Towards Valid Models and Simulations of Complex Systems. A Plant Ecology Case-Study

Complexity and complex systems have been intriguing research objectives since the discovery of the well known Aristotelian paradox: the whole is more than the sum of its parts. While defining complexity remains elusive and “emergent”, studying complex systems may be the only way towards further conceptual insights. However, current day techniques of modelling and simulating are not subject to principled methods of development, while their scientific scope is often insufficiently motivated. In addition, the way computer models are presented is often distilled into verbal descriptions that do not facilitate complete reimplementation. One way of addressing these limitations is to use rigorous software engineering methods, coupled with techniques that allow the construction of structured arguments of validity. This is not an easy task, as complex systems are through their definition decentralised and emergent. Bringing together the knowledge necessary to develop such principled methods is a first step towards their refinement. Here we review first the literature relevant to the domain of complex systems, focusing on techniques for engineering emergent properties and for validating computer simulations. We then introduce plant ecology, the domain of our case-study, together with the main types of modelling techniques that have been employed so far. Two protocols that may potentially contribute to the wider field of computer modelling and simulating, are also described. Preliminary results are available in the form of a published article on developing equivalence arguments for complex systems and the development of a large-scale, process-oriented simulation. While arguing about equivalence is a first step towards formulating arguments of validity, making use of the developed simulation will enable us to address scales that are closer to reality. In conclusion, this document reviews literature from complexity science, computer and software engineering, in an effort to structure the way towards valid computer models and simulations. We are in the initial phase of our research, but addressing hard matter subjects such as arguing validity are within reach.