No abstract

Despite the abundance of the optimization models proposed so far in the literature to improve the performance or to reduce the deployment costs of wireless mesh networks, only a few works have addressed explicitly the minimization of spectrum utilization in these networks. Considering the scarcity of this critical resource mainly caused by the tremendous recent developments in wireless technologies, we address in this paper the problem of minimizing the number of used radio channels while satisfying a set of end-to-end traffic demands. We formulate to this end a mixed-integer linear program as a cross-layer optimization approach acting on various aspects, such as channel assignment, power control, medium access and flow routing. Due to the expensive computational costs required to find exact solutions, we propose various approximation schemes that provide near-optimal results in a reasonable amount of time. We illustrate the impact of various parameters such as radio patterns, requested traffic loads, network density on the total required spectrum. Results show how a knowledgeable exploitation of spatial reuse can save significantly spectrum utilization.

No abstract

We investigate in this paper the implications of scheduling and adaptive modulation and coding (AMC) on the performance of downlink traffic in OFDMA-based IEEE 802.16 cells. To this end, we model the downlink traffic as a discrete-time Markov chain by assuming exponential interarrival times of packets at the upstream of the unidirectional unbuffered connection maintained by the base station with each associated subscriber station (SS). Various performance metrics are derived by considering two usual scheduling policies: equal-slot sharing (ESS) and equal-throughput sharing (ETS). In particular, the estimated packet access delays provide a lower bound on the achieved delays in case of buffering. Besides, our model considers the adaptivity of the modulation and coding schemes (MCSs) used by the base station. By assuming uniform distribution of the SSs over the cell and shadowing log-normal radio propagation, we derive the probability density function of the signal-to-noise ratios measured at the level of the SSs. Albeit defined in the context of WiMAX networks, this second contribution can be exploited in more generic wireless settings. Several numerical results are presented and discussed to expose the effects of system parameters and scheduling policies on performance.