T9 - z,x + 2 y-z,-2 x-y + 3 z 2-1) . Therefore, critical...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
ERG2018 Tutorial 9th November 10, 2008 Optimization of Unconstrained Multivariate Functions Critical Points and First Order Condition: f ( x ) = 0 . Classification of Critical Points and Second Order Conditions: A critical point is LocalMaximum, if eigenvalues of the Hessian matrix 2 f ( x ) are nonpositive; LocalMinimum, if eigenvalues of the Hessian matrix 2 f ( x ) are nonnegative; SaddlePoint, if eigenvalues of the Hessian matrix 2 f ( x ) have both strictly positive and negative elements. To find the eigenvalue of a matrix A , solve the equation system det ( λI - A ) = 0 . Example: Determine the critical points of f ( x,y,z ) = x 2 + xy + y 2 - 2 xz - yz + z 3 - z and classify them as local maxima, local minima, or saddle. 1. First, the gradient of f ( x,y,z ) is f ( x,y,z ) = (2 x + y - 2
Background image of page 1
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: z,x + 2 y-z,-2 x-y + 3 z 2-1) . Therefore, critical points ( x,y,z ) can be identified with equation: 2 x + y-2 z = 0 x + 2 y-z = 0-2 x-y + 3 z 2-1 = 0 Solving this, we get two group of solutions: (1 , , 1) and (-1 / 3 , ,-1 / 3) , which are the critical points of function f ( x,y,z ) . 2. The Hessian Matrix of f ( x,y,z ) is Hf ( x,y,z ) = 2 1-2 1 2-1-2-1 6 z . For the first critical point, the eigenvalues are 2 , 4-√ 10 , 4 + √ 10 , which are all positive, therefore it is a local minimum point. For the second critical point, the eigenvalues are-2 . 89 , 1 . 09 , 3 . 80 , therefore it is a saddle point. 1...
View Full Document

This note was uploaded on 11/13/2010 for the course SEEM SEEM2018 taught by Professor Chan during the Spring '10 term at CUHK.

Ask a homework question - tutors are online