Complex Agent-based Modeling for HetNets Design and Optimization

In wireless heterogeneous networks (HetNets), complexity is an intrinsic property. This paper presents agent-based modeling (ABM) as a tool to optimize complex HetNets. We introduce and analyze a HetNet ABM model that employs parallel algorithms for interference management, resource allocation, and load balancing at both micro and macro levels. Two reinforcement learning (RL) algorithms jointly work together in the model to resolve co-tier and cross-tier interferences. The first RL algorithm controls the transmission power of the small cells, whereas the second assigns the users to the sub-bands with less interference levels. Concurrently, the user association is decided by the users based on their preferences and the resources available at the cells. The model is analyzed in three different operation modes, by switching processes on and off. Results show that individual processes contribute to overall system performance, while jointly maximizing the network's aggregate signal-to-interference-and-noise ratio (SINR) and minimizing load-induced latency by efficient load balancing.