A gradient dynamical system for solving mixed equality-inequality constrained optimization problems

摘要

In this paper, a gradient dynamical system is derived for solving the optimization problems with mixed equality-inequality constraints, and is applied to the 0-1 optimization. This gradient dynamical system is the dynamical system an equilibrium point of which is equal to a local optimal solution, and is constructed from the Karush-Kuhn- Tucker optimality condition. Algorithms are also constructed for the direction vector and the variable step-size of an iteration method for solving the constrained optimization problems.