Research map / Learning theory and algorithms / Combinatorial and search optimization
DeepACO: Neural-enhanced Ant Systems for Combinatorial Optimization
Haoran Ye, Jiarui Wang, Zhiguang Cao, Helan Liang, Yong Li
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add neural-guided perturbation instead of vanilla LS with more iterations (ablation) | DeepACO · TSP500 | tour length: 16.90 -> 16.86 | Improved |
| use neural-guided instead of random perturbation (ablation) | DeepACO · TSP1000 | tour length: 23.90 -> 23.85 | Improved |
| train with LS term (W=9) instead of W=0 (ablation) | DeepACO · TSP500 | tour length: 17.05 -> 16.86 | Improved |
| train with LS term (W=9) instead of W=0 (ablation) | DeepACO · TSP1000 | tour length: 24.23 -> 23.85 | Improved |
| replace expert heuristic measures with learned ones (no NLS) | Ant System/EAS/MMAS · eight COPs | best-so-far objective | Improved |
| apply DeepACO NLS vs plain ACO | ACO · TSP500 | gap to LKH-3 (%): 6.16 -> 1.84 | Improved |
| add multihead decoder, top-k entropy loss, or imitation loss (ablation) | DeepACO · TSP20/50/100 | best-so-far objective | Improved |
| use ACO evolution with pheromone update instead of pure sampling (ablation) | DeepACO · routing | solution quality | Improved |
Closest papers by meaning
- ConfigX: Modular Configuration for Evolutionary Algorithms via Multitask Reinforcement Learning (AAAI 2025 · similarity 0.83)
- BOPO: Neural Combinatorial Optimization via Best-anchored and Objective-guided Preference Optimization (ICML 2025 · similarity 0.83)
- PARCO: Parallel AutoRegressive Models for Multi-Agent Combinatorial Optimization (NeurIPS 2025 · similarity 0.83)
- Graph-Supported Dynamic Algorithm Configuration for Multi-Objective Combinatorial Optimization (ICML 2025 · similarity 0.83)
- DesignX: Human-Competitive Algorithm Designer for Black-Box Optimization (NeurIPS 2025 · similarity 0.83)
- MetaBox-v2: A Unified Benchmark Platform for Meta-Black-Box Optimization (NeurIPS 2025 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
