Applying relevance feedback with a linear programming classifier;

An interactive linear fractional programming algorithm is presented to solve multiple attribute decision problems based on the assumption that the decision-maker has a linear utility function.

In this paper, the authors present an algorithm to handle linear fractional programming in the general form directly, which need not transform the constraints of the problem into the stan dard form.

In this paper, a numerical example of linear fractional programming (LFP) has beenconstructed in which a finite sequence of degenerate bases obtained by the LFP’s primalsimplex algorithm in referenes [1] and [2] may yield basis cycling and hence no opti-mum solution could be got.

In this paper,a class of new generalized invexity concept is defined on basis of[1],and then Mond-Weir duality theorems with the weakduality theorems,strong duality theorem and converse duality theorem are proved under this new generalized invexity condition for a classof nonconvex nonlinear fractional programming.

And then some optimal sufficient conditions are proved under these new generalized invexity sufficient conditions for a class of nonconvex nonlinear fractional programming.