An efficient genetic-algorithm-based (GA-based) technique for modeling and analysis of RF heterojunction bipolar transistors (HBTs) has been successfully established. This mixed-mode optimization method focuses on complementing the GA with a set of generalized analytic-modeling equations (GAMEs), requiring no special assumptions, to effectively extract small-signal parameters. Over a wide range of biasing conditions, consistent results between the GA-derived and measured S-parameters on the pup InGaAs collector-up HBT as well as npn SiGe HBTs, which can be used in small high-power amplifiers for mobile communication systems, clearly demonstrate the superiority of the proposed approach. (c) 2006 Elsevier Ltd. All rights reserved.