.3 G Network, Radio Access Network Optimization Design

With the development of GSM, mobile communication system is going to the 3rd generation mobile communication system-Universal Mobile Telecommunication System (UMTS). The network based on UMTS has established , and the concerned mobile telephones have appeared。 By the year of 2001, 16 countries have given the service providers 61 licenses in the world.UMTS use a new high rate radio access technology-Code Division Multiple Access (CDMA). However, in the core network, the UMTS system is based on the GSM network. And the GSM network has been confirmed to be successful. Nowadays, GSM network contains Circuit-Switched tongue and Packet-Switched data platform, and the data service supported by Packet-Switched data platform include short message, browsing mobile web page and mobile e-mail.With the explosive increase of Chinese mobile subscribers and the burgeoning of Chinese communication industry, its mobile communication networks kept spreading and upgrading, which make it a highlight to implement the network planning. How to minimize the financial cost of the infrastructure without sacrificing QOS are the key topic of the survival of mobile service provider and also the key topic of this thesis. After introducing the basic structure of UMTS and UTRAN and its related theories, the mathematical model of cost function of UTRAN was defined. And it is taken for granted that this problem is a NP-hard problem. The Genetic Algorithms were applied to solve this problem since there are many advantages in them. Two-dimension binary system was adopted as the coding method to represent the base station placement precisely. Some new operators, such as sorting operator, random number transformation operator and delete operator, were also inspired to accelerate the convergence without leading to premature.In the chapter 4, we improve the model which given in chapter 2, in order to reducing the size of search space and cutting down the search time. We compare Tabusearch algorithms, Simulated Annealing algorithms and Greedy algorithms in the improved model. All these make sense in field practice.Because of the limitation of the time and some impersonal conditions, we only use the basic Genetic Algorithms in the Assignment problem of Radio Access Network. And there are many topics which are not studied including some advance technologies and improved Genetic Algorithms such as Parallel Genetic Algorithms, NSGA-n etc. Further study on these topics will be done in future work.