Designing a Modelfor the Service Selection Aware-Quality of Service in the Internet of Things Environment

Abstract

The Internet of Things (IoT) has been evolved over the last decade to provide machine‐to‐machine (M2M) interactions. There are many services in the IoT environment with similar functional but with different non-functional properties or Quality of Service requirements. Thus, the problem of selecting and combining the optimal service that matches the end-user constraints is challenging. Therefore, this research firstly reviews and analyzes the state-of-the-art algorithms for service selection process under QoS constraints in an IoT environment. That review aims to provide the research community with guidance and knowledge to identify the trends in the services selection problem. Secondly, the research proposed a model to select the optimal service that meets the end-user preferences. The proposed solution gradates into three models, where the last model includes the completed proposed solution for solving the selection problem. The firstintroduced model used a friendly Likert type measurement method with an Improved Practical Swarm Optimization (Improved-PSO)algorithm to enhance the end-user requirements.The second model called RI-SSO, which implemented a Likert type measurement methodto evaluate the reputation of the services from the end-users. It aims to improve bio-inspired optimizing algorithms, called a Social Spider Optimization (SSO) algorithm, by adding a reputation value to member’s weight. The third model aims to classify the search space into sub-clusters. So when the end-user requires a service, the searching process is done in the appropriate cluster. For applying search space clustering, the model used a Fuzzy Logic System (FLS). Lastly, the comparative study and experimental results were done to evaluate the proposed algorithm’s performance with other related service selection models.The feasible efficiency has been obtained as output results from the proposed models in both condition maximization and minimization optimization states. They achieved better performance in terms of fitness values and execution time