I found this while looking for some primary sources online and found it quite stimulating-
http://www.iaa.bham.ac.uk/research/fieldwork_research_themes/projects/logistics/Manzikert/Index.htm
Medieval Warfare on the Grid
JISC EPSRC AHRC E-Science Programme
Principal Investigators
Professor Vincent Gaffney and Dr Georgios Theodoropoulos
4 Year Bursary available
Professor Vincent Gaffney (IAA) and Dr Georgios Theodoropoulos (Comp. Sci.) have received a major E-science project grant of 447,000. The project, entitled "Warfare on the GRID", will use the new University "Large Cluster" to explore the military logistical context of the Battle of Manzikert in 1071. Manzikert is a key historic event in Byzantine history. The defeat of the Byzantine army by the Seljuk Turks, and the following civil war, resulted in the collapse of Byzantine power in central Anatolia. An event that was so traumatic that it became known as 'The dreadful day'. Given the key position this event takes within the collapse of Byzantine power, the lack of consensus between historians on the numbers of men involved at, or even the route taken by the Byzantine Army to, Manzikert is profound. This project builds on work initiated by Professors Gaffney and Haldon through the Birmingham/Princeton Medieval Logistics Group, and seeks to address the problems associated with early military logistics through a distributed GRID analysis using an agent-based model. The results will have significant implications for study of pre-industrial societies in methodological and theoretical terms and will benefit academics with an interest in comparative military history, the cultural role of military organisation and the relationship of historical and modelled data. This project will be funded over 4 years and includes an open 4 year PhD bursary for a historic researcher to be trained in the application of GRID technologies used as part of the project. Further Information on applications to follow.
Assumed route of the Byzantine army to Manzikert
Management Committee
Professor John Haldon, Professor of Byzantine History, Princeton
Dr Lutgarde Vandeput, Director, British Institute in Ankara
Dr Warren Eastwood, Geography and Environmental Science, Birmingham
Dr Robert Fletcher, White Rose Grid and York
Professor Steve Turner, Nanyang Technological University, Singapore.
Dr Stuart Dunn Arts and Humanities E-Science Support Centre
Dr Simon Esmonde-Cleary, University of Birmingham
Professor Aaron Sloman, Birmingham
Mr Paul Hatton Information Services, Birmingham
Interim Project Description
The need for mediaeval states to collect and distribute resources to maintain armies affected all aspects of political organisation and at critical times, when armies failed, the results could prove disastrous to society. Despite this, military studies seldom progress past the study of existing texts to bear out the pragmatic consequences of military behaviour, even though military activity in terms of resource allocation and consumption is decisive in shaping pre-modern societies. This project gathers historians, archaeologists and computer scientists in a project aimed at modelling logistical arrangements relating to the battle of Manzikert (AD 1071)[1], a key event in Byzantine history. The defeat of the Emperor Romanos IV Diogenes by the Seljuk Turks, and the civil war that followed, resulted in the collapse of Byzantine power in central Anatolia. These events were so traumatic, and the association with Manzikert so profound, that the defeat became known as 'the dreadful day. Given the significance attributed to these events and, ultimately, the collapse of Byzantine power regionally, the lack of consensus between historians on the numbers of men involved at, or even the route taken by the Byzantine Army to, Manzikert is profound. Repeated debate on the arrangements leading up to this critical encounter suggest the need for alternative methodologies that can break cycles of academic claim and counter-claim, have a wider applicability to military research and appreciate the role of military studies within broader cultural studies.
This project will provide a fundamental re-analysis of the Manzikert campaign and illustrate the use of Grid-aware distributed simulation techniques to model movement and sustainability of historic armies. The study will involve multiple simulations of varied army units moving within a digital environmental database collated at Birmingham. Alongside primary and secondary source material, the data will be used to interpret events related to the battle and assess contemporary interpretation of historic sources. Such simulation methodologies have a wide applicability and allow the re-use of models and processes in comparable regional or period studies.
Project goals are:
o To establish a novel, rigorous computational frameworks for analysis of historical, military logistical data (pertaining to movement and communications; production, allocation, consumption of resources; settlement patterns) for the early medieval periods in Europe and the Near East, using environmental data held at Birmingham, and as a collaborative research programme between Birminghams Institute of Archaeology and Antiquity and Department of Computer Science.
o To create a reusable GRID-enabled simulation and data analysis infrastructure, that supports the framework outlined above.
o To utilise a GRID modelling framework to improve our understanding of the central role of warfare and conflict in the medieval periods. To use the results of multiple simulations of the Manzikert campaign to critique previous research on the battle.
o To demonstrate, using Manzikert, the opportunities provided the historical community by large-scale distributed simulation technologies and provide fundamentally new insights into established historical questions.
Research Context
Armies possessed a role within mediaeval states that extended beyond their military function. The need to maintain permanent military structures ensured that military groups were formative in structuring society, affecting most aspects of political organisation, resource allocation and consumption. Although decisive in shaping pre-modern societies and despite academic interest in military history of all periods, research seldom progresses past the study of existing texts to explore the implications of military behaviour[2]. A methodology for analysing pre-modern military logistics is required which places campaigns within their wider cultural context. To achieve this, factors in the organisation of medieval societies (communications, resource production and consumption) must be integrated within an historic framework incorporating settlement data and land-use. Although these themes have been the object of study, they are frequently characterised by unproven assertions and rarely tested against a range of evidence. This project will bring together historians with experts on archaeological settlement and environmental data, to provide a strong set of models representing the behaviour of armies, that can be implemented as part of a large distributed simulation exercise.
Whilst the outcome of the battle of Manzikert appears clear, historically, its wider significance is contentious. There is uncertainty concerning the nature and disposition of Byzantine forces prior to battle and even aspects of the route taken to Manzikert. Clearly, our interpretation of this important event is limited if we cannot establish, with some certainty, the parameters of the military forces involved, the nature of the action and, ultimately, the wider historic context of the defeat. However, repetitive argument over sparse references within mediaeval texts gets us no closer to understanding these issues, and even results produced by competent and wide-ranging scholars remain subjective. The goal of the project will be to model the movement of a range of troop dispositions suggested for Manzikert against the terrain and environment, and to compare these data with the historic account of the battle. The route from Constantinople to Manzikert is not actually explicit in the sources, although communication and resource data suggest that there were, perhaps, only two options. This is critical as logistical requirements for these routes are central to current debate on how Manzikert was fought and the numbers of troops involved, and this will be subject to multiple simulation events to identify best fit. In cultural terms there will be the requirement to consider the physical context of the campaign, historical structures evolved to meet logistical demands and logistical responses to warfare or the nature of Byzantine support for the military more generally. This project builds on work by the Birmingham/Princeton Medieval Logistics Group, and addresses these issues directly through a novel GRID-based, distributed simulation. The results will have significant implications for the study of pre-industrial societies in methodological and theoretical terms and will benefit academics with an interest in comparative military history, the cultural role of military organisation and the relationship of historical and modelled data
Research Method
Modelling and forecasting using descriptive agent-based simulation.
Agent-based simulations are increasingly used to model social systems. Running agent-based models as simulations can generate hypothetical explanations of events which challenge existing, or support new, hypotheses. Although hypotheses can be generated without simulation, e.g. using static tools such as data mining, the results from simulation can lead to unexpected hypotheses that can be tested. Models generated from undirected data mining and without domain expert knowledge are bottom-up exercises and can make predictions that are irrelevant or contain no new information. By contrast, descriptive agent-based models make full use of expert knowledge. The consequences of a models rules are not known in advance and the model, as a simulation, can generate unexpected outcomes. These properties may be used to form hypotheses that can lead to further, focused data collection or directed analysis.
This project will develop models of different actors related to the Manzikert campaign, derived from secondary and primary sources. These will be used in agent-based simulations to analyse events and generate what-if scenarios based on models of military behaviour. Simulations will run on different levels of abstraction, (soldiers and commanders, army units, geographical regions). These models will provide understanding, have explanatory value for the end user and assist in the analysis of medieval warfare.
A number of methodologies offer substantial opportunities for descriptive modelling within this environment. Foremost is Optimal Foraging Theory. Originating in classical mathematical biology[3], this emphasises the advantages of behavioural strategies that maximise net energy intake per unit time spent foraging. Although armies fit the general parameters of optimal foraging theory well, such models have rarely been applied because they rarely support any cultural constraints. However, modified foraging theory accepts important cognitive and cultural controls, including lack of information on which to base decisions. As pre-modern campaigns were often characterised by ignorance of local conditions, this suggests that ancient campaigns should approximate modified optimal foraging behaviour and that this can be modelled.
Foraging models demand a detailed approximation of the environment as well as a technical base to permit analysis and modelling. Previously, few technologies could adequately approximate 4 dimensional models (X, Y, Z values plus time/movement). Consequently, iterative path finding algorithms are central to this application. Shortest path analyses, using the A* algorithm, should be helpful here with routes to battle determined by looking at each action performed within the terrain and the new states generated by previous actions. Together with Game Theory, decision theory, probability and associated statistical methods these will provide the mathematical basis for decision-making. All such procedures require substantial historic and environmental databases from Birmingham, and data on the physical context, historic administrative structures supporting logistical demands and the social responses to warfare. These sources provide essential data pertaining to communications; production/allocation, consumption of resources; and settlement patterns in the early Byzantine periods. The expertise provided by the PIs and advisory group supports the generation of behavioural rules and historic constraints for the model.
Simulation Interoperability Frameworks and Grid Technologies
Development of complex simulation applications is typically an interdisciplinary enterprise requiring collaborative effort from researchers with different expertise. Occasionally, an appropriate model will be available but it is likely that a new model will have to be developed. Creating a new model for each new problem is wasteful; an alternative envisages the researcher drawing on existing work, combining and adapting components developed by others. For example, models of terrain, vegetation or even people can be input to create a new model. In doing so, researchers reduce development time and draw on the expertise of scientists who created the adapted components.
The last decade witnessed an increasing interest in distributed simulation, not simply to speed up simulations but also to link disparate simulation components and data sources at multiple locations to create a common virtual environment. The culmination of this has been the development of High Level Architecture (HLA), a framework for simulator reuse and interoperability recently adopted as an IEEE standard. HLA is being widely adopted within the simulation community; HLA-compliance will be an increasingly important feature of simulators.
Using HLA, a simulation links a number of geographically distributed simulation components (Federates), into a single larger simulation (Federation). An Federation consists of one or more Federates, a Federation Object Model (FOM), and the Runtime Infrastructure (RTI). Each Federate can model a single entity (e.g. an agent), a number of entities or it may have a different purpose (e.g. a data logger or a viewer used to steer a simulation, or even act as a surrogate for a human participant, reflecting the state of the larger simulation to some user interface and conveying decisions from the participant to the rest of the simulation). The FOM defines types and relationships between data exchanged by the Federates. The RTI middleware provides common services to the Federates and communication between Federates and between Federations is via the RTI. The emergence of Grid technologies provides new opportunities for distributed simulation, enabling collaboration and use of distributed computing resources, whilst facilitating access to geographically distributed data sets. Recently, there has been an increasing interest in utilising Grid technologies to execute HLA simulations over the Internet. An important initiative here is the HLA-Grid system, prototyped at NTU, Singapore and developed further in Birmingham. In HLA-Grid, Federates are instantiated as Grid services to facilitate communication between Federates and the RTI.
This project will develop a generic, reusable Grid-enabled infrastructure which will integrate descriptive agent-based simulations, environmental models, data sources and visualisation facilities to enable the analysis of medieval warfare events, using Manzikert as a case study. As a simulation toolkit, we anticipate using RePast, a Java-based toolkit for development of lightweight agents and agent models. RePast was developed at Chicagos Social Science Research Computing division and is derived from the Swarm toolkit. It has become a popular and influential toolkit, assessed as the most effective development platform available for large-scale simulations of social phenomena.
Bursary Information
4 Year PhD Studentship
JISC/EPSRC/AHRC E-Science Programme
Medieval Warfare on the Grid
Principal investigators
Professor Vincent Gaffney and Dr Georgios Theodoropoulos
The Institute of Archaeology and Antiquity (University of Birmingham) is offering a 4 year PhD studentship as part of the JISC/EPSRC/AHRC E-Science Programme funded project Medieval Warfare on the GRID. The thesis will have as its core the development of the primary behavioural model defining the requirements and functions of a military campaign simulation using historic data relating to the Battle of Manzikert (1071) and incorporating the wide range of types of data from historical and environmental sources for the battle available at Birmingham. Particular emphasis will be placed on the development of GRID aware simulations under the guidance of the Department of Computer Science at Birmingham. The award will be offered to a scholar with appropriate historical skills and will include training in the use of GRID-based technologies. Dissemination of core simulation skills to the wider Arts community will be central to the project and research. The studentship will be part of the team managing the Medieval Warfare on the GRID project and contribute to the activities of the wider Birmingham/Princeton Medieval Logistics Research Group.
The primary objective of this research is to assist in the design a GRID-based framework to address research issues focussing on the logistical context of the Battle of Manzikert in 1071. Manzikert is a key historic event in Byzantine history. The defeat of the Byzantine army by the Seljuk Turks, and the following civil war, resulted in the collapse of Byzantine power in central Anatolia. An event so traumatic that it became known as 'The dreadful day'. Given the key position this event takes within the collapse of Byzantine power, the lack of consensus between historians on the numbers of men involved at, or even the route taken by the Byzantine Army to, Manzikert is profound. This project will implement a detailed modelling programme which seeks to break the current cycle of academic claim and counter-claim. The student will be based in the Birmingham Visual and Spatial Technology Centre - part of the Institute for Archaeology and Antiquity.
Application forms can be obtained from the Postgraduate Office, School of Historical Studies, University of Birmingham, Edgbaston, Birmingham B15 2TT,
tel: 0044-121-414 3189 or email L.A.Robinson.1@bham.ac.uk. Applications for the studentships should be returned by 10 August 2007. Late applications may be considered in certain circumstances. Informal enquiries can be made to Professor Vince Gaffney (tel 0121 414 7632, email V.L.Gaffney@bham.ac.uk).