Reliability of N-version programming software with testing effort

Abstract

N-Version Programming (NVP) based software has become popular owing to its tolerance of design faults. In the present paper, we propose a Software Reliability Growth Model (SRGM) for NVP-based software systems along with detection and removal efficiency of the faults. When a software fault is detected during testing and debugging, debugging activities are executed to remove the fault. In this investigation, the influence of testing effort on the cost of the software development during testing phase is explored. The maintenance cost analysis is presented to determine the optimal testing time under cost and cost-reliability constraints. A numerical illustration is facilitated to perform the sensitivity analysis for the reliability indices of 3-VP system. We employ fuzzy interference system to use neuro-fuzzy technique which helps to examine the SRGM computationally by approximating the system parameters. Furthermore, comparative discussion on analytical results and approximate results is presented with the help of supervised learning process.