Search papers, labs, and topics across Lattice.
This paper introduces AutoPref, an innovative framework that automates the discovery of task-specific preference objectives for neural combinatorial optimization (NCO), addressing the inefficiencies of existing manually specified approaches. By factorizing the preference objective into a pairwise loss program and a set-aware weighting program, AutoPref enables a more nuanced and scalable approach to learning from solution quality. The framework significantly outperforms traditional hand-designed baselines across various combinatorial optimization problems, showcasing its effectiveness and adaptability in real-world applications.
Automated preference-objective discovery can dramatically enhance the efficiency and performance of neural combinatorial optimization, outperforming traditional methods.
Combinatorial optimization problems (COPs) underpin many real-world decisions, but their exponentially large search spaces make high-quality solutions costly to obtain. Neural combinatorial optimization (NCO) learns fast construction policies, typically with reinforcement learning (RL), while preference-based NCO improves sample efficiency by learning from relative solution quality. However, existing preference objectives combine two distinct design choices in manually specified, one-size-fits-all formulations: what learning signal to extract from each solution pair and how to weight each pair relative to the sampled set. We present AutoPref, the first LLM-guided framework for automated preference-objective discovery in NCO. AutoPref factorizes the objective into a pairwise loss program, which defines the learning signal, and a set-aware weighting program, which determines each pair's relative contribution. Their composition forms a unified programmatic objective space containing existing preference objectives as special cases. To make its search tractable, we introduce a staged conditional search strategy with behavioral gates that filter inadmissible programs before short-horizon training and evaluation. Across TSP, CVRP, FFSP, and JSSP, AutoPref consistently outperforms strong hand-designed baselines across problem scales, demonstrating the benefits and scalability of automated objective discovery for NCO.
