J. W. Cangussu (USA)
Software test process, state variable model, parameter cali bration, system identification, feedback control.
This paper reports an evaluation of the calibration algo rithm for the state variable model of the software test pro cess. The data to evaluate the convergenceof the estimation of the parameters of the model is generated using simula tion. The analysis exposes that even under the assumption of an initial over/under estimation of the parameter, the al gorithm presents a fast convergence. The results show that in almost 90% of the cases the algorithm uses data from a short period of time to obtain an accuracy of, at least, 15% in the prediction of the time to reach a desired level in the error reduction. These outcomes show the efficiency of the calibration algorithm and further demonstrate the applica bility of the state variable model.
Important Links:
Go Back
