چكيده به لاتين
Recent advances in ICT have led to the emergence of a new concept called the Internet of Things (IoT). IoT is a modern technology in which any creature (human, animal or object) can send and receive data through communication networks. One of the challenges in IoT consists in service composition. Today, different services are offered in IoT according to the needs of users, sometimes users have requests that can not be answered with a single and simple service and need combined services. A composite service is a set of abstract or single tasks or services that are interconnected and can do private work. Therefore, the main issue in composition IoT services is choosing the best candidate service from services that do the same job but have different quality criteria. The issue of service integration, in addition to being able to address users' concerns about access to the need or service they want. It also faces challenges such as security, addressing incomplete resources, describing and measuring service quality features, and interdependence between services, so there must be a way to combine services that meet the needs of users, such as quality parameters.
Studies in this field have been reviewed from different points of view, the results of these studies show that each of the methods studied, in addition to having advantages, also has a series of basic limitations, limitations that can affect the quality of IoT network services. In most of the reviewed methods, the issue of IoT network energy, which is one of the main limitations of this network, has not been considered. In some cases, the delay and time of service combination and response to the user is not considered, which in its place can overshadow the life of the network.
To address these challenges, the aim of this dissertation is to provide a service quality-aware service combination method that, in addition to considering the service quality parameters of users, considers the energy consumption of equipment and the IoT. And provide the desired output compared to the previous methods in accordance with the stated goals. The computational time of the proposed method has been reduced by 4.26%, 3.47% and 4.1%, respectively, compared to the particle swarm optimization, ant optimization and genetic optimization algorithms.
