Ranqueamento de recursos em IoT: uma abordagem explorando análise de consenso Fuzzy

Abstract

The Internet of Things (IoT) is a distributed computing paradigm whose growth has attracted attention both for the opportunity for a new impact on the digital economy and for the potential research fronts to be explored. This phenomenon is driven by the growing supply of resources connected to the Internet and brings with it a greater complexity in its operations, especially in the procedures for discovering, classifying and selecting suitable resources to meet customer demands. The proper selection of services, considering the availability of a large number of services able to serve you and whose availability may change over time, comprises the scope of this work. Specifying customer preferences is huge task involving inaccuracies and uncertainties, depending on previous knowledge and experience for the correct parameter definitions as measurement scales. This article aims to address the uncertainties in specifying and processing customer preferences when classifying a set of resources on the IoT. For this purpose, both definitions of fuzzy consensus measures and consensus measures on fuzzy sets are presented basing on strict equivalence relations and on aggregations as arithmetic and exponential means. The EXEHDA-RR model collaborates with the discovery and classification of resources in the IoT, considering not only non-functional attributes of QoS but also high scalability of resources. The proposal stands out in the classification of IoT resources, exploring the treatment of uncertainty in processing preferences using both fuzzy logic approaches: T1FL and T2FL. In addition, scenarios containing resource request simulations and applying different customer preferences can be demonstrated in the EXEHDA-RR resources. The main results obtained are presented, adjusting the selection of the most relevant resources according to the preferences specified in linguistic terms for each QoS attribute.