A low hit rate on dc_objects is a perfectly normal symptom of the use of public synonyms. If SCOTT parses the statement "SELECT * FROM DUAL;" then the session will first attempt a get against the dc_objects cache for SCOTT.DUAL. This will of course result in a miss. Thus the hit rate on dc_objects is more a reflection on the extent to which objects are referenced via public synonyms, rather than the effectiveness of the dictionary cache.

Yes, it could be dc_objects but the question still remains the same.

Are you sure that you don't mean dc_objects rather than dc_object_ids?

It is common to have an acceptable overall dictionary cache ratio while the individual dictionary cache component (i.e. dc_object_ids) can have a poor hit ratio. I can increase the shared_pool_size, but sometimes this will not significantly improve the individual component in question. Is there any other way to tune this?