How wenniebyfoxmail/rmab-rdt works
RMAB-RDT is a research prototype for road digital-twin maintenance scheduling. Its main workflow configures road-segment arms and transition dynamics, simulates competing update policies under limited resources, and analyzes AoII outcomes across synthetic and Ontario-calibrated cases. The graph includes the documented experiment branches and the model, policy, calibration, and benchmark layers. Several links are supported by the SOURCE INDEX or README rather than complete source excerpts; those limits are reflected by omitting uncertain relationships.
Main entry points: config.py Simulation Config · environment.py RMAB Environment · 01_main_experiments.py Main Experiments · 02_regime_map.py Regime Map
Components 19
Experiments
- Main Experiments
01_main_experiments.py - Regime Map
02_regime_map.py - Time-Varying Runs
03_time_varying.py - Indexability Checks
04_indexability.py - Noise Sensitivity
05_noise_sensitivity.py - LP Benchmark
lp_comparison.py
Simulation Runtime
- Simulation Config
config.py - RMAB Environment
environment.py - Parallel Runner
parallel_utils.py
Models and Policies
- NHGP Transitions
nhgp_builder.py - Whittle Solver
whittle_solver.py - Scheduling Policies
policies.py
Calibration
- Ontario Data Loader
ontario_data_loader.py - Ontario Calibration
ontario_calibration.py - LTPP Calibration
ltpp_calibration.py
Evaluation
- Statistical Analysis
06_statistical_analysis.py - Convergence Analysis
07_convergence_analysis.py - Experiment Outputs
results
Other components
- Researcher
Connections 41
- Researcher → Main Experiments: runs experiments
README.md - Researcher → Regime Map: runs analysis
README.md - Main Experiments → Simulation Config: loads settings
01_main_experiments.py - Main Experiments → RMAB Environment: simulates arms
01_main_experiments.py - Main Experiments → Whittle Solver: computes indices
01_main_experiments.py - Main Experiments → Scheduling Policies: compares policies
01_main_experiments.py - Main Experiments → Parallel Runner: parallelizes runs
01_main_experiments.py - Regime Map → Simulation Config: loads settings
02_regime_map.py - Regime Map → RMAB Environment: simulates arms
02_regime_map.py - Regime Map → Whittle Solver: computes indices
02_regime_map.py - Regime Map → Scheduling Policies: compares policies
02_regime_map.py - Regime Map → Parallel Runner: parallelizes runs
02_regime_map.py - Time-Varying Runs → NHGP Transitions: builds transitions
03_time_varying.py - Time-Varying Runs → Simulation Config: loads settings
03_time_varying.py - Time-Varying Runs → RMAB Environment: simulates arms
03_time_varying.py - Time-Varying Runs → Whittle Solver: computes indices
03_time_varying.py - Time-Varying Runs → Scheduling Policies: compares policies
03_time_varying.py - Indexability Checks → Simulation Config: loads settings
04_indexability.py - Noise Sensitivity → Simulation Config: loads settings
05_noise_sensitivity.py - Noise Sensitivity → RMAB Environment: simulates arms
05_noise_sensitivity.py - Noise Sensitivity → Whittle Solver: computes indices
05_noise_sensitivity.py - Noise Sensitivity → Scheduling Policies: compares policies
05_noise_sensitivity.py - LP Benchmark → Simulation Config: uses arm classes
lp_comparison.py - LP Benchmark → Whittle Solver: compares indices
lp_comparison.py - LP Benchmark → Scheduling Policies: compares policies
lp_comparison.py - LP Benchmark → RMAB Environment: models instances
lp_comparison.py - LP Benchmark → Parallel Runner: parallelizes runs
lp_comparison.py - Simulation Config → NHGP Transitions: loads NHGP classes
config.py - Simulation Config → Ontario Data Loader: loads Ontario data
config.py - Ontario Calibration → Simulation Config: uses model settings
ontario_calibration.py - Ontario Calibration → NHGP Transitions: validates transitions
ontario_calibration.py - Ontario Data Loader → NHGP Transitions: calibrates transitions
- LTPP Calibration → Simulation Config: uses model settings
ltpp_calibration.py - LTPP Calibration → NHGP Transitions: builds calibrated model
ltpp_calibration.py - RMAB Environment → Experiment Outputs: records outcomes
- Statistical Analysis → Experiment Outputs: analyzes outputs
- Convergence Analysis → Simulation Config: loads settings
07_convergence_analysis.py - Convergence Analysis → RMAB Environment: runs trajectories
07_convergence_analysis.py - Convergence Analysis → Whittle Solver: computes indices
07_convergence_analysis.py - Convergence Analysis → Scheduling Policies: runs policies
07_convergence_analysis.py - Convergence Analysis → Experiment Outputs: writes diagnostics