IMPLEMENTATION OF BRANCH AND BOUND METHOD FOR CONVEX OPTIMIZATION PROBLEM

Victor Hariadi, Rully Soelaiman

Abstract


Convex optimization is one of important areas in nonlinear programming, both from
the mathematical and application viewpoints. Numerous real world problems are
naturally expressed as quadratic programming problems. In non linear programming,
it is necessary to use a method to get global optimal value in order to not get trapped in
local optimal value. Branch and bound method in non linear programming known as
one of global optimization technique.In this research we implement branch and bound
method with relaxation process in the problem to get global optimization from convex
programming. The boundary of sub problem will be considering using reformulation
from quadratic with KKT condition. Experiment was accomplished on quadra1tic
programming problems. The results showed that using reformulation from quadratic
with Karush-Kuhn-Tucker (KKT) condition for bounding the problems helps to get
optimal solution in branch and bound method.

Keywords


convex optimization, quadratic programming, branch and bound, Karush- Kuhn-Tucker (KKT) condition.

Refbacks

  • There are currently no refbacks.