In this work we introduce an adaptive architecture for supporting news information services in a dynamical web environment. In the news provider problem, a system continuously generates service instances (e.g.the news of a online newspaper) which should meet the requirements of a market of readers/agents which are supposed unknown. The only way for the system to know the evaluation of a service instance, i.e. a newspaper instance, is the feedback obtained after delivering it to the reader. In a dynamical environment the domain of the service offered is changing and the readers/agents are changing their preferences over the time. The proposed algorithm for newpapers production employs typical genetic operators in order to optimize the product delivered and to adapt it to the environment feedback and domain evolution. Experiments shows the capability of dynamical adaptation of the algorithm..