The minimum risk approach to max-min bilinear programming. (English) Zbl 0628.90074

The max-min bilinear programming problem is investigated and the minimum risk approach for it is formulated. The authors reduce this problem under certain hypotheses to a max-min bilinear fractional problem which is deterministic and does not depend on the distribution function of the random variables. A parametrical algorithm for solving the max-min bilinear fractional problem is proposed. The convergence of the proposed algorithm as well as the finiteness of an approximate variant is based on the theorem stated in the paper.
Reviewer: R.G.Vachnadze


90C31 Sensitivity, stability, parametric optimization
65K05 Numerical mathematical programming methods