The purpose of this study is to provide a query optimization tool for multi-version XML documents.

In the paper, multi-version XML documents mean user-generated documents which are written by XML and save all past versions created by updating such as office document and/or wiki contents. Each version is preserved not at the document level but at the element node level, and we can perform fine-grained version retrieval such as retrieving a subdocument of any versions from very large XML documents.

We propose a number of query rewriting rules which can exchange between document axes and version axes, where document axes represent document structures and version axes represent history information. The rewriting rules enable us to swap path subexpressions between document axes and version axes to optimize queries.