Estimation of Ready Queue Processing Time using Efficient Factor Type Estimator (E-F-T) in Multiprocessor Environment

The ready queue processing time estimation problem appears when many processes remain in the ready queue after the sudden failure. The system manager has to decide immediately how much further time is required to process remaining jobs in the ready queue. In lottery scheduling, this prediction is possible with the help of sampling techniques. Ratio method, existing in literature of sampling, was previously used by authors to predict the time required provided highly correlated source of auxiliary information is available and used. This paper suggests two new estimation methods which are compared in terms of estimating the total processing time. Under large sample approximation, the bias and m.s.e of proposed estimators have been obtained in the set up of random sampling applicable to lottery scheduling. Performance of both is compared in terms of mean squared error. The confidence intervals are calculated for the estimate and they provide strong numerical support to the theoretical findings.