A hybrid GA-PSO approach for reliability under type II censored data using exponential distribution

Abstract

Today software plays an important role and its application is used in each and every domain. In software testing phase, quality risk, reliability and fault detection are used to analysis and remove the failure of the item. Owing to time constraints and limited number of testing unit, we cannot fix the experiment until the failure. In order to observe the failure during testing phase, censoring becomes significant methodology to estimate model parameters of exponential distributions. The most common censoring schemes do not have the flexibility to identify the failure in the terminal point. The most commonly used censoring schemes are Type I and Type II censoring schemes. To identify the optimum censoring scheme and overcome these problems optimal technique is used in this paper. Thus, optimal scheme will improve the output of testing phase with the aid of specific optimal constraints. Entropy and variance are used as optimal criterion. Determination of Optimal schemes will be done by Hybrid Genetic Algorithm (GA) and Particle Swarm Optimization (PSO). The risk values are estimated of the selected optimal censoring scheme. 2482 K. Kalaivani and S. Somsundaram