A Retraction Theorem for Distributed Synthesis

Abstract: We present a general theorem for distributed synthesis problems in coordination games with $\omega$-regular objectives of the form: If there exists a winning strategy for the coalition, then there exists an "essential" winning strategy, that is obtained by a retraction of the given one. In general, this does not lead to finite-state winning strategies, but when the knowledge of agents remains bounded, we can solve the synthesis problem. Our study is carried out in a setting where objectives are expressed in terms of events that may \emph{not} be observable. This is natural in games of imperfect information, rather than the common assumption that objectives are expressed in terms of events that are observable to all agents. We characterise decidable distributed synthesis problems in terms of finiteness of knowledge states and finite congruence classes induced by them.