Safavi - Dept. of Electrical Engineering, Shiraz University Shiraz,I.R. of Iran
Zakeri - Dept. of Electrical Engineering, Shiraz University Shiraz,I.R. of Iran

The technologicala dvancemenitn the computerc ontrolled systems and other areas in the process induitries makes availablec ontinuottss tream ofdata to be exploitedu tilizing on-line systems environment. This paper investigates procedures for on-line learning and improvement with 'v)ave-nets utilizing such a stream of data. l[/ave-nets are wavelet-based neural networks with localized and hierarc hical multi-resoltttion Iearning. The multi-res oluti o n framework of wave-net and its learning algorithms are briefly explained. Then based on the learning algorithms, pr o c e dur e s for o n- I ine w av e -ne t impr ovements are proposed and the verifications are performed with some case studies The results show effectivenesso f the melhods.

Wave-Netso, nline learning,n eural networks, CSTR