Pitfalls of JESS for Dynamic Systems

We considered different varieties of inference engines for a sub-system of Mission Control Technologies (MCT) being developed at NASA Ames Research center. One inference engine, Jess, is attractive due to its benchmark results, Java API, and being a stable software product. The outstanding issues of Jess with respect to MCT are the way its Java Bean defines ’defclass’ and slots, and converting between or mapping to MCT interfaces. The slot names and ’defclass’ are static in Jess rules, and at run time in a given knowledgebase it cannot be modified. Distributing Jess to outside federal agencies and licensing is another issue.

• Type: Conference
• Year: 2007
• Where: AIPRF
• Authors: Rajkumar Thirumalainambi
• Keywords: AIPRF 2007, Artificial Intelligence, Inference Engine, Jess Rules, NASA Ames Research Center