《《1998 A Globally Convergent Primal-Dual Interior Point Method for Constrained Optimization》.pdf

A Globally Convergent Primal-Dual Interior Point Metho d for Constrained Optimization Hiroshi Yamashita Abstract This pap er prop oses a primal-dual interior p oint metho d for solving general nonlinearly constrained optimization problems. The metho d is based on solving the barrier Karush-Kuhn-Tucker conditions for optimality by the Newton metho d. To globalize the iteration we intro duce the barrier-p enalty fucntion and the opti- mality condition for minimizing this function. Our basic iteration is the Newton iteration for solving the optimality conditions with resp ect to the barrier-p enalty function which coincides with the Newton iteration for the barrier Karush-Kuhn- Tucker conditions if the p enalty parameter is suciently large. It is proved that the metho d is globally convergent from an arbitrary initial p oint that strictly satises the b ounds on the variables. Implementations of the given algorithm are done for small dense nonlinear programs . The metho d solves all the problems in Ho ck and Schittkowskis textb o ok eciently. Thus it is shown that the metho d given in this pap er p ossesses a go o d theoretical convergence prop erty and is ecient in practice. Key words: interior p oint metho d, primal-dual metho d, constrained optimization, non- linear programming  Mathematical Systems, Inc., 2-4-3, Shinjuku, Shinjuku-ku, Tokyo, Japan hy@msi.co.jp 1 1 Intro duction In this pap er we prop ose a primal-dual interior p oint metho d that solves general nonlin- early constrained optimization problems. To obtain a fast algorithm for nonlinear opti- mization problems it is fa