Search

Browse

• Communities
• Authors
• Year
• Titles
• Subjects

My Space

• Login

Implementation of a software server and of an autonomous agent for the cooperative strategic game "Pandemic"

Voloudakis Konstantinos

Full record

---

Summary

Artificial Intelligence (AI) has in our times made considerable progress, and manifests itself in various devices and infrastructure used in our everyday lives. It has thus become even more important than ever to measure AI’s progress via pitting intelligent agents against humans in board games, following a decades-long tradition. Still, testing AI in board games that require the players’ cooperation without any element of competition among them during a game instance, is extremely rare. At the same time, this is important since the ability to cooperate effectively is a key component of intelligence. Against this background, in this thesis we develop the necessary server infrastructure that enables the participation of autonomous agents in the multiagent cooperative board game "Pandemic"; along with seven autonomous agents that can compete in the game, each of which incorporating different strategies that we also devised. Our server has already been used in the context of an AI-related course at the Technical University of Crete, and can provide the basis for a richer infrastructure that can be potentially employed for holding an international challenge for Pandemic-playing autonomous agents. We complement our work with a preliminary experimental evaluation of our agents’ strategies, drawing lessons on their effectiveness in Pandemic domains of varying difficulty.

Available Files

Services