KnowSoS: A Software Architecture for Knowledge Discovery in Systems-of-Systems

• Autor(a) principal: Silva, Bruno Sena da
• Data de Publicação: 2020
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Biblioteca Digital de Teses e Dissertações da USP
• Texto Completo: https://www.teses.usp.br/teses/disponiveis/55/55134/tde-22072020-153547/
• topic: Arquitetura de software; Descoberta de conhecimento; Knoweldge discovery; Sistemas-de-Sistemas; Software architecture; Systems-of-Systems

Systems-of-Systems (SoS) have been playing an important role in the industry and academia as an answer to the ever growing complexity of software systems. This class of systems emerges from the interoperability of operational and managerial independent systems (called constituents) that work together to achieve more complex missions, not achievable by any of them individually. Constituents of an SoS generate at runtime a huge amount of data with high velocity and that is sometimes highly variable, characterizing a big data scenario. However, SoS have not taken advantages from the knowledge that could be discovered from such data, which could provide an overall understanding of the SoS behavior, support missions fulfillment, and enable the system to react to it. In this scenario, the main contribution of this Masters project is to provide means to support knowledge discovery in SoS, that is, to support the entire data cycle in SoS, from the collection in the constituents to the extraction of knowledge that will be useful for the system operation. For this, we propose KnowSoS, a Software Architecture for Knowledge Discovery in Systems-of-Systems, which assists in the design of a system that can be integrated into the SoS to control data flow and enable knowledge discovery. Besides that, we propose a set of guidelines that support SoS understanding and the development of the system designed by using KnowSoS. To evaluate this proposal, we present a proof of concept using the architecture and the guidelines to design and develop a knowledge discovery system in two different SoS domains, i.e., human resources and emergency management domain. As main results, the system improved the recommendation of job openings to candidates who are looking for a job in the human resources domain, and the precision of alerts for natural disasters to emergency stakeholders (i.e., police, hospitals, and firefighters) in the emergency management context. Such results indicate that, even as an initial approach, KnowSoS can be used to bring benefits in different domains where knowledge discovery is necessary to fulfill SoS missions. As future works, we intend to apply this approach in new domains and conduct more robust experimental studies.