A testability growth model and its application

Testability growth is the enhancement in system testability through design modifications and/or other corrective actions performed throughout a system´s life cycle. A testability growth model can help system designers to plan and execute a testability progression process, and to achieve the specified system testability metrics in minimum time and/or cost. A Markov chain-based testability growth model that tracks and projects a user-defined composite testability growth metric is proposed. A Bayesian approach and a hybrid genetic algorithm, coupled with a particle swarm optimization method, are used to learn the parameters of the testability growth model from evolving data, and use the estimated model to track and project the testability metric. Validation of the theory is provided via simulated data. Results show that the testability growth model is reasonable, and the accuracy of the method is quite good.