Recent research studies which purport to integrate distributed information have been concentrated on developing efficient mediators that provide a high degree of autonomy for a local user. Since a view in a mediator is defined by views in other mediators or base tables in several information resources, semantically related mediators form a hierarchical structure based on the definition of views in each mediator. It has been shown that the query evaluation cost of a mediator can be greatly reduced if the results of frequently requested queries are materialized in the mediator. Unlike a database system, the schema of a mediator can be evolved. In addition, the evaluation frequency of each query as well as the update frequency of each related information resource can be easily varied as time goes by. Consequently, they should be adaptively monitored. The effect of such materialization is greatly increased in a set of mediators in a hierarchical mediator system. This paper, proposes an algorithm for selecting the optimized set of materialized queries in each mediator such that the overall query evaluation cost of a hierarchical mediator system can be minimized at any time. In addition, the performance of the proposed algorithm is analyzed by a series of experiments to identify their various characteristics.