A PRIMER ON THE DESIGN AND SCIENCE OF COMPLEX SYSTEMS
In both Engineering and Science, the term 'complex system' is used to characterize an entity that is either being designed or observed. This often means that the system has an analytically challenging number of interacting elements, which are described at di erent levels and which need to be understood from di erent perspectives. When the relationships between these di erent levels and perspectives are not well-de ned (or are subject to change), the system can be seen as exhibiting unexpected behaviours, sometimes referred to as 'emer- gence'. Such emergent behaviours might correspond to unanticipated failures or to robust 'self-organising' patterns that can be exploited. It might be tempting to see emergence, self-organisation and other aspects of complexity as inherent to some systems. However, this primer makes no such assumption. Instead, our starting premise is that any system can be described in a multitude of ways. What distinguishes a complex system from a non-complex system is that we do not understand that system well enough to realise our objectives. In other words, 'complexity' is subjective; it describes the stance that is being taken towards a system1. That complexity can itself be characterised in many di erent ways (e.g. emergence) depending on the di erent ways in which this shortfall in under- standing is manifest (e.g. unpredictability).