A wavelet-balance approach for steady-state analysis of nonlinear circuits

In this paper, a novel wavelet-balance method is proposed for steady-state analysis of nonlinear circuits. Taking advantage of the superior computational properties of wavelets, the proposed method presents several merits compared with those conventional frequency-domain techniques. First, it has a high convergence rate O(h/sup 4/), where h is the step length. Second, it works in time domain so that many critical problems in frequency domain, such as nonlinearity and high order harmonics, can be handled efficiently. Third, an adaptive scheme exists to automatically select proper wavelet basis functions needed at a given accuracy. Numerical experiments further prove the promising features of the proposed method in solving steady-state problems.