| Software reliability is one of the most important indicators of software quality.With theevolution of software engineering and software development mode, a rapid progress tookplace in software reliablility models. Non-Homogeneous Poisson Process software reliabilitymodel became a research focus. In order to make the software reliability model capable ofguiding software development and providing reliable software quality assurance, more andmore research focused on the software index distribution technology. This thesis propose anew method of reliability index distribution and prediction model.In the early period of software development, with use of Musa execution time model, anew method of software reliability index distribution was proposed in this paper. Andcorporating with image processing technology, software reliability index can turn into color, itcan provide direct observation to the software engineers in software development process.Then analysed the current situation of Non-Homogeneous Poisson Process software reliabilitygrowth models in domestic and abroad. With the analysis of the assumed condition, aimproved model was proposed. Total fault number, fault detection rate and troubleshootingrate were considered as time functions in this model. So it can be more practical and capablefor the actual operation aspect. With the improvement of software reliability growth models,the fitting effect of forecasting the failure rate has been inhanced.This article use the software reliability index distribution technology combined with adeposit banking system, and set the mapping between reliability index and color in imageprocessing technology first. Then introduces the previous research and concepts involved, andraise the improvement of assumed condition of Goel-Okumoto model (one of the mostfamous Non-Homogeneous Poisson Process software reliability growth models), proved itsaffect finally.Main research and results in this essay includes,(1) Take advantage of Musa execution time model with software reliability indexdistribution of component base software system. Assess software reliability indexwith Musa execution time model, and calculate the distribution weight. Make pertinent modification to Musa execution time model.(2) Map software reliability index into gray scale,then turn gray scale into hue value ofthe HSI model. Through fix the saturation and intensity of HSI model, calculate thevalue of RGB model. Achieve the realization of mapping from software reliabilityindex to specific color.(3) Introduce color processing technology to fields of software reliabilty indexdistribution. By research on atmosphere warning mechanism and chromatology, andthe feature of human vision system, continous function curves from softwarereliability index to hue value of HSI model was proposed.(4) Integrate and develop current approaches of G-O model improvement. Proceed withimprovement of G-O model’s assumed condition.Consider total fault number,faultdetection rate and troubleshooting rate as time functions, achieve a furtherenhancement of G-O model. |
PDF Full Text Request