On finding primal and dual optimal bases英文文献.pdfVIP

On Finding Primal and DualOptimal Bases Nimro d Megiddo revised June Abstract We show that if there exists a strongly p olynomial time algorithm that nds a basis which is optimal for b oth the primal and the dual problems given an optimal solution for one of the problems then there exists a strongly p olyno mial algorithm for the general linear programming problem On other hand we give a strongly p olynomial time algorithm that nds such a basis given any pair of optimal solutions not necessarily basic for the primal and the dual problems Such an algorithm is needed when one is using an interior p oint metho d and is interested in nding a basis which is b oth primal and dualoptimal Sub ject classication Programming Linear Algorithms and Theory Intro duction The reader is referred for example to for information ab out standard results in linear programming which are used in this work The simplex metho d for linear programming has the nice prop erty that if the problem has an optimal solution then a basis is found which is b oth primaloptimal and dualoptimal ie b oth of the basic solutions which are dened by such a basis in the primal and the dual problems are optimal in their resp ective problems For brevity we call such a basis an optimal basis An optimal basis IBM Research Almaden Research Center San Jose CA and Scho ol of Mathematical Sciences Tel Aviv University Tel Aviv Israel is a useful ob ject for p ostoptimality analysis The known p olynomialtime algorithms for linear programminggenerate optimal solutions which are not necessarily basic Given any primaloptimal solution not necessarily basic it is easy to nd a primaloptimal ba