Uma abordagem metodológica de apoio à decisão para resoluções de incidentes em TI centrada na combinação de raciocínio baseado em casos e baseado em regras

• Autor(a) principal: MELLO, Thiago Dias Freire de
• Data de Publicação: 2015
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UNIFACS
• Texto Completo: http://teste.tede.unifacs.br:8080/tede/handle/tede/483
• topic: Raciocínio Baseado em Casos; Raciocínio Baseado em Regras; Tomada de decisão; Abordagem Metodológica de apoio à decisão; ITIL; Ciências Exatas e da Terra

In present day, the services offered by Information Technology (IT) has become a key factor for organizations to achieve success. The unexpected interruption of these services, known as incidents, causing damage to organizations that can reach millions of dollars. Decisions and actions are constantly taken for support teams to repair and to normalize the services caused by these incidents. The process of efficient decision-making requires timing and standardization among all members of a team. Many organizations adopt Information Technology Infrastructure Library (ITIL) to structure and facilitate their decision-making. On its turn, two important models of artificial intelligence, Case-Based Reasoning (CBR) and Rule Based Reasoning (RBR), has aided the decision-making process using reasoning by analogy and rule, respectively. On the other hand we must structure at what time and when to use each of these arguments in decision support, defining and ordering a set of steps. This paper presents a methodological approach to support decision making in resolving IT incidents, described through a general decision process respecting a consolidated decision model, the model Simon decision, aided by the combination of techniques Case Based Reasoning (CBR) and Rules Based Reasoning (RBR), since ITIL does not propose the best practices for incident resolution analogy nor by rule. Therefore, the aim is to support decision-making of the support team through a methodological approach, in order to solve the incidents, considering the reasoning by analogy and reasoning by rule. A prototype was built to support decision-making with these two cited arguments relying on frameworks JColibri and Expert Sinta. A case study was carried out within an organization in order to evaluate the proposal of this work, being resolved four cases of incidents using the prototype with the built methodological approach.