An interior-point method for finding a DC solution of nonlinear circuits

Some problems in DC nonlinear circuits and steady state simulations demonstrate slow convergence using the global Newton-Ralphson (NR) method. This paper presents a new technique for dealing with the convergence problems that Newton-Ralphson based methods encounter. This method is based on an interior-point concept and an affine scaling used in the linear programming. Issues related to the implementation of the algorithm in the content of solving nonlinear equations are discussed. This algorithm is then compared to the Newton-Ralphson based solver currently used in some circuit simulators. Using various examples, the comparison shows improved convergence behavior using this algorithm.