Multi-Agent System Simulator

We investigate a new DCOP_MST model that incorporates elements of ML in distributed search problems.

Algorithms

TODO:

  • Surveillance and Search & Detect teams from Harel's thesis
  • Algorithms from "DCOPs Meet the Real World" paper
  • Implement different metrics (Neptune)
  • Fictitious play
  • Rational learning

Implemented:

Metrics:

  • ER Loss
  • DCOP_MST Objective

Screenshot

Credits