Português English
Contato

Proposta de Tese de João Guilherme Faccin


Detalhes do Evento


Aluno: João Guilherme Faccin
Orientadora: Profª. Drª. Ingrid Oliveira de Nunes

Título: Promoting Resilience with Automated Management of Remediation Actions

Linha de Pesquisa: Inteligência Artificial

Data: 22/01/2019
Horário: 10h
Local: Prédio 43412 – Sala 215 (sala de videoconferência) do Instituto de Informática

Banca Examinadora:
– Prof. Dr. Edison Pignaton De Freitas (UFRGS)
– Profª. Drª. Genaína Nunes Rodrigues (UnB – por vídeoconferência)
– Prof. Dr. Valter Vieira de Camargo (UFSCar – por vídeoconferência)

Presidente da Banca: Profª. Drª. Ingrid Oliveira de Nunes

Abstract: Services provided by software systems have an increasing impact on our lives, ranging from information management to the automation of complex tasks such as driving. It is thus necessary that such systems remain operational for as long as possible, being able to resist and recover from unpredicted abnormal situations. This ability is called resilience. One way to achieve resilience is through a remedial problem-solving strategy in which the effects of a problem are minimised while its causes are identified and, when possible, resolved. Currently, solutions capable of providing this remedial behaviour to software systems are designed and implemented focusing on specific domains and problems. Therefore, the recurrent solutions developed to control the rmedial behaviour are not exploited in a reusable way across different systems. In addition, existing solutions require remediation actions adopted by the system to be previously coordinated, which results not only in an increased need for domain knowledge by developers, but also in limited system adaptability to situations unpredicted at design time. In this thesis, our goal is to improve the way resilient systems with remedial behaviour are developed. Therefore, we propose a domain-independent approach that abstracts the reasoning related to this behaviour from the particularities of each problem addressed. This proposal explores the concept of autonomous software components and comprises the development of techniques to automate not only the action management process arising from the remedial problem-solving strategy, but also the extraction of causal knowledge required by it and the identification of problematic events to be handled. As a consequence, the resulting solution will be able to be employed in different application domains, minimizing the need for human intervention to adapt the behaviour of resilient software systems.

Keywords: Resilience. remediation actions. software agents. BDI architecture. causality.