MIT 公开课 - 高等算法 lec9.pdfVIP

MIT OpenCourseWare 6.854J / 18.415J Advanced Algorithms Fall 2008 �� For information about citing these materials or our Terms of Use, visit: /terms. 18.415/6.854 Advanced Algorithms October 6, 2008 Lecture 9 Lecturer: Michel X. Goemans 9 Linear Programming Linear programming is the class of optimization problems consisting of optimizing the value of a linear objective function, subject to linear equality or inequality constraints. These constraints are of the form a x + · · · + a x {≤, =, ≥} b, 1 1 n n where a , b ∈ R, and the goal is to maximize or minimize an objective function of the form i c x + · · · + c x . 1 1 n n In addition, we constrain the variables xi to be nonnegative. The problem can be expressed in matrix form. Given these constraints Ax {≤, =, ≥} b x ≥ 0, maximize or minimize the value of T c x, where x ∈ Rn , A ∈ Rm ×n , b ∈ Rm , c ∈ Rn . Linear programming has many applications and can also be used as a proof technique. In addition, it is important from a complexity point-of-view, since it is among the hardest of the class of polynomial-time solvable problems. 9.1 Algorithms Research in linear programming algorithms has been an active area for over 60 years. In this class, we will discuss three major (classes of ) algorithms: • Simplex method (Dantzig 1947). – Fast i