Applications of Continuous Amortization to Bisection-based Root Isolation

Continuous amortization is a technique for computing the complexity of algorithms, and it was first presented by the author in [6]. Continuous amortization can result in simpler and more straight-forward complexity analyses, and it was used in [6, 5, 52] to provide complexity bounds for simple root isolation algorithms. This paper greatly extends the reach of continuous amortization to serve as an overarching technique which can be used to compute complexity of many root isolation techniques in a straight-forward manner. Additionally, the technique of continuous amortization is extended to higher dimensions and to the computation of the bit-complexity of algorithms. In this paper, six continuous amortization calculations are performed to compute complexity bounds (on either the size of the subdivision tree or the bit complexity) for several algorithms (including algorithms based on Sturm sequences, Descartes’ rule of signs, and polynomial evaluation); in each case, continuous amortization achieves an optimal complexity bound.