Bounded Error Parameter Estimation: A Sequential Analytic Center Approach

E. Bai, Y. Ye and R. Tempo

In this paper, a sequential analytic-center approach for bounded error parameter estimation is proposed. The analytic center minimizes the ``average'' output error and allows an easy-to-compute sequential algorithm. With little computational effort two ellipsoids centered at the analytic center can be obtained as well: One inscribes and the other outscribes the so-called membership set. Finally, a sequential algorithm is presented along with its convergence proof. The complexity of this algorithm in terms of the maximum number of Newton iterations is shown to be {\em linear} in the number of observed data points. The result is is new in the literature of linear programming as well.

Working Paper, Dept. of Management Science, University of Iowa, IA, 52242, USA.