A new algorithm for generalized fractional programs

摘要: A new dual problem for convex generalized fractional programs with no duality gap is presented and it shown how this can be efficiently solved using a parametric approach. The resulting algorithm seen as “dual” to the Dinkelbach-type since approximates optimal objective value of (primal) from below. Convergence results are derived an easy condition achieve superlinear convergence also established. Moreover, under some additional assumptions recovers at same time solution primal problem. We consider variant algorithm, based on scaling function. numerical results, in case quadratic-linear ratios linear constraints, show that performance its scaled version superior algorithms. From computational appears contrary approach, approach less influenced by scaling.