I'm a PhD student in the institute of Geography. Having a background in Computer Science, I moved to GeoSciences to pursue my interest in agent-based ecological modelling. I generally keep myself very busy with other projects, music, motorcycles, sport and of course drinking!
From a young age I've been interested in computers and their seemingly unlimited capacity to facilitate human expression. When I first discovered my interest in computers in high school I started (like most kids do) by creating games. These games became more and more complex until I found myself creating physics models and 3D graphics engines.
Taking on a BSc degree in Computer Science at university provided an opportunity to further develop my computing skills. The course touched on a wide variety of fields including:
The final year involved a dissertation entitled "A Visualisation Tool for Pasture Management" with the original objective of producing a system to visualise a livestock farmer's fields for management of rotational grazing. Being very interested in the whole subject area I eventually underpinned the 'visualisation' with a biophysical model of pasture growth and a behavioural agent-based model of grazing ruminants. This project was my first 'real' introduction to computer modelling. A specialism which was set to develop.
My current PhD was intended to be a continuation of my undergraduate research, but as I dug deeper into the literature I discovered some more pressing needs for research. It appeared that biophysical models have received much attention and success. Meanwhile modelling of human decision making in crucial socio-environmental systems such as agriculture and urban planning was far less mature, with many models adopting the unrealistic profit maximisation approach. My research addresses this issue by investigating approaches used to model human decision making in socio-environmental models. My research involves the development of a BDI agent-based model with experiments involving human subjects and participatory simulation.
With greater political attention to land use change and the growth of agent-based modelling an increasing number of agent-based land use models are being used in land use research. With no obvious method for simulating human decision making, modellers are using a multitude of techniques including utility maximisation and mathematical programming. What are the merits of each and are any particularly suited to this domain? This research project aims to shed light on these questions by employing an agent-based model of land use to compare the behaviour of human subjects with a set of artificial agents.
The ability of the artificial agents to mimic their human counterparts will be measured using both quantitative output metrics and qualitative analysis of decision narratives. Preliminary studies using human subjects in a simplified land use scenario indicate that the BDI (Belief-Desires-Intentions) model provides a more intuitive representation of the human thought process than traditional decision theoretic approaches. BDI has the further advantage of supporting a fairly straightforward means of encoding human decision behaviour from semi-structured interview data. Further investigation with a more realistic case study using dairy farmers as subjects will provide a real world grounding in which to properly evaluate the research method and findings.
C. Rider & F. Reitsma, PastureSim: A visualisation tool for pasture management, Innovations in GIS 13, 2008