Some special situations in linear fractional programming problem: An algorithmic comparison (Q1825774)

A comparative study of algorithms due to \textit{A. Charnes} and \textit{W. W. Cooper} [Nav. Res. Logist. Q. 9, 181-186 (1962; Zbl 0127.369)], \textit{B. Martos} [ibid. 11, 135-155 (1964; Zbl 0131.185)] and \textit{H. Wolf} [Oper. Res. 34, 930-937 (1986; Zbl 0627.90085)] for solving linear fractional programming problems, when the feasible set F is unbounded, is presented. Furthermore, the three algorithms are compared with respect to identification of so-called bad points (\^x\(\in F\) is called a bad point if \[ \lim_{x\to \hat x}\frac{n(x)}{d(x)}=\infty \] where \(\frac{n(x)}{d(x)}\) is the objective function) and singular points (a singular point is defined as a point \(\hat x\in F\) at which n(\^x)\(=d(\hat x)=0\) provided \(\hat x\) is no bad point). It is observed that Wolf's algorithm is as good as Charnes-Cooper's algorithm for the case when F is unbounded and is comparatively better than the other two if identification of bad points and singular points is concerned.