Removing inessential points in c-and A-optimal design

A design point is inessential when it does not contribute to an optimal design, and can therefore be safely discarded from the space. We derive three inequalities for detection of such points in c-optimal design: first two are direct consequences equivalence theorem c-optimality; third one derived a second-order cone programming formulation design. Elimination rules A-optimal obtained as byproduct. When implemented within optimization algorithm, each inequality gives screening test that may provide substantial acceleration by reducing size problem online. Several examples presented with multiplicative algorithm illustrate effectiveness approach.