A note on primal-dual stability in infinite linear programming (Q2228398)

The authors examine consistency as well as inconsistency of linear programming problems and their corresponding duals defined in infinite-dimensional Banach spaces. They scrutinize these problems under sufficiently small perturbations of the data. The authors analyse different possibilities of data under distinct conditions and obtain the results for the same. The consistency of fixed data for primal-dual stability is presented in case of linear infinite programming. This result is significantly being applied to the study of the stability of duality gap functions. Theorems and results are stated and well proved to obtain necessary and sufficient conditions for the situation where optimal values of primal and dual problems are equal under small perturbations. This condition is stated as zero stability. It is also shown that the property dealing with dual stability with respect to consistency and primal stability with respect to inconsistency in linear programming and linear semi-infinite programming problems holds true in finite-dimensional spaces but does not hold in infinite-dimensional spaces. This is because there is variation in the topological properties of cones in finite- and infinite-dimensional Banach spaces, therefore, the results differ considerably for finite-dimensional problems.