Analyzing the effects of stochastic perturbation and fuzzy distance transformation on Wuhan urban growth simulation

Abstract Stochastic perturbation and fuzzy distance transformation (FDT) have been widely introduced into the derivation of transition rules to improve the simulation capability of urban cellular automata (CA) models. However, their effects on urban growth simulation have not been revealed. In this article, we compare their effects on urban growth simulation for the city of Wuhan by a sensitivity analysis of the simulation accuracy and simulated urban patterns. We reveal the relationship between the two components and propose an optimized stochastic variable to improve the simulation capability of urban CA models. The results show that the stochastic perturbation is the main factor affecting the simulation accuracy and simulated urban patterns of Wuhan, in that it significantly reduces the simulation accuracy and increases the degree of fragmentation and the proximity and shape complexity of the simulation results. In addition, the change caused by the stochastic perturbation increases with its intensity. Although FDT can restrain the effect of the stochastic perturbation, its inhibitory effect becomes negligible as the stochastic perturbation intensity increases. The optimized stochastic variable can significantly reduce the negative impact of the original stochastic variable, and its improvement is more obvious when the stochastic perturbation intensity is higher.