Article ID: | iaor20002436 |
Country: | Netherlands |
Volume: | 31 |
Issue: | 4 |
Start Page Number: | 479 |
End Page Number: | 487 |
Publication Date: | Mar 1999 |
Journal: | Engineering Optimization |
Authors: | Mazumder S.K., Das N.C., De Kajal |
Keywords: | information theory |
The paper presents an information-theoretic algorithm for solving constrained non-linear programming problems based upon the principle of minimum cross-entropy and surrogate mathematical programming. It is an alternative to the earlier work by Templeman and Li based on the maximum entropy principle and so far as the rate of convergence of the solution is concerned it is an improvement upon it. The merit of the present technique is illustrated with two numerical examples.