ABSTRACT
This chapter explains a basic Quadratic Programming (QP) algorithm. It addresses the problem of finding an initial feasible point. We formulate this as a linear programming problem for which an initial feasible point is available by inspection. Since a linear programming (LP) is a special case of a QP, the LP can be solved using the QP algorithm. The complete solution procedure then consists of two passes of the QP algorithm. The first finds an initial feasible point and the second, beginning with the initial feasible point obtained by the first, solves the QP. The chapter illustrates a variation of the basic QP algorithm. The modification substantially reduces the total number of iterations required to solve a QP. Each of the two QP algorithms developed in this chapter require that all quasistationary determined by the algorithms be nondegenerate in order to guarantee finite termination. The chapter formulates a method to resolve degeneracy.
