Green network planning and operational power consumption optimization in LTE-A using artificial intelligence

Demands for high bandwidth and data rates put wireless communication industry in a leading position as energy demanding industry which contributes in carbon dioxide volume. Therefore, Green wireless network concept has emerged to provide the foundation of energy-efficient wireless network. Base stations in wireless network are considered as the major part responsible for the network power consumption. Therefore, this work is dedicated to introduce a new perspective for the heterogeneous network power efficiency improvement through network planning optimization and relay station switching. A cascaded multi-objective genetic algorithm network optimization (CMOGANO) is developed to optimize the network number of base station, their location and configuration in the first stage to provide full coverage. The second stage of the developed algorithm optimizes the average number of relay station per base station and their distance from the respective base station to meet the capacity constraint of the network operator. To optimize the relay station switching, a detailed mathematical model assuming linear power consumption model for the transmitters in the network is developed. In this model, the rate of active relay stations is defined and integrated in the model as a varying function in time. Optimization of the rate of active relay stations, in this model, is treated as a variation concept. A simplified fuzzy logic solution to the optimization of the rate of active RS is introduced. The CMOGANO optimization showed that a power reduction up to 40% is feasible in the network by reducing the number of base stations by 47% as compared to the operator plan. This reduction of the number of base stations is achieved without losses in the network coverage. The second stage of CMOGANO added a total of 516 relay stations in the network to improve its capacity up to 98%. The rate of active relay station optimization process revealed that the optimum rate of active relay station obeys a linear first order ODE. Moreover, the optimum rate of active relay is a function of the traffic pattern, average relay station load factor, their derivatives, and the relative RS to BS capacity factor. The mathematical model yielded a significant power saving up to 30% and 46% can be achieved in RS idling and sleeping modes respectively. Moreover, the amount of power saving is a function of the number of relay stations per base station. The effect of the traffic rate derivative is to activate the relay stations for longer time resulting in less power saving in the network. Similarly, power saving up to 41% can be achieved by the fuzzy logic solution. In conclusion, complying with green network concept in heterogeneous network systems is feasible. This compliance starts from adopting the green network concept in the planning stage of the systems and continues through the optimization of the network power expenditure during its operation. However, the introduction of the relay base stations in heterogeneous network significantly increased the complexity of the network. This complexity can be handled by developing a sophisticated heuristic mathematical tools that account for diverse parameters in the network such that the demanded Quality of Service (QoS) is achieved under the umbrella of green network operation.

Text FK 2015 78 - IR.pdf Download (1MB)

Actions (login required)

View Item