On engineering smart systems

Abstract

A smart system exhibits the four important properties: (i) Interactive, collective, coordinated and efficient Operation (ii) Self -organization and emergence (iii) Power law scaling under emergence (iv) Adaptive. We describe the role of fractal and percolation models for understanding smart systems. A hierarchy based on metric entropy is suggested among the computational systems to differentiate ordinary system from the smart system. Engineering a general purpose smart system is not feasible, since emergence is a global behaviour (or a goal) that evolves from the local behaviour (goals) of components. This is due to the fact that the evolutionary rules for the global goal is non-computable, as it cannot be expressed as a finite composition of computable function of local goals for any arbitrary problem domain.