How testable are BDI agents? An analysis of branch coverage

Abstract

Before deploying a software system, it is important to assure that it will function correctly. Traditionally, this assurance is obtained by testing the system with a collection of test cases. However, since agent systems exhibit complex behaviour, it is not clear whether testing is even feasible. In this paper we extend our understanding of the feasibility of testing BDI agent programs by analysing their testability with respect to the all edges test adequacy criterion, and comparing with previous work that considered the all paths criterion. Our findings include that the number of tests required with respect to the all edges criterion is much lower than for the all paths criterion. We also compare BDI program testability with testability of (abstract) procedural programs.