| A. Lodi. Personal communication. |
....which variables will contribute to the discrepancy. In the worst case they can be the k variables having the lowest c # i values. Then the best (i.e. the highest) value among these can be chosen, which corresponds to L[k] A better bound improvement can be achieved by the following theorem [22]. n , LB 0 i=1 L[i] is a valid lower bound for the subproblems corresponding to discrepancy k. Proof. The proof is based on the concept of additive bounding procedures [6, 7] that states as follows: first we solve a relaxation of a problem P . We obtain a bound LB, in our case LB 0 ....
A. Lodi. Personal communication.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC