Full-duplex access points (APs) deployment can significantly affect network performance of a wireless local area network (WLAN). Unlike in traditional half-duplex networks, location of a full-duplex AP will affect network coverage quality as well as full-duplex transmission opportunities. However, the effect of full-duplex AP deployment on network performance and the differences between half- and full-duplex AP deployment have not been well investigated yet. In this paper, we first theoretically analyze the effect of full-duplex AP deployment on WLAN throughput. Exact full-duplex transmission probability is derived in presence of Rayleigh fading with different AP locations. Our analysis reveal that a good AP deployment profile can exploit more full-duplex transmission opportunities and greatly improve network performance. The full-duplex AP deployment problem is then formulated as an integer linear programming (ILP) problem in which our objective is to obtain optimized network throughput. Then we develop a heuristic algorithm to solve the formulated problem and optimal deployment profile can be produced. Simulation results validate that the WLAN throughput as well as full-duplex transmission opportunities can be significantly improved by our generated full-duplex AP deployment profile.

GPUs have become the dominant computing units to meet the need of high performance in various computational fields. But the long operation latency causes the underutilization of on-chip computing resources, resulting in performance degradation when running parallel tasks on GPUs. A good warp scheduling strategy is an effective solution to hide latency and improve resource utilization. However, most current warp scheduling algorithms on GPUs ignore the ability of long operations to hide latency. In this paper, we propose a long-operation-first warp scheduling algorithm, LFWS, for GPU platforms. The LFWS filters warps in the ready state to a ready queue and updates the queue in time according to changes in the status of the warp. The LFWS divides the warps in the ready queue into long and short operation groups based on the type of operations in their instruction buffers, and it gives higher priority to the long-operating warp in the ready queue. This can effectively use the long operations to hide some of the latency from each other and enhance the system's ability to hide the latency. To verify the effectiveness of the LFWS, we implement the LFWS algorithm on a simulation platform GPGPU-Sim. Experiments are conducted over various CUDA applications to evaluate the performance of LFWS algorithm, compared with other five warp scheduling algorithms. The results show that the LFWS algorithm achieves an average performance improvement of 8.01% and 5.09%, respectively, over three traditional and two novel warp scheduling algorithms, effectively improving computational resource utilization on GPU.

This paper studies the impact of integrating pricing with connection admission control (CAC) on the congestion management practices in contention-based wireless random access networks. Notably, when the network is free of charge, each self-interested user tries to occupy the channel as much as possible, resulting in the inefficient utilization of network resources. Pricing is therefore adopted as incentive mechanism to encourage users to choose their access probabilities considering the real-time network congestion level. A Stackelberg leader-follower game is formulated to analyze the competitive interaction between the service provider and the users. In particular, each user chooses the access probability that optimizes its payoff, while the self-interested service provider decides whether to admit or to reject the user's connection request in order to optimize its revenue. The stability of the Stackelberg leader-follower game in terms of convergence to the Nash equilibrium is established. The proposed CAC scheme is completely distributed and can be implemented by individual access points using only local information. Compared to the existing schemes, the proposed scheme achieves higher revenue gain, higher user payoff, and higher QoS performance.

This letter has retrained an MQDF classifier on the retraining set, which is constructed by samples locating near classification boundary. The method is evaluated on HCL2000 and HCD Chinese handwriting sets. The results show that the retrained MQDF outperforms MQDF and cascade MQDF on all test sets.