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:
“The love of the famous, like all strong passions, is quite abstract. Its intensity can be measured mathematically, and it is independent of persons.”
—Susan Sontag (b. 1933)
“It was a cruel city, but it was a lovely one, a savage city, yet it had such tenderness, a bitter, harsh, and violent catacomb of stone and steel and tunneled rock, slashed savagely with light, and roaring, fighting a constant ceaseless warfare of men and of machinery; and yet it was so sweetly and so delicately pulsed, as full of warmth, of passion, and of love, as it was full of hate.”
—Thomas Wolfe (1900–1938)
“If we will admit time into our thoughts at all, the mythologies, those vestiges of ancient poems, wrecks of poems, so to speak, the world’s inheritance,... these are the materials and hints for a history of the rise and progress of the race; how, from the condition of ants, it arrived at the condition of men, and arts were gradually invented. Let a thousand surmises shed some light on this story.”
—Henry David Thoreau (1817–1862)