Medieval Military Logistics: A Case for Distributed Agent- based Simulation Bart Craenen Georgios Theodoropoulos Vinoth Suryanarayanan School of Computer Science The University of Birmingham United Kingdom { gkt, b.g.w.craenen, vys} @cs.bham.ac.uk Vincent Gaffney Philip Murgatroyd Institute of Archaeology and Antiquity University of Birmingham United Kingdom {v.l.gaffney, psm703}@bham.ac.uk John Haldon History Department Princeton University U.S.A jhaldon@Princeton.edu ABSTRACT Historical studies are frequently perceived to be characterised as clear narratives defined by a series of fixed events or actions. In reality, even where critical historic events may be identified, historic documentation frequently lacks corroborative detail that supports verifiable interpretation. Consequently, for many periods and areas of research, interpretation may rarely rise above the level of unproven assertion and is rarely tested against a range of evidence. Simulation provides an opportunity to break cycles of academic claim and counter-claim. This paper discusses the development and utilisation of large scale distributed Agent-based simulations designed to investigate the medieval military logistics in order to generate new evidence to supplement existing historical analysis. The work aims at modelling logistical arrangements relating to the battle of Manzikert (AD 1071), a key event in Byzantine history. The paper discusses the distributed simulation infrastructure and provides an overview of the agent models developed for this exercise. Categories and Subject Descriptors I.6 [Simulation And Modeling]: Types of Simulation - Discrete event, Distributed, Gaming. Simulation Support Systems - Environments I.2.11 [Distributed Artificial Intelligence] Multiagent systems General Terms Algorithms, Design, Experimentation Keywords Agent-based modelling, distributed simulation, historical studies, medieval, military logistics 1. INTRODUCTION The analysis of humanities data sets offers considerable chal- lenges to computational science [9]. Large, complex and often characterised as partial or fuzzy, their interpretation is frequently presented in the form of assertion and the prospect of formal analysis is often dismissed as mechanistic and inappropriate to the complexities of human action or behaviour. Historical studies are a good example of the difficulties associated with such research. For many, historical interpretation is associated with clear narra- tives defined by a series of fixed events or actions. In reality, even where critical historic events may be identified, contemporary documentation frequently lacks corroborative detail that supports verifiable interpretation. Consequently, for many periods and areas of research, interpretation may rarely rise above the level of unproven assumptions, rarely or never tested. Constant, subjec- tive, argument over the same texts hampers a deeper understand- ing of the finite evidence that is available to historians. In such an academic context there is an imperative to provide alternative, novel paths toward interpretation. Computer simula- tion provides an opportunity to break cycles of academic claim and counter-claim. Having made such a statement it would not be true to suggest that there is no prior quantitative, or computa- tional, base to historical studies. Many areas of research have voluminous data sets, although their study, generally, provides abstract numeric outputs and provides a limited insight into de- tailed or individual action, and this limitation has been the object of heated debated within historical disciplines for several decades [15]. Where significant computational linkage exists the technol- ogy of choice for many historical disciplines has been GIS [1][3]. The applications of GIS have been variously significant or in- sightful but, most often, can be characterised as static models, frequently dependent on a limited, economic database, even where they may have aspirations towards the explanation of larger be- havioural patterns[10]. The application of agent-based modelling, as a means to explore the effect of individual action, has recently emerged as an area of interest for the historical disciplines. Application of such tech- nologies has, however, been extremely limited. The analysis of resource exploitation by Mesolithic hunter gatherer groups, for instance, was an early example of such work [18], whilst the process of evolution of complex societies in the fertile crescent has also been the object of sustained study [27]. Examples such as these which build on more traditional historical and archaeologi- cal work often have the benefit of being able to use certain well established known points as validation for their models. It is nota- ble, however, that most recent studies have generally involved the analysis of small-scale groups at individual or household level rather than larger societies [17]. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. DISIO 2010 March 15, Torremolinos, Malaga, Spain. Copyright 2010 ICST, ISBN 78-963-9799-87-5.