Triply articulated modelling in complex systems

We argue that traditional systems engineering (TSE) is inadequate for engineering complex systems. This inadequacy becomes particularly problematic when the system in question is a System of Systems (SoS) whose behaviour depends on how human users of the component systems, anticipating the consequences of their own behaviour, interact with those systems and with each other. Considering its users to be part of the SoS renders such a system anticipative, and therefore necessarily complex. To be effective, the designer of an SoS must incorporate an understanding of its users' models of their contexts of use. However, the process of design then becomes reflexive, since it must include within itself models of its users' anticipations of its relation to the SoS's environment, including those of its designers. As a result, complex systems engineering (CSE) has to become a distinct discipline, involving explicit processes that occur at the time of use of an SoS, not only for negotiating shared meaning between component systems within peer-to-peer relationships, but also for considering the dynamic effects of differing uses of component systems on the behaviour of the SoS as a whole. By modeling users' relations to demand as well as to the behaviour and organisation of component systems, triply articulated modeling supports the negotiation of shared meaning and accommodates the reflexivity required of a CSE regimen.