Designing turbine wheels for automotive turbochargers one is faced with a multidisciplinary design problem with many input and output parameters. Especially in the automotive industry short development cycles for high quality products in a competitive environment are daily routine. For meeting these requirements optimization algorithms can be a powerful tool in the design process. This paper presents the multidisciplinary optimization of an automotive mixed flow turbine wheel used in a 4 cylinder 1.6 l spark ignition engine. Before describing the optimization workflow in detail, the requirements for turbines operating in an automotive environment under pulsating flow conditions and during an engine load step are discussed. From there objectives for a multidisciplinary optimization are derived. The turbine wheel is optimized with respect to maximizing efficiency in two design points and minimizing its moment of inertia. For the optimization process, an algorithm based on evolution theory is used. As constraints, the operating points are fixed and the natural frequencies are limited to ensure the mechanical strength of the turbine. To speed up the optimization process meta models based on neural networks are applied. Three designs of the Pareto frontier are chosen and their characteristics are discussed. Using statistical methods, the interaction of the input variables and their impact on turbine performance are presented.