This article considers a topology construction problem involving a fixed-wing autonomous aerial vehicle (AAV), a set of quad-rotor AAVs and mobile ground users. The constructed topology must maximize the max-min flow rate of ground users over a given planning horizon. To this end, we outline a mixed integer linear program (MILP) that jointly optimizes the trajectory of the fixed-wing AAV, placement of quad-rotor AAVs, and routing of traffic from each source-destination ground user pair. Solving the MILP is challenging because it requires an exhaustive collection of topologies. To this end, this article outlines a solution called rollout to determine the network topology in each time slot of a given planning horizon iteratively. The main idea of rollout is to generate a sequence of future decisions using a heuristic, where a decision corresponds to the placement of quad-rotor AAVs. Further, each sequence of decisions has a cost-to-go value. Rollout then selects the sequence with the highest cost-to-go value. The results show that the max-min flow rate achieved by rollout is on average 81% that of MILP.

This paper considers a network design problem using Unmanned Aerial Vehicles (UAVs). It aims to create a network to provide communication and computation service to a set of source-destination ground node pairs. The main performance metric is the minimum amount of computed data among a set of source-destination pairs. To optimize this metric, we outline two mixed Integer Linear Programs (MILPs), namely S-MILP and NS-MILP, which are designed respectively for splittable and non-splittable traffic flow models. They jointly optimize the placement of UAVs, assignment of Virtualized Network Functions (VNFs), and routing of unprocessed and processed flow. Further, NS-MILP optimizes the path selection of each source-destination pair. A key challenge is that these MILPs require an exhaustive collection of network topologies. To this end, this paper outlines two heuristic algorithms, called Resource-Aware Location Selection (RALS) and Resource-Aware Path and Location Selection (RAPLS), respectively for each traffic flow model. The simulation results show that RALS and RAPLS achieve on average 83% and 80% of the amount of computed flow of S-MILP and NS-MILP, respectively. Lastly, RALS and RAPLS require 45% and 53% less computation time as compared to S-MILP and NS-MILP, respectively.

With the rapid development of Machine Learning (ML), Machine-Learning-as-a-Service (MLaaS) clusters appear in large numbers to support cloud platforms services, which adopt virtual machine (VM) to improve the availability, resilience and security. However, low energy efficiency is a major problem in such clusters. Previous work focused on reducing the number of physical machines by centralizing resources migration. Nevertheless, for ML tasks with frequent memory switching, blind migration is not worth the cost because the remaining time is less than the migration time, since the migration time can not be ignore due to the memory intensive of ML tasks. Therefore, this paper explores how the remaining time and memory replacement states in ML tasks, which we summarize as residue, affect migration, and proposes an online residue-aware migration algorithm based on Lyapunov optimization. Through rigorous proof, the gap between the algorithm and the optimal solution is ensured. Extensive simulations show that the proposed algorithm is better than the previous migration.

User-perceived quality of experience (QoE) is critical as mobile video streaming experiences a substantial growth. User's demands are becoming diversified where accelerated play-back is the preference of a considerable part of users. However, the limited and fluctuate mobile bandwidth is often not capable of satisfying user's demand of watching video at 2x or higher speed because of consequential frequent rebuffering. Previous adaptive bitrate (ABR) algorithms hardly consider the variety of user playback rates. In this work, we fully exploit the relation between user-perceived, i.e., subjective video quality and the characteristic of video content. The result of our motivational experiments shows that viewers are less sensitive to the bitrate variation and playback rate alternation if there is higher degree of motion in the video. With above guidelines, we adaptively adjust the quality configuration and playback rate to significantly reduce the rebuffering while achieving similar or even higher subjective quality. Then we formulate subjective quality and playback rate adaption as a QoE maximization problem and propose the content based subjective quality and playback rate adaptation algorithm (CSP) utilizing Lyapunov optimization technique. Via rigorous proof, the time-average QoE achieved by CSP is in$O(1/V)$gap compared to optimal value, where$V$is the control parameter. Extensive evaluations confirm the superiority of our proposed algorithm over other state-of-the-art algorithms under both normal and accelerated playback rate.