A Genetic Algorithm-Based Approach for Fluctuating QoS Aware Selection of IoT Services

Abstract

In recent years, the Internet of Things (IoT) has evolved at an exceptional speed, which enables to interconnect a very large number of heterogeneous, distributed, and mobile devices. This number will exceed 70 billion by 2025 according to Statista a . Therefore, with this huge amount of connected objects, the fulfillment of complex IoT applications, which usually requires a combination of several IoT objects, remains a real challenge. Besides, several requirements of Quality of Service (QoS) must be fulfilled, which makes the problem of selecting the appropriate IoT services NP-hard. In the literature, two main techniques for QoS-driven service selection are proposed: global selection characterized by a poor performance in dynamic and distributed huge environments and local selection which considers pre-defined local QoS constraints. Mainly, the existing works consider static QoS. However, in real life scenarios, QoS of IoT services can be fluctuating. To enhance the reliability of IoT applications, it is of paramount importance to consider the fluctuation dimension. In this context, we propose a QoS fluctuation-aware selection approach of IoT services. To do so, we propose a near-to optimal distributed approach that relies on decomposing the global QoS into distinct local constraints that serve as upper/lower bounds for selection while enhancing the reliability of the resulting composition by considering the QoS fluctuation of the candidate IoT services. The approach, we propose is based on a multi-objective evolutionary algorithm (MOEA) to solve the global QoS decomposition problem. Then a local selection using the obtained local constraints is performed in a parallel and distributed way. The performance of the proposed approach is evaluated and validated via experiment series.

Optimal Control Strategy of Construction Machinery Motor Considering Genetic Algorithm

“…Intelligent optimization algorithm refers to determining the optimal scheme according to the change rule of the relationship between the studied objects (such as genetic factors) through the analysis of input and output information in the system [11][12]. The following issues shall be considered in controller design.…”

Intelligent Drone Positioning via BIC Optimization for Maximizing LPWAN Coverage and Capacity in Suburban Amazon Environments

“…GAs employ a population-based approach to problem-solving, where a set of possible solutions, called individuals, are evolved over several generations using selection, crossover, and mutation operators. It is known they are employed in many IoT applications in the literature [37][38][39].…”

Web Information Systems Engineering – WISE 2023

“…With IoT, real-time data can be collected, processed, and analyzed to enable smart decision-making. The Web of Things (WoT) has further propelled theses advances by enabling physical objects to be connected to the Web through their virtual representations [7], enabling them to be discovered and used as IoT services. Although WoT has brought great opportunities to provide new value added services that can leverage a wide range of potentially heterogeneous IoT devices to fulfill complex IoT needs, it also brings several challenges that remain open, such as IoT service discovery, selection and composition.…”