Generic Algorithm for Generalized Fractional Programming

We propose a unified framework to study various versions of Dinkelbach-type algorithms for solving the generalized fractional programming problem. Classical algorithms used carefully selected iterate points and incorporated subtle convergence proofs. Our generic algorithm, however, is natural and simple. Moreover, the convergence analysis can be carried out through geometric observations and fundamental properties of convex functions. Consequently, the classical results are either refined or strengthened with new insights.