Optimisation problems for dynamic concurrent task-based systems

One of the most critical bottlenecks in many novel multi-media applications is their very dynamic concurrent behaviour. This is especially true because of the quality-of-service (QoS) aspects of these applications. Prominent examples of this can be found in MPEG4 and JPEG2000 standards and especially the new MPEG21 standard. In order to deal with these dynamic issues where tasks and complex data types are created and deleted at run-time based on non-deterministic events, a novel system design paradigm is required. Because of the high computation and communicating requirement from these kind of applications, multi-processor SoC (system on chip) is accepted as a solution (e.g., TriMedia), which is also promised by the advance of the processing technology. It is different from traditional general-purpose computers because it is application specific and it is an embedded system, which means costs like energy consumption are of major concern. This paper focused on the new requirements in system-level synthesis. In particular a "task concurrency management" problem formulation is proposed, with special focus on the formal definition of the most crucial optimization problems. The concept of Pareto curve based exploration is crucial in these formulations.