Considering Signal Constraints in Search-Based Testing of Continuous Systems

This paper presents a technique to consider constraints on input signals of continuous systems when applying search-based testing. The signal constraints are described using a logic based on Signal Temporal Logic. We developed a distance-oriented evaluation technique for these constraints that provides an exact rating of the amount of constraint violation, thus allowing a ranking of the generated solutions in terms of constraint violation. An adaptive penalty function is then used to incorporate the evaluation results into the optimization. Finally, the overall method is shown to be capable of considering signal constraints appropriately when experimentally applied to search-based black-box testing of a MATLAB SIMULINK model of an automatic transmission controller.