Decidable Semantics
Dr. Anderson has recently become interested in the topic of decidable semantics and how it can be used to model the behaviors of large-scale software systems. In the past, programs with decidable semantics have been limited to “toy problems” of limited use. However, there are application domains involving large software systems in which the behavior of the software system is decidable, that means, given the system's inputs, it is possible to predict what the system's output will be. Dr. Anderson is currently working with a software company that is bringing together a variety of techniques such as constraint logic programming, subsumption reasoning, and numerical methods, to make it possible to model the decidable semantics of large software systems from particular application domains (such as finance). The research topics of this project involve the classic tradeoff between the expressivity of a specification language and the decidability of its semantics and how subsumption reasoning techniques can aid the software quality assurance process of software systems.