In this paper, we will use the current reasoning capabilities of OWL. This reasoning process could be simplified as the following formula, CQ=R(O) where CQ stands for the results of classification and query used during this process, R stands for the actual process of  Tbox and Abox reasoning, O stands for the OWL ontology sets used for the current reasoning. After OWL reasoning, there may involves the changes of agent state σ, e.g. new beliefs obtained. Therefore, we could define a new belief update action αub：αub =(UpdateBelief, B’,B,β,△t), the action type is UpdateBelief，β is parameter for belief, △t is the interval for updating, B is the original belief，B’＝F(CQ) is new beliefs after formatting the reasoning results CQ.

Then the results will be parsed, and these new facts will be propagated and added to the belief base in related ADFs (all persons sharing the same hobby sport within reasonable distance) . This process is based on the assumption that the ADFs are equipped with customization slots (using XVCL <break> commands with ‘xvcl’ as namespace). Take the Person agent as an example, we add break points to allow customization of hobby sharing by enclosing original hobby sharing information

Other evolution scenarios could be handled in principle in the same way and will be shown in detail in an upcoming paper.