A recommender system based on semantic web technologies and on an adaptive hypermedia architecture is shown in this paper. The system uses a stochastic algorithm to provide recommendations to users. The paper presents the system architecture based on the semantic Web technologies and explains a simulated annealing algorithm performing the recommendations. A mobile application for the tourism domain proving the feasibility of this system is described at the end of the paper, some benchmarks are presented. In this application, the recommendations are defined as combinations of tourism products, which are linked to each other. The paper is mainly focused on the architecture and the recommendation process of the system.