Game Theory for Reliable Computer Systems

My recent research develops game theory techniques and applies them to the problem of constructing reliable computer systems. Game theory is gaining importance in many disciplines, including economics, sociology, political and sciences. In my research, I have used games to automatically construct, correct, and analyze computer systems and their specifications. The games I study are turn-based games. Each game is played on a finite automaton; each state in the automaton is owned by one of the players. Initially, a token is put on a state. The player owning the state with the token can move the token along any of the outgoing edges to a next state, and the next turn starts. In general, plays are infinite sequences of turns. The number of players and their objectives vary with the application. The fundamental techniques used in this branch of game theory go back to Church [Chu62], McNaughton [McN66], Büchi and Landweber [BL69], and Rabin [Rab69]. These results have shown that game theory is in principle applicable to problems such as synthesis of systems from specification. However, previous techniques failed to yield practical algorithms. With my collaborators, I made the following contributions to the area of game theory and its applications to computer systems: