Ever since highschool I have been a big fan of classics, and in particular Greek history and literature. I took some papers in Greek history and ancient warfare during my degree in computer science and came across an opening for an interesting piece of cross-disciplinary research that allowed me to use my computer science 'toolkit' to study the subject of hoplite battle...
The research is an application of agent-based modelling to the study of hoplite battle. Hoplite battle is the form of heavy-infantry warfare practised by the Greeks of the 5th and 6th centuries BC.
To get a quick idea of what the research is all about, download this video clip (9.7MB .avi) of one of the battles fought by the 'virtual hoplites' that I created. The view of battle is from directly overhead, looking down on individual hoplites who interact while acting autonomously in accordance with five simple rules and in response to their own immediate situation. Each hoplite knows nothing about the wider progress of the battle, and there is no centralized organization. Everything that happens in the battle is entirely attributable to the interaction of the hoplites. The battle demonstrates that the sort of complex behaviour that we associate with hoplite battle - such as low casualties, a drift to the right, each side winning on their right, localized advantage held by each army at different parts of the line, the collapse, and a relationship between cohesion and survival - can be explained in terms of the interaction of individual hoplites. It also shows that the othismos - which is traditionally thought of as massed shoving - can be explained as a consequence of a large number of individual hoplites striving to gain forward ground.
The research is documented in my Honours dissertation (1.1MB) which was presented to the Institute of Information Sciences and Technology at Massey University in November 2004.
The research has been very successful and I was fortunate enough to get the opportunity to present it to the Australasian Society For Classical Studies at the ASCS XXVI Conference held in Dunedin during 30th January - 3rd February 2005. The presentation I gave is available in MS Powerpoint here (0.6MB).