/prismToMILP

Primary LanguagePython

ambiguousPref

  1. Edit PRISM file "simple_map.prism" and export model files (tra, sta, lab, rew) using the command:
    prism simple_map.prism -exportmodel .tra,sta,lab -exporttransrewards .rew
    (You need to rename the rewards files to make sure it is in the format of simple_mapi.rew where i is the reward index)
  1. Parse the model files into data structures
    python parse.py simple_map THRESHOLD_LOWER THRESHOLD_UPPER
    should generate simple_map.py
  1. Compute bfs costs pre and post decision points
    python bfs_decision.py
    (it should generate pre-decision.props and post-decision.props and one model file for each decision point, i.e. simple_map$i$.prism where i is the state number for each decision point)
  1. Check the cost of pre and post decision point in properties
    prism simple_map.prism pre-decision.props -exportresults pre-decisionLog.txt
    prism simple_map$i$.prism post-decision.props -exportresults post-decisionLog$i$.txt (general case where i is the decision point number)
  1. Generate lp formulation by
    python generate-lp.py and get lp-output.txt
  1. Go to https://online-optimizer.appspot.com/?model=builtin:empty.mod and copy the content in lp-output.txt
    Solve the model and view the result in variables section