Distributed Modeling Architecture of a Multi-Agent-Based Behavioral Economic Landscape (MABEL) Model

Abstract

The authors discuss a distributed modeling architecture in a multi-agent-based behavioral economic landscape (MABEL) model that simulates land-use changes over time and space. Based on agent-based modeling methodologies, MABEL presents a bottom-up approach to allow the analysis of dynamic features and relations among geographic, environmental, human, and socioeconomic attributes of landowners, as well as comprehensive relational schematics of land-use change. The authors adopt a distributed modeling architecture (DMA) in MABEL to separate the modeling of agent behaviors in Bayesian belief networks from task-specific simulation scenarios.Through a client-server infrastructure, MABEL provides an efficient and scalable decision request-response mechanism among heterogeneous agents, scenarios, and behavioral models. As an important part of the land-use change model, a market-bidding system and an adaptive land partition algorithm for land transactions are also discussed.