Combinatorial Sequence Testing Using Behavioral Programming and Generalized Coverage Criteria

Abstract

This paper tackles three main issues regarding test design: (1) it proposes a new way to model what to test; (2) it offers a framework for specifying coverage criteria that generalizes previous types of coverage; (3) it outlines a Bayesian approach to an informed exploitation-exploration balance in the context of testing.In addition to the theoretical contribution, we present an empirical evaluation with a proof-of-concept tool that we have developed to support the conceptual advantages and to illustrate practical benefits.