In alloy design, one seeks to identify compositions where certain functions have optimal values under various constraints. Such functions include liquidus temperature, amounts and compositions of phases (either as-cast or after annealing), densities, vapour pressures, and viscosities. We report here on the development of new tool to perform such calculations. The tool incorporates the MADS (Mesh Adaptive Direct Search) algorithm which is designed to solve nonsmooth optimization problems where the objectives and constraints are typically outputs of computer simulations. Numerical results for several examples will be presented. (C) 2013 Elsevier B.V. All rights reserved.