Skip to main content

Distributed Games and Strategies

Glynn Winskel

I will introduce distributed (concurrent) games and strategies, outline their original motivation - as a new foundation for semantics - and the current state of research on them. In a distributed game it is imagined that a team Player and a team Opponent compete in a game distributed over possibly many different locations. The distributed nature of a game is not modelled by describing the team players or locations explicitly; rather a more abstract view is taken that the history of a play of a game should no longer be described by a sequence of moves (as is the tradition), but instead by a partial order. In a nutshell, distributed games and strategies can be described as a redevelopment of game semantics and game theory in which plays take the shape of partial orders of causal dependency between moves. The work is grounded in event structures, a model of distributed/concurrent computation, which can roughly be thought of as trees in which branches are partial orders.