Network Optimization Methods for Self-Organization of Future Cellular Networks

Abstract

This chapter discusses network optimization methods for enabling self-organization in current cellular networks such as Long Term Evolution (LTE)/LTE-Advanced (LTE-A), and the upcoming 5G networks. Discrete and continuous optimization models are discussed for developing distributed algorithms for self-configuration and self-optimization. The focus is on Self-Organized Networking (SON) problems, which are relevant to small cell networks. Examples include Physical Cell-ID (PCI) assignment, Primary Component Carrier (PCC) selection, Inter-Cell Interference Coordination (ICIC), and network synchronization. A conflict-graph model is considered for PCI assignment and PCC selection problems, which paves the way for different graph coloring algorithms with self-organizing properties. Algorithms for self-organized ICIC and network synchronization are also developed in a principled manner, through a network utility maximization framework. This systematic approach leads to a variety of algorithms which adhere to self-organization principles, but have varying requirements in terms of inter-cell coordination and computation complexity. Fully distributed self-organizing algorithms do not involve any inter-cell dedicated message-passing, and thus are faster and more scalable than the ones that are distributed but require local coordination via exchange of messages between cells. However, local coordination enables higher network utility and better convergence properties.