Cost based optimization for strategic mobile radio access network planning using metaheuristics

Resumo

Mobile communication technologies have experimented a great evolution since their appearance in the early 90’s. This evolution has been motivated by two main factors: the new applications and necessities demanded by the end-user, and the technology progress. Mobile services have evolved from the classical voice and short message services (SMS), to more attractive and therefore better accepted services such as video-telephony, video-streaming, online gaming, and the mobile broadband access service (MBAS), which provides internet access on global mobility. All this improvements are a reality due to the appearance of new channel access methods, modulation and coding schemes, application of multiple input multiple output (MIMO) transmission and reception techniques, etc. The liberalization of the sector has also played an important role in the evolution and current situation of mobile communications. Before this liberalization process, mobile communications’ market presented a monopolistic scheme, based on public network infrastructure. However, liberalization caused the appearance of multiple companies providing services, with the necessity of new network infrastructure deployment. This new liberalized framework makes necessary the existence of a National Regulatory Authority (NRA) to establish a fair competitive market situation. One of the main objectives of the NRA is to determine the interconnection costs. Interconnection makes possible the communication between users from different mobile operators, as well as the access to the whole set of services. Few mobile operators make use of network of a third partner to provide a specific service to the end-user. The goal of this Ph. D. Thesis is to minimize the investment cost associated to the network equipment. The interconnection charges determined by the NRA depends, to some extent, on this invested cost, as we show in the different chapters of this work. To achieve this objective, the thesis is divided into two main parts: first, a set of network dimensioning algorithms is developed. The second part is focused on the design and implementation of metaheuristic optimization algorithms to solve the optimal service distribution over the considered technologies problem (OSDP). The planning and dimensioning module is based on four different algorithms. These algorithms are able to perform the dimensioning and planning process of a mobile radio access network (RAN), in a multi-technology (second (2G), third (3G) and fourth (4G) generations are considered), multi-user profile, based on the existence of three different type of users (customer, business and premium users), and multi-service environment (covering from the classical voice and SMS services, low binary rate data services (64-144 Kbps) to the mobile broadband access service). The second part of this thesis is focused on solving the OSDP. The goal of this part is to distribute the traffic demand over the different technologies to be deployed. To achieve this objective, an efficient use of the technologies considered to reduce the investment costs incurred on network equipment, is needed. This is due to the technical features of each specific mobile generation systems, i.e., second generation systems are mainly focused on the provision of voice service and SMS, while third generation systems are data-oriented networks. Finally, the enhancement of the third generation system, known as High Speed Data Access (HSPA), together with 4G systems are focused on the provision of high speed data services such as the MBAS. We have tested the proposed algorithms for solving the OSDP in a set of different scenarios. The results provided by these algorithms have been used to carry out techno-economical studies such as the performance analysis of the HSPA and 4G systems providing the MBAS, the LTE deployment simulation in the 800 MHz frequency band, this study is based on two real scenarios reproducing the coverage and capacity conditions of the two largest mobile operators in Spain, Movistar and Vodafone. Moreover, a study simulating the deployment of a LTE network with different spectrum allocation in the 800 MHz, 1800 MHz, and 2600 MHz has been carried out. Improvements, in terms of investment cost and therefore a reduction on services provision costs, have been observed for all the simulated cases. The experiments performed in this thesis are focused on the spanish case. However, the algorithms implemented can be easily adopted to carry out similar studies in any other european country. In fact, the dimensioning algorithms of the first part of this thesis have been successfully applied to several international regulatory projects.