Case-Based Reasoning Model in Process of Emergency Management

Autor: Jiri Krupka, Miloslava Kašparová, Pavel Jirava
Předmět:
Zdroj: ResearcherID
Man-Machine Interactions ISBN: 9783642005626
ICMMI
Popis: In this paper a modelling of a system, in the conventional structure, of the emergency service is presented. The model is based on the case-based reasoning (CBR) algorithm, and it focuses on the fire rescue service. The algorithm has been realized in MATLAB, and their outputs represent effective recommendations for a commander of a fire protection unit. In the first part of this paper, information about CBR algorithm, and a decision process in the tactical level of fire rescue service, are briefly introduced. The second part is aimed at the system approach definition of the formation of ‘cases’ of decision making, by the fire protection unit commander in charge during the handling of emergency situations, on the basis of guidelines in the fire protection unit field manual. The manual contains a set of methodical sheets; every sheet is described by a set of attributes. It exploits characterization of a set of cases. Generally this real decision is executed under pressured time, contains a lot of information, and it is unique for each situation; it is very important to compare the decision of new cases with the base decisions of previous cases. In the third part, the possibility of applying CBR algorithm to new real cases is analysed. We modified the guidelines of cases on the basis of commander’s know how, and on expert recommendations. We analyzed some metrices into a comparison of cases, and we realized CBR algorithm as a graphic user interface. Finally we discussed the possibility of using Soft CBR, it means using theories of fuzzy logic and rough sets for a description of vagueness and uncertainty during the description of knowledge in the fire rescue service.We will consider an implementation of CBR algorithm on digital signal processors.
Databáze: OpenAIRE
Externí odkaz: