Cap model is a well-known model for defining the yield surface of porous materials. In this work, the constants of the model are determined by a combined experimental/numerical/optimization technique for aluminum 7075, iron and copper powder. The compaction of powder is conducted using Instron testing machine. The numerical simulation of powder compaction is performed using the Ls-Dyna hydrocode. Optimization is carried out using genetic algorithm. The objective function is defined as the difference between the experimental and numerical load-displacement curves of the compaction. The constants of the model correspond to the case when this difference is optimized. A second order Maclaurin polynomial is assumed for the objective function. The results indicate that the cap model can reasonably predict the powder compaction of iron and aluminum. The model however, proves not to be accurate for copper powder. The results also show dependency of the model's constants on strain rate. (C) 2015 The Society of Powder Technology Japan. Published by Elsevier B.V. and The Society of Powder Technology Japan. All rights reserved.