Abstract: The hidden neurous in quantum neural network adopt the multi-level actioation function;so,with fixed fuzziness,the quantum network can reasonably distribute the uncertain data of a decision into each mode,thus to reduce the uncertainty of mode recognition and raise the accuracy of the mode recognition.The soft fault diagnosis of tolerance simulated circuit based on wavelet and quantum neural network is proposed.The simulation experiment shows that,compared with BP neural network,the quantum neural network can increase the fault diagnosis accuracy to 24.78%.

Key words: quantum neural network( QNN), quantum interval, wavelet analysis, tolerance analysis