Technique for polynomial factorisation employing optimised iterative method

An iterative procedure based on the Newton-Raphson process for the factorisation (splitting) of a real-coefficient polynomial into two lower-degree polynomials is reviewed, and the method extended to the case where the Jacobian matrix is not updated at each iteration (a ´quasi-Newton´ process). This can result in a faster process. In the case of a particular class of problems, an optimal rate of update of the Jacobian matrix is determined, which results in a further reduction in computation time. This represents a significant improvement in performance, especially if the algorithm is to be used as part of an iterative network design package.