Research map / Reinforcement learning / Policy gradient and actor-critic
MS-PPO: Mean Standard Deviation Proximal Policy Optimization for Reliable Parking Space Search in Structured Environments
Haoming Chen, Hongliang Guo
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 PPO clipped mean-std advantage (MS-PPO vs MS-AC) (ablation) | MS-AC · Shenzhen/Singapore/Chicago, zeta=10 | mean + zeta*std search time: Shenzhen 123.04 -> 111.96; Singapore 124.23 -> 113.42; Chicago 139.20 -> 129.34 | Improved |
| use full MS-PPO instead of MS-TD (epsilon-greedy) (ablation) | MS-TD · Shenzhen/Singapore/Chicago, zeta=10 | mean + zeta*std search time: Shenzhen 128.81 -> 111.96; Singapore 127.61 -> 113.42; Chicago 139.02 -> 129.34 | Improved |
| use full MS-PPO instead of MS-PG (ablation) | MS-PG · Shenzhen/Singapore/Chicago, zeta=1 | mean + zeta*std search time: Shenzhen 64.62 -> 58.71; Singapore 51.58 -> 49.78; Chicago 64.42 -> 61.99 | Improved |
| use full MS-PPO instead of variants at small zeta on large networks (ablation) | MS-TD/MS-AC/MS-PG · Singapore, zeta=0.1 | mean + zeta*std search time: 43.41 vs 44.78/44.41/44.43 | No clear effect |
| MS-PPO vs ISSA heuristic baseline | ISSA · Chicago, zeta=10 | mean + zeta*std search time: 127.38 -> 129.34 | Mixed |
| use MS-PPO instead of mean-minimizing baselines at zeta=10 | DQN/A3C/ACO/DABST · Shenzhen, zeta=10 | mean + zeta*std search time: 138.81 -> 111.96 vs DQN | Improved |
| small reliability coefficient zeta=0.1 | MS-PPO vs other baselines · three parking networks | mean + zeta*std search time | No clear effect |
Closest papers by meaning
- Combinatorial Optimization with Policy Adaptation using Latent Space Search (NeurIPS 2023 · similarity 0.84)
- PPO in the Fisher-Rao geometry (NeurIPS 2026 · similarity 0.83)
- One Step is Enough: Multi-Agent Reinforcement Learning Based on One-Step Policy Optimization for Order Dispatch on Ride-Sharing Platforms (NeurIPS 2026 · similarity 0.82)
- Simple Policy Optimization (ICML 2025 · similarity 0.82)
- Colored Noise in PPO: Improved Exploration and Performance through Correlated Action Sampling (AAAI 2024 · similarity 0.82)
- Occupancy-based Policy Gradient: Estimation, Convergence, and Optimality (NeurIPS 2024 · similarity 0.82)
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.
