Applying Evolutionary Algorithms to Solve the Automatic Frequency Planning Problem
Summary
Frequency assignment is a well-known problem in operations research for which different mathematical models exist depending on the application-specific conditions. However, most of these models are far from considering actual technologies currently deployed in GSM networks, such as frequency hopping. In these networks, interferences provoked by channel reuse due to the limited available radio spectrum result in a major impact of the quality of service (QoS) for subscribers. Therefore, frequency planning is of great importance for GSM operators. We here focus on optimizing the frequency planning of a realistic-sized, real-world GSM network by using evolutionary algorithms (EAs). Results show that a (1+10) EA developed by the chapter authors for which different seeding methods and perturbation operators have been analyzed is able to compute accurate and efficient frequency plans for real-world instances.
