Strong Wolfe Condition On Curvature
The Wolfe conditions, however, can result in a value for the step length that is not close to a minimizer of . If we modify the curvature condition to the following,
- iia)
then i) and iia) together form the so-called strong Wolfe conditions, and force to lie close to a critical point of .
The principal reason for imposing the Wolfe conditions in an optimization algorithm where is to ensure convergence of the gradient to zero. In particular, if the cosine of the angle between and the gradient,
is bounded away from zero and the i) and ii) hold, then .
An additional motivation, in the case of a quasi-Newton method is that if, where the matrix is updated by the BFGS or DFP formula, then if is positive definite ii) implies is also positive definite.
Read more about this topic: Wolfe Conditions
Famous quotes containing the words strong, wolfe and/or condition:
“Oh, yes, I’d do it all again; the spirit is willing yet; I feel the same desire to do the work but the flesh is weak. It’s too bad that our bodies wear out while our interests are just as strong as ever.”
—Susan B. Anthony (1820–1906)
“On Wall Street he and a few others—how many?—three hundred, four hundred, five hundred?—had become precisely that ... Masters of the Universe.”
—Tom Wolfe (b. 1931)
“My profession brought me in contact with various minds. Earnest, serious discussion on the condition of woman enlivened my business room; failures of banks, no dividends from railroads, defalcations of all kinds, public and private, widows and orphans and unmarried women beggared by the dishonesty, or the mismanagement of men, were fruitful sources of conversation; confidence in man as a protector was evidently losing ground, and women were beginning to see that they must protect themselves.”
—Harriot K. Hunt (1805–1875)