• rate
• robust optimization
• wireless
• wlan

This Thesis aims to evaluate robust optimization under users mobility and rate uncertainty in Dense WLANs; this work is a continuation of what has already begun by A. Massida. In that Thesis, he shall exploit the robust optimization paradigm to deal with uncertainty aspects in real application contexts, such as the Green WLAN. The concept of Local Area Network is the first step to the WLANs, which introduce the use of mobile devices and, as a consequence, the deployment of this kind of infrastructure. These networks are implemented with several access points (APs) that provide wireless connectivity to user terminals (UTs) located in the area. In this scenario, it is possible to switch off some APs guaranteeing the same UTs coverage as if the network was working at full power. The idea is to connect UTs to the minimum number of APs, in order to power off some of them to save energy. This problem must take into account two uncertain factors, like wireless instability and user mobility; hence, we should take into account that users roam across the service area. This last feature has a notable impact to channel capacity, which in turn strongly depends on the UT-AP distance. The models described in this work are evaluated exploiting CPLEX solver, it enables rapid development and deployment of decision optimization models using mathematical and constraint programming. First of all, we investigate the problem of switch off some APs to minimize power consumption in dense WLANs, granting power savings.
The just mentioned problem can be formulated as an Integer Linear Programming (ILP) and solved by CPLEX. Anyway, we can have diffculties to solve this problem in short amount of time, as required in real WLAN environments; a proper setting of CPLEX parameters can help us to compute solutions in a reasonable time with reasonable power savings. Then we investigate a heuristic to manipulate the number of access points that are not switched off, and finally, we implement a simple reallocation algorithm to increase the number of feasible future solutions after CPLEX evaluation.