Parametric Solution of Bicriterion Linear Fractional Programs

We show how certain bicriterion fractional programs can be reduced to a one parameter linear program and a series of one-dimensional maximizations. The resulting algorithm is easily implemented using the PARAROW option of MPSX, and has readily solved problems having up to 300 variables and 150 constraints. As an illustration, we present a simple finite procedure for maximizing the sum of two linear fractional functions subject to linear constraints.