A Computational study of strategy switching in large games [HBNI Th40]

Abstract

The aim of this work is to study large games: games with a large number of players and/or with a large temporal, spatial or logical structure, using techniques from automata theory, logic and game theory. It is argued that the traditional way of analysing games is not only unwieldy but also cannot be properly motivated for such games. As a lot of uncertainties are present in such a game, players usually do not strategise in a single go but do so dynamically as they observe the outcomes of the game. They compose simple strategies to build more and more complex ones. They also employ several heuristic strategies. Switching between strategies forms a central part in such compositions. The notion of strategy switching in games is formally studied.