An agent-based model of trade in the East Roman Empire (25BC-150AC)

Lecture by Simon Carrignon (Barcelona Supercomputing Center, CASE).

2017.09.04 | Christina Levisen

Date Tue 10 Oct
Time 12:00 13:00
Location Centre for Urban Network Evolutions (UrbNet) Aarhus University Moesgård Allé 20, DK-8270 Højbjerg Denmark Building 4230-232

 Abstract

Traces of the processes driving economic activity in past societies are scarce and unevenly distributed. The archaeological record shows changes in the volume or in the kind of goods that were traded, but it is difficult to infer from it the social, economic and cultural mechanisms that resulted in these changes.

On the other hand, economists have developed models that link observable patterns of trade with the underlying economic principles. As a result a huge arsenal of tools and models exists nowadays that address a wide range of research question. However, the challenge of incorporating key aspects of past societies into such models remains unresolved.

Here, we present an agent-based model that fills this gap by bringing economic models and historical and archaeological knowledge together. The overarching goal is to develop a tool to test and compare different hypothesis regarding the socio-economic processes that have resulted in the observed patterns in the archaeological record.

In this particular study we want to explore the social mechanisms behind changes observed in the distribution of different types of tableware used in the East of the Roman Empire from 25 BC to 150 AC. These changes have been identified in an extensive dataset of 5121 tableware from 222 different sites in the Eastern Mediterranean, and involved four major types of tableware used during this period in the region. Our aim is to illustrate that simple social interactions, such has the frequencies of social interactions between cities can generate the pattern observed in the dataset.

To that end, the original economic model has been modified to reflect the particularities of the region and the time period of interest. In the first step, we run and record a series of experiments where the probability of social interactions between the cities varies. We show how the results of those simulations can be related to the pattern observed in the dataset.

 

 

History and archaeology, IT, computer science and mathematics