Multiagent approach for resource management in Smart Environments

摘要

The dynamic nature of Smart Environments, enabled thanks to the Internet of Things (IoT) paradigm, requires the application of innovative resource discovery/selection services. Recommender systems provide useful and customized information, properly selected and filtered, for users and services. This paper proposes a multiagent approach for building a recommendation system in smart environments. Smart objects, enhanced IoT objects interacting among them and with users, are mapped through real-valued vectors obtained by word embedding techniques. The vectors are assigned to bio-inspired agents that, working autonomously, bring out a set of similarity grouped entities – clusters of smart objects with similar features – thus enabling an efficient and dynamic recommendation/selection system. Experiment results on synthetic and real datasets prove the effectiveness of the approach.