Many large virtual 3D scenes are not structured evenly, for example because they exhibit a highly varying density distribution of their polygons. For such heterogeneous data, there is no single algorithm that constantly performs well with any type of scene and that is able to render the scene at each position fast with the same high image quality. For a small set of scenes, this situation can be improved, if an experienced user is able to manually assign different rendering algorithms to particular parts of the scene. We introduce the Multi- Algorithm-Rendering method which automatically deploys different rendering algorithms simultaneously for a broad range of scene types. The method divides the scene into suitable subregions and measures the behavior of the different algorithms for each region in a preprocessing step. During runtime, this data is utilized to compute an estimate for the quality and running time of the available rendering algorithms from the observers point of view. By solving an optimizing problem embedded in a control cycle, the image quality can be optimized by the appropriate assignment of algorithms to regions, while keeping the achieved frame rate almost constant. In an experimental evaluation, we compare our methods running time and image quality with that of standard rendering algorithms.