Rule-based system for media setup optimization

The use of a rule-based system (RBS) to offload media setup for complex media and network conditions is described. An example application uses a neural network acting as an associative search network (ASN) to implement a fuzz-pattern-matching front end to select the best media. The ASN is sensitive to the current context of the network, the load being demanded of the media, and the status of the media. Using this information, the ASN makes a choice of the best media for the current situation. The media selection is forwarded to the RBS, where the same information, current network context, media load, and media status are used to determine the optimum setup for the chosen media. The RBS has been coded using the Ada programming language. The knowledge-acquisition process to produce the pattern-directed modules for particular media setup advice is formulated. The integration of the neural-network simulation and the RBS as well as the recoding of the simulation in the Ada programming language is being implemented