queue-aware transmission scheduling for cooperative wireless communications with sub-fading-block scheduling to better balance load and capacity in low mobility environments is investigated. The scheduling problem for...
详细信息
queue-aware transmission scheduling for cooperative wireless communications with sub-fading-block scheduling to better balance load and capacity in low mobility environments is investigated. The scheduling problem for joint cooperation scheduling and resource allocation is formulated as a constrained nonlinear integer optimization problem over an integer convex set based on a source buffer queueing analysis. It is shown that with queue-aware scheduling, the state transition matrix of the source buffer queue has a highly dynamic form. As a result, the objective function of the optimization problem does not have an analytic form in general. The constrained discrete Rosenbrock search algorithm, which is a gradient-free directed discrete search algorithm, is employed to solve the nonlinear integer problem. The output of the directed integer search algorithm is used for queue-aware transmission scheduling for the cooperative system. Numerical results are presented which show that, for cooperative transmission scheduling, the Rosenbrock search based queue-aware algorithm significantly outperforms the equal partitioning, random partitioning, and gradient-based algorithms under quasi-static channel assumptions. Under practical system conditions with unsaturated traffic, the proposed queue-aware scheduling scheme achieves the true optima, and maintains a large stability region for the buffer queue, over a wide range of channel and traffic conditions. It is also shown that when fading channel dynamics are taken into consideration, the performance of the proposed queue-aware scheduling algorithm significantly outperforms fixed relaying and fixed direct transmission channel-awarescheduling strategies.
The recent development in mobile edge computing necessitates caching of dynamic contents, where new versions of contents become available around-the-clock, thus timely update is required to ensure their relevance. The...
详细信息
The recent development in mobile edge computing necessitates caching of dynamic contents, where new versions of contents become available around-the-clock, thus timely update is required to ensure their relevance. The age of information (AoI) is a performance metric that evaluates the freshness of contents. Existing works on AoI-optimization of cache content update algorithms focus on minimizing the long-term average AoI of all cached contents. Sometimes, user requests that need to be served in the future are known in advance and can be stored in user request queues. In this paper, we propose dynamic cache content update scheduling algorithms that exploit the user request queues. We consider a use case, where the trained neural networks (NNs) from deep learning models are being cached in a heterogeneous network (HetNet), as a motivating example. A queue-aware cache content update scheduling algorithm based on constrained Markov decision process (CMDP) is developed to minimize the average AoI of the dynamic contents delivered to the users. By using enforced decomposition technique and deep reinforcement learning, we propose two low-complexity suboptimal scheduling algorithms. Simulation results show that our proposed algorithms outperform the periodic cache content update scheme and reduce the average AoI by up to 30%.
Recently, attention has been paid to the integration of opportunistic communications, whether based on opportunistic user selection (OUS) or opportunistic antenna selection, with interference alignment (IA) in order t...
详细信息
Recently, attention has been paid to the integration of opportunistic communications, whether based on opportunistic user selection (OUS) or opportunistic antenna selection, with interference alignment (IA) in order to improve the performance of wireless networks. In OUS, users that have the best operational conditions are usually selected. However, fairness among users is another important aspect that should be considered in scheduling users. In this paper, a queue-aware two-stage opportunistic IA (OIA) algorithm is proposed for the downlink multicell multiuser multiple-input-multiple-output system. In the first stage, inter-cell interference is eliminated using one pair of precoding/postcoding matrices. Whereas in the second stage, two user selection polices are proposed namely, capacity-based selection (CBS) and queue-based scheduling (QBS), to select a group of users and minimize the inter-user interference among them using another pair of precoding/postcoding matrices. In the QBS-OIA case, a joint scheduling, resource allocation, and IA optimization problems are formulated, and a low complexity heuristic is proposed to solve it. Comparisons are conducted with other OIA algorithms in terms of achieved sum rate, achieved degrees-of-freedom (DoFs), number of served users, queue overflow probability, and computational complexity. Simulations show that the two proposed CBS-OIA and QBS-OIA algorithms outperform other schemes in terms of sum rate and DoFs. Moreover, the proposed QBS-OIA is capable of serving more users, in some cases, and achieves lower overflow probability with much reduced complexity on the expense of achieving a bit lower sum rate than the CBS-OIA in some cases.
暂无评论