Yannick Delbecque — site personnel

Information Flow in Game Semantics

Abstract

To successfully model an array of different programing languages, game semantics uses the detailed interactions between a system and its environment. To fine-tune such a model to a particular language, the set of strategies available to the players is limited using various conditions. In many cases these restrictions can be thought as the result of limiting the information available to the players to elaborate their strategies.

The aim of this paper is to study two frameworks to represent this partial knowledge explicitly. The first add to games as defined in game semantics equivalence relations identifying the position that cannot be distinguished by the players. The resulting structure is related to the process of restricting the player’s strategies by a Galois connection. The second framework is based upon a network game representation derived from coloured Petri nets. The games of the first framework are used to study the dynamics of this network game representation, so it can also be connected to game semantics. This relation is then used to give a characterisation of total information games in the network representation. A sketch of the basic constructions needed to define a category of games and strategies using the network representation is given.