Improved minimum-distance method for mining transition rules of cellular automata

A method was presented to retrieve transition rules of cellular automata (CA) using an improved minimum-distance method. The core of CA is how to define transition rules in a consistent way. In most situations, transition rules of CA are defined by using heuristic methods which may be subject to some uncertainties. For example, the multicriteria evaluation (MCE) may be used to determine the parameters of transition rules. However, this method can be influenced by expert preferences. The proposed modelis to define transition rules by using an improved minimum-distance method, in which the weights are determined by entropy. This method has been applied to the simulation of a fast growing city, Dongguan, in the Pearl River Delta. Comparison indicates that this method can produce better simulation accuracies than the general minimum-distance method. It model structure is transparent and can be easily understood by users.