Strategic Reasoning with Artificial Intelligence

Jacob Berg


Supervised by Federico Cerutti; Moderated by Matthew J W Morgan

* Background Humans are incredible problem solvers. With experience, we typically get more efficient, and effective, at the problem-solving process. However, strategic experience is not easily gained meaning other mechanisms are required to build up strategic experience for strategic success. As far back as 3000 B.C.E., the game of Wei Hai employed coloured stones to represent opposing forces. The game of chess similarly evolved out of military war games and references are found as early as 500 B.C.E. in India. Kriegspiel is a Chess variation in which the players have uncertain and partial information: they are not informed of their opponent's moves, thus they can only guess the position of enemy pieces. Each player tries to guess the position of the opponent's pieces as the game progresses: a referee accepts legal moves and rejects illegal ones.

* Project goals 1. Open source implementation of the Kriegspiel game---ICC Variant Wild 16 from the ICC https://www.chessclub.com/user/help/Kriegspiel---compatible with the PNG format---https://en.wikipedia.org/wiki/Portable_Game_Notation 2. Analysis of risks with collusion and adversarial reasoning between referee and opponent 3. Documentation and reflective analysis of the achievements

* Requirements - Python/Java programming: you are more than welcome to chose the programming language of your preference - The code must be released under MIT Licence https://opensource.org/licenses/MIT - The documentation must be released under CC-BY https://creativecommons.org/licenses/by/4.0/

* References: - Ciancarini P, DallaLibera F, Maran F. Decision making under uncertainty: a rational approach to Kriegspiel. Adv Comput Chess. 1997;8:277-298. - Hill RR, Tolk A. A History of Military Computer Simulation. 2017:277-299. doi:10.1007/978-3-319-64182-9_13.

Initial Plan (05/02/2018) [Zip Archive]

Final Report (11/05/2018) [Zip Archive]

Publication Form