On the improvement per iteration in Karmarkar's algorithm for linear programming (Q918862)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | On the improvement per iteration in Karmarkar's algorithm for linear programming |
scientific article |
Statements
On the improvement per iteration in Karmarkar's algorithm for linear programming (English)
0 references
1990
0 references
Let \(\delta_ n(\alpha)\) be the greatest possible guaranteed potential drop at an iteration of Karmarkar's projective method for linear programming problems in dimension n, when the method uses step length parameter \(\alpha\). The paper shows that \(\delta_ n(\alpha)\) equals \(\ln (1+\alpha)-\ln (1-\frac{\alpha}{(n-1)})\) for n sufficiently large.
0 references
potential function
0 references
Karmarkar's projective method
0 references