Multi-Agent Adversarial Search

This project was created keeping in mind the various Pac-man game variations that were created by Berkley University (http://ai.berkeley.edu/project_overview.html). This specific project assumes the Pacman and associated ghosts in the game each to be an agent in the game in a multiagent competitive deterministic environment.

Language used for implementation - Python 3.6

The following utilities were implemented as a part of this project -

  • Implementation of multiagent minimax algorithm with Alpha Beta pruning
  • Implementation of Expectimax algorithms (with evaluation functions)
Code can be viewed on Github on request