Advancing Robust Optimization Methodologies to Foster Energy Efficiency in Future Wireless Networks, Funded by DFG, (2012-2015)

Future wireless networks consume a drastically increasing amount of energy and therefore, contribute a growing amount to the CO2 emission. Besides efficient radio resource management, a sophisticated planning is essential to exploit the potential for energy savings already in the planning phase of a wireless network. So far, existing methodologies for wireless network planning do not incorporate non-deterministic factors of the real problem. However, data uncertainty arises frequently as users move around, data rate requirements fluctuate, or path losses between transmitter and receiver change by external influences. Robust optimization is a novel mathematical methodology to regard these uncertainties in the planning model. This project advanced robust optimization methodologies for the application to the planning of energy-efficient future wireless networks comprising uncertain factors. The main objective is the advancement of recoverable robustness, a two-stage robust optimization approach, since we believe that this approach has a high potential for energy savings. To this end, we aim at studying different aspects of robust optimization, such as gamma-robustness, and combining the achievements in a recoverable robust network planning. Additionally, we tackled the severity of the high complexity of the resulting optimization models via a stepwise consideration of the required variables, i.e., via column generation. Finally, since the incorporation of interference is challenging, we studied this aspect separately.



Anke Schmeink

© INDA at RWTH Aachen