Modélisation de type multi-agents en archéologie : l'expansion des premiers agriculteurs Balkaniques : adaptation du modèle OBRESOC : manipulation et exploration des données simulées

Abstract : A topic of great importance in archaeological research throughout the last decades concerns the expansion of the first farmers from Anatolia through the Balkans. The standard archaeological approaches allowed the understanding of the path and timing of this expansion; however, they lack explanation of what is unobservable in the archaeological record: in particular, the socio-economic structure of a prehistoric farming society. Throughout this thesis, an agent-based model was built in order to explore those elements which are hidden in archaeology. This model, called BEAN (Bridging European and Anatolian Neolithic), is an adaptation of the OBRESOC model (Un OBservatoire REtrospectif d'une SOCiété archéologique). OBRESOC was created to simulate the expansion of the LBK farmers in central Europe, and was adapted to the Balkan archaeological context. The expansion of the first Neolithic farmers in the Balkans was simulated by combining the archaeological records to ethnohistoric and paleodemographic inferences. A realistic environment has been modelled where the areas of optimum farming are determined by meteorology and soil fertility estimates. An agent corresponds to a household; agents interact on this landscape, following socioeconomic partial intermediate models. For instance: households composed of a nuclear family; intensive farming system on small plot completed by hunting-gathering; expansion determined by scalar stress at the hamlet scale; family clan solidarity; shortages and famines caused by meteorological events). Thus, the model simulates the functioning of the Neolithic farming society and its geographic expansion. Several simulations have been executed, testing different combinations of the key parameters, identified through a sensitivity analysis. The goodness of fit of simulated data to the archaeological data is measured mostly on geographic criteria : the best simulation is the one that produces the expansion pattern that better fits to the archaeological data. Specific procedures have been developed in order to process the large amount of data produced by the model. The observation of this data permitted to explore some aspects that are invisible in archaeological record : for example, the model helped to investigate some archaeological beliefs, based on assumptions that could not be verified. The model also permitted the exploration of other topics, such as the comparison between the pioneer front of colonization and the zones of previous occupation, as well as the effect of meteorology on the expansion of the farming system. The model produced an expansion pattern that corresponds geographically and chronologically to the expansion suggested by the archaeological evidence. The exploration of socio-economic outputs permitted the formulation of new hypothesis that could not be made using purely archaeological record. Even when there's a large gap between what is found in archaeology and what is produced by the model, this agent-based modelling approach helps to raise new questions, adding new ideas and perspective to the actual state of research.
Complete list of metadatas

Cited literature [303 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-02102276
Contributor : Abes Star <>
Submitted on : Wednesday, April 17, 2019 - 10:58:11 AM
Last modification on : Friday, April 19, 2019 - 1:30:34 AM

Identifiers

  • HAL Id : tel-02102276, version 1

Collections

Citation

Andrea Zanotti. Modélisation de type multi-agents en archéologie : l'expansion des premiers agriculteurs Balkaniques : adaptation du modèle OBRESOC : manipulation et exploration des données simulées. Archéologie et Préhistoire. École pratique des hautes études - EPHE PARIS, 2016. Français. ⟨NNT : 2016EPHE3057⟩. ⟨tel-02102276⟩

Share

Metrics

Record views

67

Files downloads

55