On regularization of lexicographic problems by the method of minimum concessions

✍ Scribed by V.N. Nefedov

• Book ID: 104263209
• Publisher: Elsevier Science
• Year: 1983
• Weight: 703 KB
• Volume: 23
• Category: Article
• ISSN: 0041-5553
• DOI: 10.1016/s0041-5553(83)80072-x

✦ Synopsis

For comparison we give below the results of solving the same problem by the method with programmed step control.

The initial approximation was the same.

In the asymptotically optimal /8/ programmed step control O,=|/l(s+a) , the parameter I must be equal to the least eigenvalue of the Hessian of the function /(x), i.e., /='/30; we select a=|0. The results for the averaged values of the coordinates and function at the 161-st to 170-th iterations are: 5,=39.9574, E2=6.9554, E,=t.5i52, E~=42.3507, E~=24.0324, 7(x)=98.5438.

If, in the programmed step control, we take different values of the parameters I and a, say, I=I~ a=|, then, averaging the data from the 161-st to 170-th iterations, we have ~,=