Abstract
M.Sc.
This dissertation aims to describe a context-aware, service-sharing, generic smart- computing environment that uses a model to create a local network profile of all connected devices and available services. The local profile is uploaded and synchronized with the context-aware, service-sharing, generic smart environment system, which has no geographical boundaries or limitations. The model detects new devices and creates a local profile for each device. The data within the profile depicts the services available. The context-aware, service-sharing, generic smart-computing environment is a combination of multiple networks and local network profiles used to control the way users interact with the devices within their physical and virtual environments.
The absence of shared services among devices is a continuous problem in the Internet of Things. The model thus attempts to solve this problem by facilitating the sharing of services across multiple networks in a smart environment. A multi-agent system in which each agent profiles a device in the smart environment is used to develop service profiles in each network in the environment. The profiling of services by each agent is implemented by using Immune-inspired algorithms. The agent context-aware, service-sharing, generic smart-computing environment model uses a negative selection service detection model to identify available services on an agent and shares this profile with all connected agents.