An AND-OR Fuzzy Neural Network and Ship Application
Jianghua Sui, Guang Sheng Ren · 2006
A novel multilayer feed-forward AND-OR fuzzy neural network (AND-OR FNN) and a piecewise optimization approach are proposed in this paper. The equivalent is proved between the architecture of AND-OR FNN and fuzzy weighted Mamdani inference. The main superiority is shown in not only reducing the input space by special inner structure of neurons, but also auto-extracting the rule base by the structure optimization of network. The optimization procedure consists of two phases, first the blueprint of network is reduced by GA (genetic algorithm) and PA (pruning algorithm); the second phase, the parameters are refined by ACO (ant colony optimization). The AND-OR FNN ship controller system is designed based on input-output data to validate this method. Simulated results demonstrate that the number of rule base is decreased remarkably, the performance is much better than ordinary fuzzy control and the approach is practicable, simple and effective.