Study on some combined forms of fuzzy neural network system and their application on simulation modeling

摘要

Fuzzyj, neural network (NN) and fuzzy neural network (FNN) technique have been attached more and more importance in increasing research fields. Their application research is also in progress. In the system simulation field, these techniques are similarly studied in theory and applied to practical projects widely. In this paper, four combined forms of NN and fuzzy system are discussed with respect to their merits and demerits. Three FNNs whose consequence is singleton are discussed in details and their learning algorithms are derived. The FNNs' modeling precision is studied by modeling a ballistics model with air resistance. It is concluded that FNN is a better approximator of nonlinear function than NN.