droftware / Medusa

Simulation of hide and seek game
BSD 2-Clause "Simplified" License
2 stars 1 forks source link
ai multiagent-systems python simulation

Medusa

Developing strategies for a massive hide and seek game

Hide and seek is a game in which players who are hiders try to hide themselves and the players who are seekers try to find all of them. We study this game by formulating, simulating and physically validating various strategies for both the hiders as well as seekers. It is a hierarchical decision making problem in which the team of hiders and seekers need to figure out efficient strategies and tactics as well as do optimal micro-management at an individual robot level so as to win the game. It requires artificial intelligence concepts such as multi agent control in an adversarial environment, plan recognition from partial observations and decision making under uncertainty.