Simulation of urban growth using agent-based modeling and game theory with different temporal resolutions

Population increase and industrialization, particularly in developing countries, has led to the rapid growth of urban areas. To adequately direct urban growth, provide access to urban facilities and preserve the environment, accurate planning is required. As part of urban planning, computer-based simulation models can be used to predict the natural growth of urban areas. The objective of this research is to develop an agent-based simulation model, with different temporal resolutions and reinforced by game theory, to predict the growth of Zanjan city between 2011 and 2016. In the model, three types of land developer, classified according to their income level, are considered as agents, with different attributes and behaviors. Agents search the cellular environment, collect data and select proper sites for development in accordance with their criteria and preferences. When several agents select a single cell for development, their competition is modeled by game theory. The agent-based model is implemented and tested in two scenarios: with game theory and without it. In addition, based on the fact that urban growth is a spatio-temporal phenomenon, different temporal resolutions of 6, 12, 30 and 60 months are considered in the implementation. Kappa statistics, Percent Correct Match, and figure of merit are used to compare the results of modeling scenarios with the urban map of 2016. The best result of 81.1%, 97.32% and 6.38% for Kappa statistics, Percent Correct Match and figure of merit are obtained, when game theory and 6 months resolution is used. The results showed that the usage of game theory and higher temporal resolution have both positive effects on the accuracy of the model. With higher temporal resolutions, the gradual developments of neighboring cells can well be considered by agents. Using game theory can help in modeling the competition-based interactions and behavior of agents.