Numerical Techniques for Finding Estimates which Minimize the Upper Bound of the Absolute Deviation