An artificial neural network for oil spill cleanup: the MUSE system

An extensive and comprehensive shoreline cleanup knowledge-base was developed as part of the implementation of a decision support system for teams involved in oil spill cleanup operations. The system, ShoreClean, was designed to provide environmental emergency response teams with a decision support system that would allow them to improve their efficiency. It was implemented with a backward chaining rule-based system. An expert system was a judicious choice for the development of the system's knowledge base, but the constant updating of the knowledge-base and the incapacity of the system to respond to fuzzy questioning caused the author to migrate toward a distributed artificial neural network (DANN). This article describes the grafting of the rule-based system and DANN.