Abstract:

As a typical complex system, man-land system is known as a coupled human and natural system. The complexity of man-land system can be divided into three critical dimensions - temporal complexity, spatial complexity and decision-making complexity. In modeling complex systems, traditional models are deficient in displaying data in multiple dimensions, and thus require additional research. Recent studies suggest that the agent-based models (ABM) would provide insights on exploratory analysis and serve as one of the key tools for complex system studies. In contrast to traditional models, ABM pays more attention to the study of 'people', focusing on assessing the influence of human activities on the environment, and can reflect it in a spatially explicit way. The models usually contain three parts: (1) environmental layer, which is composed of natural/social attribute such as terrain slope, land price and traffic condition; (2) agent layer, consisting of one or more agent types with specific attributes; (3) behavior rule, standardizing the mutual consultation and decision-making mode of agents. ABM adopts a 'bottom-up' approach by applying the relevant actors and decisions at the micro-level to producing an observable macro-phenomenon, and displaying high complexity values in three dimensions. Currently, ABM approaches are widely used to model human-environment interactions in various fields, including transportation, financial markets and tourism management. After the basic principles of agent-based simulation are briefly introduced, this paper reviews the application of ABM in ecological process, ecological resource management and land use/cover change. However, as a new method, ABM is still at an exploratory stage, faced by issues including replication potential, empirical parameterization and model validation, individual decision making, and integration with other models. Although there are many challenges, the recent developments reflect an encouraging trend towards developing a new methodology for dynamic spatial modeling of human-environment interactions. The outlook of ABM is promising.

Key words: man-land system, agent-based models, complexity, agent