User-centric learning for multiple access selections

Abstract

We are in the age where business growth is based on how user-centric your services or goods is. Current research on wireless system is more focused on ensuring that user could achieve optimal throughput with minimal delay, disregarding what user actually wants from the services. Looking from con-nectivity point of view, especially in urban areas these days, there are multiple mobile and wireless access that user could choose to get connected to. As people are looking toward machine automa-tion, we understand that the same could be done for allowing users to choose services based on their own requirement. This paper looks into unconventional, non-disruptive approach to provide mobile services based on user requirements. The first stage of this study is to look for user association from three new perspectives. The second stage involved utilizing a reinforcement learning algorithm known as q-learning, to learn from feedbacks to identify optimal decision in reaching user-centric requirement goal. The outcome from the proposed deployment has shown significant improvement in user association with learning aware solution © BEIESP.