Saddle-point type optimality criteria for generalized fractional programming (Q1090248)

Generally speaking, one cannot expect that there exists a Lagrangian saddle point even for a linear fractional program. Using the functions \(GF(x,r_ 0,r)\) and GK(x,u) somewhat like the Lagrangian functions, we present saddle-point type optimality criteria for generalized fractional programming under carefully selected assumptions. In addition, we point out conditions which ensure that a local optimal solution of the program mentioned is global.