MEFISTO: A pragmatic metaheuristic framework for adaptive search with a special application to pickup and delivery transports. We present MEFISTO, a pragmatic framework for transport optimization algorithms that has been jointly developed by the authors and is integral part of logistics planning solutions provided by PTV AG. We present design aspects that have led to the architecture of the framework using a variant of Granular Tabu Search. For the case of vehicle routing problems with pickup and delivery transports, we present a specialized local search procedure. Summarized results on benchmark and real world problems are given.