Research For Emergency Decision Making Methods Based On Case Reasoning

Emergency decision-making is a process where decision maker formulates adisposal solution for a current event from the situation and from some disposalexperience related to the current event. It needs to derive a solution to the currentevent from a group of some related cases in emergency decision-making based onCBR. It is essentially a process where the disposal solution will be derived from agroup of uncertain emergency cases using logic reasoning, and there two problemsinvolving in the process: knowledge representation of emergency cases and reasoningbased on strong relevant logic with uncertainty information. This paper focuses on theknowledge representation and logic reasoning with uncertainty information, and thefollowing major works were completed in this paper:（1）Rough strong relevant logic based on strong relevant logic was proposed tobe the basis of knowledge representation and reasoning of uncertain emergencycases.Rough set theory was introduced into rough strong relevant logic to defineupper approximation operator, lower approximation operator, and rough similarityrelation, and to extend strong relevant logic’s ability in representing uncertainconcepts.According to the need of knowledge representation of emergency cases, theconcrete domains was extended in the rough strong relevant logic, and somepredicates were defined in rough strong relevant logic, such as spatial topologicalrelation, spatial direction relations, temporal relation, and some spatial logicaltheorems and temporal logical theorems based on rough similarity relation in strongrelevant logic were proposed.（2）The ontology description language SWRL, extending from SWRL language,was designed to describe rough ontology of emergency cases. CBR-based emergencycase ontology model (CECOM) was built based on the extended SUMO ontology, andfour key ontology elements of the model were described, such as concepts, relations,axioms and instances. RSWRL and CECOM were validated by an instance.（3）On the basis of RSWRL and rough strong relevant logic, an emergencydecision making model based on CBR was proposed. The rough concept similarityalgorithm was proposed by improving the semantic concept similarity algothim, and then on the basis of it, the emergency cases similarity algorithm and the emergencycases retrieval algorithm based on rough similarity relation were proposed. Thesealgorithms were validated by the experiment.（4）A prototype system was designed and developed. A framework ofemergency decision based on case-based reasoning was established and an emergencydecision making prototype system based on case-based reasoning was developedbased on the framework, and it was verified using some instances.