lecture21 - CSE 6740 Lecture 21 How Do I Optimize With...

Info iconThis preview shows pages 1–12. Sign up to view the full content.

View Full Document Right Arrow Icon
CSE 6740 Lecture 21 How Do I Optimize With Constraints? (Constrained Optimization) Ravi Sastry and Alexander Gray agray@cc.gatech.edu Georgia Institute of Technology CSE 6740 Lecture 21 – p. 1/3 4
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Today 1. Convex (Constrained) Optimization Problems 2. Convex (Constrained) Optimization Methods 3. SMO: An Active Set Method CSE 6740 Lecture 21 – p. 2/3 4
Background image of page 2
Convex (Constrained) Optimization Problems Problems with a convex objective function and, if there are constraints, convex constraints. CSE 6740 Lecture 21 – p. 3/3 4
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Linear Programming When the objective and constraint functions are all affine, the problem is called a linear program (LP), which has the form Find x * = arg min x R D c T x + d (1) subject to Gx h (2) Ax = b. (3) CSE 6740 Lecture 21 – p. 4/3 4
Background image of page 4
Quadratic Programming When the objective function is quadratic and the constraint functions are affine, the problem is called a quadratic program (QP), which has the form Find x * = arg min x R D 1 2 x T Px + q T x + r (4) subject to Gx h (5) Ax = b. (6) CSE 6740 Lecture 21 – p. 5/3 4
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Quadratically Constrained QP If the constraints are also quadratic, the problem is called a quadratically constrained quadratic program (QCQP): Find x * = arg min x R D 1 2 x T Px + q T x + r (7) subject to 1 2 x T P i x + q T i x + r i 0 , i = 1 ,... ,M (8) Ax = b. (9) CSE 6740 Lecture 21 – p. 6/3 4
Background image of page 6
Second-order Cone Programming A closely related problem is called a second-order cone program (SOCP), which has the form Find x * = arg min x R D f T x (10) subject to || A i x + b i || 2 c T i x + d i , i = 1 ,... ,M (11) Fx = g. (12) A constraint of this form is called a second-order cone constraint. CSE 6740 Lecture 21 – p. 7/3 4
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Geometric Programming A geometric program (GP) is a problem of the form Find x * = arg min x R D f ( x ) (13) subject to c i ( x ) 1 , i = 1 ,... ,M (14) d i ( x ) = 1 , i = 1 ,... ,N (15) (16) where f and the c i have the form log p s k e a T 1 k x + b ik P (17) and the d i have the form e g T i x + h i . CSE 6740 Lecture 21 – p. 8/3 4
Background image of page 8
Semidefinite Programming A semidefinite program (SDP) has the form Find x * = arg min x R D c T x (18) subject to x 1 F 1 + ... + x n F n + G 0 (19) Ax = b. (20) CSE 6740 Lecture 21 – p. 9/3 4
Background image of page 9

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Relationships Generality: LP < QP < QCQP < SOCP < SDP. Computational cost: LP < QP < QCQP < SOCP < SDP. CSE 6740 Lecture 21 – p. 10/3 4
Background image of page 10
Optimization Methods The interior-point method. CSE 6740 Lecture 21 – p. 11/3
Background image of page 11

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 12
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 04/03/2010 for the course CSE 6740 taught by Professor Staff during the Fall '08 term at Georgia Institute of Technology.

Page1 / 34

lecture21 - CSE 6740 Lecture 21 How Do I Optimize With...

This preview shows document pages 1 - 12. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online