Stigmergic Linking for Optimising and Reasoning Over Information Networks
Kieran Greer, Matthias Baumgarten, Maurice D. Mulvenna, Kevin J Curran, Chris Nugent · 2008
Abstract- This main contribution of this report will be to describe and evaluate a set of tests, used to measure the effectiveness of stigmergic linking with regard to query optimisation. It will be shown that the results of queries can be autonomously fed back through a network to allow it to self-optimise. The type of network that is being considered is a network of information sources, such as a ‘knowledge network’. This is a construct that will self-organise distributed knowledge in a way that allows it to be efficiently retrieved and used. Key elements thereof are its lightweight, reference-based structure and its autonomous nature. The network may have permanent organisational structures based on semantics. However, the stigmergic links will add an extra dimension to the self-organisation and allow it to optimise the query process effectively. These links will in effect create temporary views that reflect the use of the system and complement any permanent organisation created through semantics. Some tests on reasoning over the stored information will also be provided and shows that higher levels of reasoning can also be obtained.