This is a list of benchmark datasets and problems for research in control, optimization, and networks.

Control

• Control Gallery
  • Collection of standard system models
  • In development
  • Developed by UT Dallas

System Identification

• Nonlinear system identification benchmarks

Reinforcement Learning

• Gym
  • RL benchmark problems
  • Developed by OpenAI
• Safety Gym
  • RL benchmark problems w/ focus on safe training
  • Developed by OpenAI
• Baselines
  • Reference implementations of RL algorithms
  • Developed by OpenAI
  • Stable Baselines
    • Third-party fork w/ improved user friendliness
• RoboNet
  • Joint work from UC Berkeley, Stanford, UPenn, and CMU
  • Blog post from BAIR
• Flow
  • RL for automobile traffic control
  • Compatible w/ SUMO
  • Developed by UC Berkeley

Machine Learning

UCI Machine Learning Repository

• Datasets w/ a focus on traditional supervised ML
• Maintained by UC Irvine