On the Exploitation of Parallelism in a Test Generation Method for LOTOS-Specifications

Abstract

A major problem in the field of automatic test derivation from LOTOS and other formal description techniques is that resulting complete tests become too large to be useful in practice. This paper describes how a new automatic test derivation method exploits parallelism in the specification to overcome this problem. The method uses a fault model in order to define possible faults that the test has to investigate. Thus the method avoids the well-known state space explosion problem, and it, therefore, generates short test sequences with high fault coverage. The length of the test sequence grows only linear with the number of subbehaviours in a parallel expression.