HARL: A knowledge Representation Language

HARL is a general-purpose knowledge representation language for building expert systems. It represents reasoning knowledge with the clustered structures whose centers are the hypotheses, organizes reasoning actions in multiple ways according to the control meanings of associative links of the hypotheses; therefore it provides the strong power of representation: focusing the attention of reasoning control highly and making the strategies for problem solving flexibly. The inference engine of the HARL system adopts a general control structure based on the control meanings of the associative links, the marks indicating the modes for selecting rules from rule groups, and the schedule information in the blackboard. Because of allowing the combination of multiple control meanings, the use of multiple kinds of knowledge units and the acquisition of data from multiple sources, HARL can adapt itself to extensive domains.