Loading…
CAA Conference has ended

Thursday, March 29 • 2:00pm - 4:00pm
1 - Applying the ODD protocol in agent-based modeling of past socio-ecological dynamics

Sign up or log in to save this to your schedule and see who's attending!

The paper proposes a critical reflection on the use of the ODD (Overview, Design concepts, and Details) protocol (Grimm et al. 2010) to model and simulate past socio-ecological dynamics. We present the use of the ODD protocol to implement an agent-based model (ABM) within a specific case study: hunter-gatherer persistence in arid margins, the case of Northern Gujarat (India). We choose this specific case study as in North Gujarat hunter-gatherer presence is attested for a long period of time (7th to 2nd mill. BC) and it is postulated that they have co-existed with communities of pastoralists and agriculturalist. The aim of the proposed case study is to model resource management and decision making among hunter-gatherer groups in Northern Gujarat for exploring evolutionary trajectories in relation to a) the appearance of other specialized groups (i.e. pastoralists and agriculturalists) during the mid-Holocene and b) environmental variability. As complex system theory becomes a fundamental framework to understand the interactions between ecological and social dynamics, the purpose is always to reduce the complexity in a way that makes the problem tractable. Such a reduction can be achieved by improving model assumptions, by elimination of variables, reducing the relevant subsystems, etc. In any case such a reduction of complexity has implications for the validity and the precision of the theoretical findings. To ensure it, stringent protocols are needed to communicate and make explicit the reduction in a consistent and coherent way. The ODD protocol is one of the available solutions for such formalization. The presented case study is developed within a research project (SimulPast - www.simulpast.es) that aims at developing a transdisciplinary methodological framework to model and simulate ancient societies and their relationship with environmental transformations. This project integrates people from several fields (archaeology, anthropology, sociology, computer science, environmental studies and physics), and communication is a fundamental object within the project. In this framework, using the ODD protocol can help the integration researchers coming from different fields in those case-studies where ABM is applied. The ODD protocol was initially proposed in 2006 to standardize the published descriptions of ABM (Grimm et al. 2006). A recent review of ODD models by Grimm et al. (2010) illustrates their main advantages in ABM, namely i) to improve the rigorous formulation of models and ii) to make their theoretical foundations more visible. That is, to make model descriptions more understandable and complete and ABMs less subject to criticism for being irreproducible. The ODD is increasingly accepted for social simulation, supported by domain expert communities (e.g OpenABM) and used for publication in scientific journals (e.g. ecological modeling, JASSS). However, the ODD protocol remains unused in archaeology. We argue that the ODD protocol (with some improvements to better adapt it at social simulations) represents a useful tool to support a dialectical approach in agent-based modelling of past socio-ecological dynamics between specialists coming from different background, supporting in particular: a) transfer from conceptual to formal models, b) semantic understanding among people with different expertise, c) reduction of complexity in a coherent and consistent way, d) model replication, modification and extension.


Speakers
CL

Carla Lancelotti

CASES, IMF and CCHS Spanish National Research Council (CSIC)
BR

Bernardo Rondelli

Institució Milà i Fontanals, Spanish National Research Council (IMF-CSIC)
XR

Xavier Rubio-Campillo

Barcelona Supercomputing Centre - Centro Nacional de Supercomputación | Twitter: @xrubiocampillo


Thursday March 29, 2012 2:00pm - 4:00pm
Building 65, Lecture Theatre C

Attendees (0)