Dwell Time Refinement

The abstraction of a hybrid system by eliminating one or more of its continuous variables is in general too coarse. In this situation, dwell time refinement may be useful. The idea is to impose explicit constraints on the time that may resp. must be spent (or “dwelled”) between successive mode switches. We show how the required constraints can be inferred automatically (with acceptable cost). The approach is general in that it can be superposed on a range of verification methods for hybrid systems. We exemplify its practical potential by experiments on previously unsolved instances of the stability verification problem.