🤖 AI Summary
Timed parity games suffer from unintuitive semantics and a lack of efficient solving algorithms for controller synthesis. To address this, we propose a refined semantic model that enhances both the modeling fidelity of timed interactions and alignment with intuitive reasoning. We further introduce, for the first time, a zone-based symbolic algorithm that integrates the UppAal zone library with parity-objective-driven state-space exploration. This approach overcomes the longstanding challenge of adapting parity objectives to zone abstractions in timed games—a key bottleneck in prior work. Experimental evaluation confirms the feasibility and effectiveness of our method. Notably, it constitutes the first scalable, zone-based framework for automated synthesis of complex real-time control strategies in timed systems.
📝 Abstract
This paper revisits timed games by building upon the semantics introduced in "The Element of Surprise in Timed Games". We introduce some modifications to this semantics for two primary reasons: firstly, we recognize instances where the original semantics appears counterintuitive in the context of controller synthesis; secondly, we present methods to develop efficient zone-based algorithms. Our algorithm successfully addresses timed parity games, and we have implemented it using UppAal's zone library. This prototype effectively demonstrates the feasibility of a zone-based algorithm for parity objectives and a rich semantics for timed interactions between the players.