A computational comparison between algorithms for linear fractional programming

A computational experience on random generated instances of the linear fractional problem is reported and discussed. The algorithm by Charnes and Cooper and the one by Cambini and Martein are compared, with particular attention for the case of unbounded feasible region. Indeed theoretical considerations allow the comparison of the two algorithms also in this situation. From the results obtained it can be observed that the algorithm by Cambini and Martein gets better results, on average, than the other one.