State space compression in history driven quasi-static scheduling

This paper presents efficient compression techniques to avoid the state space explosion problem during quasi-static task scheduling of embedded, reactive systems. Our application domain is targeted to one-processor software synthesis, and the scheduling process is based on Petri net reachability analysis to ensure cyclic, bounded and live programs. We suggest two complementary compression techniques that effectively reduce the size of the generated schedule and make the problem tractable for large specifications. Our experimental results reveal a significant reduction in algorithmic complexity (both in memory storage and CPU time) obtained for medium and large size problems.