Extensions of Lagrange Multipliers in Nonlinear Programming

摘要: defined for x > 0, u 0. A saddle point of (1.2) is to be a (x*, u*), x* Q, u* 0 such that +(x, u*) < +(x*, _ u) all ? One the Kuhn-Tucker results of(1.2) provides solution (1.1). Under cer-eain additional regularity assumptions equivalence problem and program (1.1) has been shown [1], [10]. In particular, if feasible set in an interior (a nonnegative g(x) b),3 objective function concave constraint functions are convex, then only there (1.2). This so-called theorem nonlinear programming. this paper Lagrange multipliers generalized from usual constants (possibly nonlinear) multiplier functions. leads statement several general under quite weak assumptions. From somewhat different view, Everett [4] drops differentiability discusses procedure solving programs terms