Under different terms – e.g. platform, hierarchic, modular, complex – systems made of interdependent components managed by interfaces have been subject for many studies in the economic and management literature. In this paper, we consider though that these studies are more focus on the components than on the interfaces managing interdependences. Consequently, based on the recent formalizations of the industrial dynamics related to complex products – as “platform architecture” (Gawer, 2009) and “dominant design model” (Murmann and Frenken, 2006), this paper is a very first proposition to formalize our project of interface analysis in these perspectives. The aim is to produce a strong enough set of hypotheses related to interfaces, in order to implement them into computational experiments of system simulations. On the technological side of complex product systems, we argue that interfaces are to be analysed independently from the technological components. Then, on the organisational side, interfaces have counterparts, via the definition of property rights, the active diffusion of technological knowledge and the hierarchical validation. These counterparts are embedded into higher-level standards which co-evolve with technologies and low-level interfaces, according to strategic and economic concerns of the complex product system stakeholders. Finally, we organize this idea according to the representation followed by evolutionary computations, as the first sketch of one simulation model. This paper is a first step in constructing an agent-based model of architectural innovation faced by complex products. Meanwhile, we attempt to keep the intuitive and simple representation of NK models; this is the reason why we have made this step back to theoretical foundations.