Parallel and Distributed Optimization of Dynamic Data Structures for Multimedia Embedded Systems

Energy-efficient design of multimedia embedded systems demands optimizations in both hardware and software. Software optimization has no received much attention, although modern multimedia applications exhibit high resource utilization. In order to efficiently run this kind of applications in embedded systems, the dynamic memory subsystem needs to be optimized. A key role in this optimization is played by the Dynamic Data Types (DDTs) that reside in every reallife application. It would be desirable to organize this set of DDTs to achieve the best performance in the target embedded system. This problem is NP-complete, and cannot be fully explored. In these cases the use of parallel processing can be very useful because it allows not only to explore more solutions spending the same time, but also to implement new algorithms. In this work, we propose a method that uses parallel processing and evolutionary computation to explore DDTs in the design of embedded applications. We propose a parallel Multi-Objective Evolutionary Algorithm (MOEA) which combines NSGA-II and SPEA2. We use Discrete Event Systems Specification (DEVS) to implement this parallel evolutionary algorithm over Service Oriented Architecture (SOA). Parallelism improves the solutions found by the corresponding sequential algorithms, and it allows system designers to reach better solutions than previous approximations