How xiangjiexiao/resched works
ReSched is a reinforcement-learning application for JSSP, FJSP and FFSP scheduling, with PPO and REINFORCE implementations. A user selects an algorithm and problem configuration, then runs training or pretrained-model evaluation; the runner coordinates the environment, policy, trainer and evaluator. The environment represents scheduling state and executes actions, while generators and configured datasets supply instances. CP-SAT is an optional baseline. Relationships below follow README statements and SOURCE INDEX references; unsampled implementation details are not inferred.
Main entry points: PPO/SchedulingRunner.py PPO runner · REINFORCE/SchedulingRunner.py REINFORCE runner · SchedulingEvaluator.py PPO evaluator · cp_sat.py PPO CP-SAT
Components 25
Experiment entry
- PPO entry
PPO/SchedulingMain.py - REINFORCE entry
REINFORCE/SchedulingMain.py - PPO configs
PPO/configs - REINFORCE configs
REINFORCE/configs
PPO workflow
- PPO runner
PPO/SchedulingRunner.py - PPO policy
PPO/SchedulingModel.py - PPO trainer
PPO/PPOTrainer.py - PPO evaluator
PPO/SchedulingEvaluator.py
REINFORCE workflow
- REINFORCE runner
REINFORCE/SchedulingRunner.py - REINFORCE policy
REINFORCE/SchedulingModel.py - REINFORCE trainer
REINFORCE/REINFORCETrainer.py - REINFORCE evaluator
REINFORCE/SchedulingEvaluator.py
Scheduling domain
- PPO environment
PPO/SchedulingEnvironment.py - REINFORCE environment
REINFORCE/SchedulingEnvironment.py - PPO generators
PPO/SchedulingGenerator.py - PPO SD1 generator
PPO/SD1FJSPGenerator.py - REINFORCE generators
REINFORCE/SchedulingGenerator.py - REINFORCE SD1 generator
REINFORCE/SD1FJSPGenerator.py - PPO CP-SAT
PPO/cp_sat.py - REINFORCE CP-SAT
REINFORCE/cp_sat.py
Data and baselines
- PPO utilities
PPO/utils.py - REINFORCE utilities
REINFORCE/utils.py - Problem datasets
data - Model checkpoints
ckpt
Other components
- Experiment user
Connections 36
- Experiment user → PPO entry: runs
README.md - Experiment user → REINFORCE entry: runs
README.md - PPO entry → PPO runner: starts
PPO/SchedulingMain.py - REINFORCE entry → REINFORCE runner: starts
REINFORCE/SchedulingMain.py - PPO entry → PPO configs: selects
README.md - REINFORCE entry → REINFORCE configs: selects
README.md - PPO runner → PPO environment: initializes
PPO/SchedulingRunner.py - PPO runner → PPO policy: creates
PPO/SchedulingRunner.py - PPO runner → PPO trainer: creates
PPO/SchedulingRunner.py - PPO runner → PPO evaluator: uses
PPO/SchedulingRunner.py - PPO runner → PPO utilities: calls
PPO/SchedulingRunner.py - PPO trainer → PPO policy: evaluates policy
PPO/PPOTrainer.py - PPO trainer → PPO utilities: uses
PPO/PPOTrainer.py - PPO evaluator → PPO policy: evaluates policy
PPO/SchedulingEvaluator.py - PPO evaluator → PPO utilities: uses
PPO/SchedulingEvaluator.py - PPO environment → PPO generators: constructs
PPO/SchedulingEnvironment.py - PPO environment → PPO SD1 generator: constructs
PPO/SchedulingEnvironment.py - PPO CP-SAT → PPO generators: uses
PPO/cp_sat.py - PPO CP-SAT → PPO SD1 generator: uses
PPO/cp_sat.py - REINFORCE runner → REINFORCE environment: initializes
REINFORCE/SchedulingRunner.py - REINFORCE runner → REINFORCE policy: creates
REINFORCE/SchedulingRunner.py - REINFORCE runner → REINFORCE trainer: creates
REINFORCE/SchedulingRunner.py - REINFORCE runner → REINFORCE evaluator: uses
REINFORCE/SchedulingRunner.py - REINFORCE runner → REINFORCE utilities: calls
REINFORCE/SchedulingRunner.py - REINFORCE trainer → REINFORCE policy: evaluates policy
REINFORCE/REINFORCETrainer.py - REINFORCE evaluator → REINFORCE policy: evaluates policy
REINFORCE/SchedulingEvaluator.py - REINFORCE environment → REINFORCE generators: constructs
REINFORCE/SchedulingEnvironment.py - REINFORCE environment → REINFORCE SD1 generator: constructs
REINFORCE/SchedulingEnvironment.py - REINFORCE CP-SAT → REINFORCE generators: uses
REINFORCE/cp_sat.py - REINFORCE CP-SAT → REINFORCE SD1 generator: uses
REINFORCE/cp_sat.py - PPO configs → Problem datasets: configures loading
README.md - REINFORCE configs → Problem datasets: configures loading
README.md - PPO runner → Model checkpoints: loads or saves
README.md - REINFORCE runner → Model checkpoints: loads or saves
README.md - PPO CP-SAT → PPO evaluator: baseline comparison
- REINFORCE CP-SAT → REINFORCE evaluator: baseline comparison