Banca de DEFESA: THIAGO DE JESUS INOCÊNCIO

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
DISCENTE : THIAGO DE JESUS INOCÊNCIO
DATA : 16/12/2020
HORA: 14:00
LOCAL: Google Meet (meet.google.com/nqc-hdwe-wqg)
TÍTULO:

A DATA MINING-BASED METHODOLOGY FOR DESIGNING SYSTEMS-OF-SYSTEMS USING EMERGENT BEHAVIORS DERIVED FROM BUSINESS PROCESSES

PÁGINAS: 80
GRANDE ÁREA: Ciências Exatas e da Terra
ÁREA: Ciência da Computação
SUBÁREA: Metodologia e Técnicas da Computação
ESPECIALIDADE: Engenharia de Software
RESUMO:

Context: A system-of-systems (SoS) is a class of systems that is characterized by the union of independently operated and managed systems that do new and valuable things when together. Each constituent of an SoS contributes with their specific functionalities to build new behaviors that could not be performed by any one of these constituents in isolation, these behaviors are known as emergent behavior. Because of the complex nature of this class of systems, as new constituents join or leave the SoS, unexpected emergent behaviors can appear, making modeling and simulation a significant role in the development of these systems. Business processes are widely adopted in organizations to improve agility and decision making. In many cases, the activities present in a business process are carried out through systems. Problem: If a business process activity cannot be performed efficiently by a single system, there may be sets of systems that together present emergent behaviors capable of meeting this activity. The search for constituents can be carried out in systems repositories, however, depending on the requirements defined for the constituents, a manual search is an expensive and possibly fault-oriented task. In this sense, data mining techniques are characterized as an important tool in the knowledge discovery process in large systems databases. Objective: Thus, the main objective of this master's dissertation is to present a methodology capable of creating models of SoS based on emerging behaviors derived from business processes using data mining techniques. Methods: A systematic mapping study was carried out in order to analyze the existing literature dealing with emerging behaviors in SoS in order to characterize the current state of the art in relation to existing works in the SoS field. In order to find the SoS constituent systems by their attributes, data mining algorithms were used to support this task. For this, the proposed methodology was applied in a practical case study with a view to validating it. Results: The application’s results of the methodology show that although all the data mining algorithms analyzed in this research present satisfactory results, the dbscan algorithm showed more satisfactory results in the search for systems constituents that together show emerging behaviors that meet business requirements. However, it can be seen that the dbscan algorithm showed the longest execution times among all techniques, making it impractical if the execution time is an important variable during the methodology application. In contrast, the ranking algorithm presented itself as the
best option when execution time is more relevant. Conclusions: This master's thesis presented important results for the state of the art of SoS engineering that meets the research directions identified in the systematic mapping. The main contribution concerns the finding that emergent behaviors from Sos derived from business processes are capable to carry out the activities present in these processes.

MEMBROS DA BANCA:
Presidente - Interno ao Programa - 3007914 - FLAVIO EDUARDO AOKI HORITA
Membro Titular - Examinador(a) Interno ao Programa - 2376122 - THIAGO FERREIRA COVOES
Membro Titular - Examinador(a) Externo à Instituição - EVERTON RANIELLY DE SOUSA CAVALCANTE - UFRN
Membro Suplente - Examinador(a) Interno ao Programa - 3009301 - VLADIMIR EMILIANO MOREIRA ROCHA
Membro Suplente - Examinador(a) Interno ao Programa - 3008017 - DENIS GUSTAVO FANTINATO
Membro Suplente - Examinador(a) Externo à Instituição - VALDEMAR VICENTE GRACIANO NETO - UFG