Research map / Learning theory and algorithms / Combinatorial and search optimization
Neural Lyapunov Control for Discrete-Time Systems
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use DITL instead of LQR | Inverted Pendulum · Inverted Pendulum | ROA: 14 -> 61 ± 31 | Helped |
| Use DITL instead of SOS | Path Tracking · Path Tracking | ROA: 1.8 -> 9 ± 3.5 | Helped |
| Use DITL instead of the baselines | Cartpole · Cartpole | success rate | Helped |
| Use DITL instead of the baselines | PVTOL · PVTOL | success rate | Helped |
| Replace PGD counterexample generation with MILP during learning (ablation) | DITL · Inverted Pendulum | ROA: 61 ± 31 -> 42 | Hurt |
| Replace PGD counterexample generation with MILP during learning (ablation) | DITL · Path Tracking | ROA: 9 ± 3.5 -> 1.1 | Hurt |
| Replace MILP components with dReal (ablation) | DITL · Inverted Pendulum | ROA: 61 ± 31 -> 57 ± 24 | Hurt |
| Replace MILP components with dReal (ablation) | DITL (RL) · Path Tracking | success rate: 100% -> 0% | Hurt |
| Replace MILP components with dReal (ablation) | DITL (LQR) · Path Tracking | ROA: 8 ± 3 -> 4 ± 4 | Hurt |
| Replace MILP components with dReal (ablation) | DITL · Cartpole | success rate: 100% -> 0% | Hurt |
| Replace MILP components with dReal (ablation) | DITL · PVTOL | success rate: 100% -> 0% | Hurt |
| Initialize DITL with RL rather than LQR (ablation) | DITL · Path Tracking | ROA: 8 ± 3 -> 9 ± 3.5 | Helped |
Closest papers by meaning
- Stability Analysis of Switched Linear Systems with Neural Lyapunov Functions (AAAI 2024 · similarity 0.92)
- Lyapunov-stable Neural Control for State and Output Feedback: A Novel Formulation (ICML 2024 · similarity 0.91)
- Analytical Lyapunov Function Discovery: An RL-based Generative Approach (ICML 2025 · similarity 0.90)
- Certifying Stability of Reinforcement Learning Policies using Generalized Lyapunov Functions (NeurIPS 2025 · similarity 0.89)
- Two‑Stage Learning of Stabilizing Neural Controllers via Zubov Sampling and Iterative Domain Expansion (NeurIPS 2025 · similarity 0.88)
- Learning-enabled Polynomial Lyapunov Function Synthesis via High-Accuracy Counterexample-Guided Framework (CVPR 2025 · similarity 0.87)
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.