Modeling dynamic urban growth using cellular automata and particle swarm optimization rules

This paper presents an improved cellular automata (CA) model of urban growth based on particle swarm optimization (PSO) approach with inertia weight. An innovative feature of this cellular model is the incorporation of swarm intelligence to stochastically optimize the transition rules to reduce the simulation uncertainties and improve its locational accuracy in urban modeling. The similarity between the nature of self-organization of particle swarm optimizers and the bottom-up approach of cellular automata makes PSO particularly suitable to search for the global optimum parameters of CA transition rules. The CA parameters retrieved by the PSO technique are able to express precisely the contributions of various driving forces to urban growth; hence an effective cellular model can be realized for modeling urban dynamics. The PSO based CA model was applied in Fengxian District of Shanghai Municipality, eastern China, to simulate the spatio-temporal process of urban growth from 1992 to 2008 at 30 m spatial resolution. The simulation outcomes, evaluated with error matrix and simulation accuracies, demonstrate that the PSO-CA model outperforms other spatial statistical based CA models.

► We apply a particle swarm optimization (PSO) algorithm to optimize the transition rules of an urban CA model.
► The optimized CA parameters quantify the contributions of various driving factors to urban growth.
► A case study demonstrates that the PSO-CA model performs better than a logistic regression based CA model.