Rationale
Complexity Theory is a loose cluster of theories and methodologies aimed at understanding the properties of complex adaptive systems. Complex adaptive systems (CAS) are ones characterised by: emergence; path dependency: non state equilibrium; and adaptation. The heroin drug market fits these characteristics nicely. The tools we have available to understand and model such CAS include Multi-Agent Systems, Dynamical Systems, and Network Theory. Scientists using Multi-Agent Systems tend to focus on the individual components interacting within a given system. This is a purely bottom-up approach where representations of the individual components – the agents – display a large autonomy of action. Hence, system-level behaviours and patterns emerge from a multitude of local interactions. Scientists using Dynamical Systems tend to focus on the flows of information, mass and energy within a given system. Practically, modellers describe systems as a set of modules or compartments (stocks) interlinked by flows and controls. Scientists using Network Theory tend to focus on the structure of interactions between individual components of a system.
The Complexity Theory group considered the advantages and limitations of using these approaches for DPMP. Two key issues shaped the boundaries and content of the present project: finding a case study that would contain – a priori – as much complexity as possible and would provide the information needed to build a consistent model; and fitting into the actual structure of the DPMP project in order to interact efficiently with relevant experts and to avoid undesirable overlapping with other on-going research.
Dynamical Systems were already being explored within the epidemiology team. Looking at the Australian illicit drug markets through a cross-scale approach, it seemed that urban districts constituting a ‘drug scene’ involved most of the actors (with exception of importing syndicates and production cartels) while displaying a maximal complexity. We chose the multi-agent system (rather than network theory) because most of the potential agents in the system were clearly identified but various aspects of their interdependent links were ill-defined. In addition, the high level of transdisciplinarity that was needed advocated for an intuitive modelling approach.