Paradigm-independent design of composite systems

Complex systems design methodologies are currently focused on semi-automatic generation of optimal mixture of hardware and software components by starting from algorithmic specifications. Nowadays, advanced systems include also soft-computing components. The choice of the computing paradigm (among algorithms and soft-computing approaches) to solve each component is performed a priori by the designer. The codesign is therefore applied only to the configured components within the fixed framework imposed by the given computational partitioning, thus preventing global optimization. Our purpose is to bring codesign concepts to early phases of design, reframing them to fit with paradigms instead of implementation. In this work we focus our attention on an overview of the methodology, highlighting the requirements of this innovative approach and the tools needed to perform it.