دانلود رایگان مقاله به حداقل رساندن تاثیر هنداور برای کاربران تلفن همراه در WLAN

Abstract

IEEE 802.11 based Wireless LANs are an important piece in today’s communication infrastructure in order to provide high speed wireless Internet access to static or quasi mobile users. For large WLAN deployments (i.e., Campus or enterprise WLAN), it is important to understand the impact of user mobility and handovers on the system performance. In this article, we have developed a performance model for a set of networked 802.11 based WLAN Access Points, which is based on a Mixed Integer Linear Program (MILP). The objective function tries simultaneously to maximize the total system rate while at the same time minimizing the number of handovers for a configurable handover signaling rate. Because of the conflicting nature of the two objective functions, such multi-objective optimization is difficult to explore. A detailed evaluation of the model using several scenarios involving both different numbers of static and mobile users shows that our formulation allows trading off those two objectives in a robust way.

7. Conclusion

IEEE 802.11 based WLANs are an important part of the future wireless Internet. Understanding their performance is crucial in order to gain an insight into the different system tradeoffs involved. In this paper, we have built a mathematical model that allows to study the tradeoff between the number of HOs and the total system rate for a set of IEEE 802.11 based access points, that are connected to the Internet. Our model is built on a MILP with the objective of both maximizing the total download rate while at the same time minimizing the number of HOs with a configurable HO signaling rate. Because the objective function is composed of two conflicting terms, we have studied the tradeoff between the two objectives in the optimization function, which allowed us to find a good compromise between download rate and number of HOs. An important aspect of our model is the total knowledge of all system parameters over the whole simulation duration, which is typically not known in advance. As future work, we intend to apply robust optimization techniques on our model in order to cope with unknown or erroneous predicted demands, mobility patterns and download rates. By applying robust optimization techniques we expect to get better insight into the tradeoffs involved when the exact values of several parameters of the system are not precisely known. Also, we intend to study fast heuristics that can be applied when some parameters of the system changes in order to react fast during critical situations.