computational efficiency on a range of combinatorial optimization problems, exhibiting strong cross-scale transferability.

Technology Category

• Search and Optimization: Learning to Search
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Machine Learning: Neuro-Symbolic Learning

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Adaptive Large Neighborhood Search (ALNS) is a widely used heuristic method for solving combinatorial optimization problems. ALNS explores the solution space by iteratively using destroy and repair operators with probabilities, which are adjusted by an adaptive mechanism to find optimal solutions. However, the classic ALNS adaptive mechanism does not consider the interaction between destroy and repair operators when selecting them. To overcome this limitation, this study proposes a novel adaptive mechanism. This mechanism enhances the adaptability of the algorithm through a Dual Actor-Critic (DAC) model, which fully considers the fact that the quality of new solutions is jointly determined by the destroy and repair operators. It effectively utilizes the interaction between these operators during the weight adjustment process, greatly improving the adaptability of the ALNS algorithm. In this mechanism, the destroy and repair processes are modeled as independent Markov Decision Processes to guide the selection of operators more accurately. Furthermore, we use Graph Neural Networks to extract key features from problem instances and perform effective aggregation and normalization to enhance the algorithm's transferability to different sizes and characteristics of problems. Through a series of experiments, we demonstrate that the proposed DAC-ALNS algorithm significantly improves solution efficiency and exhibits excellent transferability.

Problem

• Adaptive Large Neighborhood Search
• combinatorial optimization
• destroy and repair operators
• operator interaction
• adaptability

Innovation

• Dual Actor-Critic
• Adaptive Large Neighborhood Search
• Operator Interaction
• Graph Neural Networks
• Transferability