An AI-based system for telecommunication network planning

Today´s telecoms market is filled with severe competition and tight budget. Many network operators are striving for different ways to reduce their operating cost while still providing excellent services to their customers. One of the possible savings that can be made is to apply software planning tools to automate network designs at the lowest possible cost. In this paper, we describe an AI based system which can automate/optimize the planning process. It facilitates the implementation of different optimization algorithms to solve network design problems. A case study of physical layer design for FTTH networks is provided. FTTH is considered to be the last mile of access network which is critical to provide bandwidth-intensive services such as high speed Internet access, video on demand, online gaming, etc. In this paper, an overview of the system is provided. The problem definition of the FTTH network design and three optimization approaches (i.e. Mixed Integer Linear Programming (MILP), Ant Colony Optimization(ACO) and Genetic Algorithm (GA)) are described. Finally, different sizes of networks will be used to illustrate the performance of each approach.