The Web has become an important source for information, which are created by independent providers. Web portals provide an unified point of access to content, data, services and web applications located throughout the enterprise. However, Web users have often only an insufficient available amount of time, to effectively use the available information resources. This thesis proposes a mashup framework that automatically mashes-up web portal content with related background information. The background information are derived from information web services that are composed by an evolutionary algorithm.