In this paper, a Bonferroni confidence interval for the failure rate of the inverse Gaussian distribution is derived. It can be solved by formulating the problem as a nonlinear programming. A numerical example on repair time data is provided to illustrate the procedure.
