Gadolinia doped ceria in its doped or strained form is considered to be an electrolyte for solid oxide fuel cell applications. The simulation of the defect processes in these materials is complicated by the random distribution of the constituent atoms. We propose the use of the special quasirandom structure (SQS) approach as a computationally efficient way to describe the random nature of the local cation environment and the distribution of the oxygen vacancies. We have generated two 96-atom SQS cells describing 9% and 12% gadolinia doped ceria. These SQS cells are transferable and can be used to model related materials such as yttria stabilized zirconia. To demonstrate the applicability of the method we use density functional theory to investigate the influence of the local environment around a Y dopant in Y-codoped gadolinia doped ceria. It is energetically favourable if Y is not close to Gd or an oxygen vacancy. Moreover, Y-O bonds are found to be weaker than Gd-O bonds so that the conductivity of O ions is improved.