Author(s)

A. Cointet & C. Laval

Abstract

It has been the tendency for some years, which is still confirmed today, to try to model complex systems by means of software, of successful languages and of algorithms in order to better understand, to optimize and even to develop them. However, in many cases, this modelling attempts to describe the functioning of the existing system. To analyze an existing system by its functioning description can present certain advantages but also inconveniences as for its complete understanding within the framework of objectives of diagnosis, optimization or evolution. To be able to visualize the system considered at different detailed levels, to act on parameters and to analyze diverse behaviours are, among others, part of the advantages. To develop a system by acting on functioning parameters in the aim of optimization does not guarantee a relevant result. This modelling hides the wide register of choices reserved or not reserved, and thus the logics of decisions which led to the existing solution. Furthermore, this modelling brings no proof that:  performances of the system are strictly in a direct link with the initial requirements,  all the interactions with the environments of the system are well taken into account. Moreover, it becomes even more significant when we approach the management of the risks attached to the considered system. This paper suggests the consideration of functional analysis and a description to model the complex system considered. Keywords: functional analysis, system engineering, risk management, defence in depth.

Keywords

functional analysis, system engineering, risk management, defence in depth.