Optimal and adaptive testing with cost constraints

Abstract

This paper generalizes our previous work on optimal and adaptive testing to consider a more general scenario of software testing resource constraints. The assumption is that software testing must be stopped once the allowed testing resources are used up. The contributions of this paper are as follows. First, we show that software testing with fixed resource constraints can be handled in the framework of the controlled Markov chains (CMC) approach to software testing. Second, an algorithm is adopted to reduce the computational complexity of on-line decision making in optimal and adaptive testing. Finally, the simulation results presented in this paper further confirm the effectiveness of the idea of adaptive testing in particular, and that of software cybernetics (which explores the interplay between software and control) in general.