Match Plan Generation in Web Search with Parameterized Action Reinforcement Learning

About Environment Installation

• Match plan generation

  • Microsoft internal access

• Benchmark environments

  • referring to https://github.com/cycraig/MP-DQN

    • pip install -e git+https://github.com/cycraig/gym-platform#egg=gym_platform
      pip install -e git+https://github.com/cycraig/gym-goal#egg=gym_goal

Code Running

• pip install -r requirements.txt
  git clone https://github.com/Microsoft/nni.git

• Running

  • run_benchmark.py: manually tune hyper-parameters
  • run_benchmark_nni.py: use NNI to tune hyper-parameters

• Parameters:

  • referring to the code run_xxx.py
  • loading checkpoint: see run_benchmark.py dict:debug, load_filename = 'Goal_psac3_dl10_d100_320000_May_31_2020_16_26_12' (example)