Powerdomains and zero finding

AbstractTraditionally, powerdomains have been used to provide models for various forms of nondeterminism in semantics. We establish a similar analogy between zero finding methods in numerical analysis and powerdomains: Different powerdomain constructions correspond to different types of behavior exhibited by numerical methods for zero finding. By combining this observation with the basic quantitative paradigm provided by measurement, a simple and uniform method for analyzing zero finding algorithms is obtained