| Article ID: | iaor1994311 |
| Country: | Singapore |
| Volume: | 6 |
| Issue: | 1 |
| Start Page Number: | 77 |
| End Page Number: | 89 |
| Publication Date: | May 1989 |
| Journal: | Asia-Pacific Journal of Operational Research |
| Authors: | Shang Yi, Shao Heping |
This paper proposes a fast algorithm for linear programming which is based on direct approach to the saddle point of the Lagrangian function. The algorithm appears particularly befitting for linear problems of high dimension and for those with a great number of nonzero coefficients. Computational experience suggests that the algorithm converges much faster than the simplex method.