To model larger system and control problems effectively one needs various mechanisms for constructing systems from subsystems. In this task the consortium intends to see how important, well-understood compositional methods from various areas including process algebra, Statecharts (UML), I/O automata, etc. extend to control synthesis and validation problems in the timed automata setting. There are two major difficulties: the first is that in timed systems, time is a state variable common to all components and hence one has to be careful not to define composition operators that block the evolution of the system. The second obstacle is in the different ``identity'' of the components: if one adopts a game-theoretic view (as in Task 1.5), the question is which transitions are controllable (performed at the decision of the scheduler) and which reflect uncontrolled choices of the plant. These distinctions between transitions (including the passage of time) should be maintained by the composition operation.

Other compositionality related work, in particular, the ability to infer the behaviour of a composite system from that of its components, will be discussed in the analysis and tools workpackage.

responsible partner: KUN