ESTIMATING AGE REPLACEMENT POLICIES FROM SMALL SAMPLE DATA

AbstractIn the present paper, we consider the typical age replacement models to minimize the relevant expected costs, and formulate the statistical estimation problems with the complete sample of failure time data. Based on the concept of total time on test statistics, we show that the underlying optimization problems are translated to the graphical ones on the data space. Next, we utilize a kernel density estimator and improve the existing statistical estimation algorithms in terms of convergence speed. Throughout simulation experiments, it can be shown that the developed algorithms are useful especially for the small sample problems, and enable us to estimate the optimal age replacement times with higher accuracy.