Intelligent Task Dispatching and Scheduling Using a Deep Q-Network in a Cluster Edge Computing System.

Recently, intelligent IoT applications based on artificial intelligence (AI) have been deployed with mobile edge computing (MEC). Intelligent IoT applications demand more computing resources and lower service latencies for AI tasks in dynamic MEC environments. Thus, in this paper, considering the resource scalability and resource optimization of edge computing, an intelligent task dispatching model using a deep Q-network, which can efficiently use the computing resource of edge nodes is proposed to maximize the computation ability of the cluster edge system, which consists of multiple edge nodes. The cluster edge system can be implemented with the Kubernetes technology. The objective of the proposed model is to minimize the average response time of tasks offloaded to the edge computing system and optimize the resource allocation for computing the offloaded tasks. For this, we first formulate the optimization problem of resource allocation as a Markov decision process (MDP) and adopt a deep reinforcement learning technology to solve this problem. Thus, the proposed intelligent task dispatching model is designed based on a deep Q-network (DQN) algorithm to update the task dispatching policy. The simulation results show that the proposed model archives a better convergence performanc in terms of the average completion time of all offloaded tasks, than existing task dispatching methods, such as the Random Method, Least Load Method and Round-Robin Method, and has a better task completion rate than the existing task dispatching method when using the same resources as the cluster edge system.

Effects of Frontal Theta Rhythms in a Prior Resting State on the Subsequent Motor Imagery Brain-Computer Interface Performance.

Dealing with subjects who are unable to attain a proper level of performance, that is, those with brain-computer interface (BCI) illiteracy or BCI inefficients, is still a major issue in human electroencephalography (EEG) BCI systems. The most suitable approach to address this issue is to analyze the EEG signals of individual subjects independently recorded before the main BCI tasks, to evaluate their performance on these tasks. This study mainly focused on non-linear analyses and deep learning techniques to investigate the significant relationship between the intrinsic characteristics of a prior idle resting state and the subsequent BCI performance. To achieve this main objective, a public EEG motor/movement imagery dataset that constituted two individual EEG signals recorded from an idle resting state and a motor imagery BCI task was used in this study. For the EEG processing in the prior resting state, spectral analysis but also non-linear analyses, such as sample entropy, permutation entropy, and recurrent quantification analyses (RQA), were performed to obtain individual groups of EEG features to represent intrinsic EEG characteristics in the subject. For the EEG signals in the BCI tasks, four individual decoding methods, as a filter-bank common spatial pattern-based classifier and three types of convolution neural network-based classifiers, quantified the subsequent BCI performance in the subject. Statistical linear regression and ANOVA with post hoc analyses verified the significant relationship between non-linear EEG features in the prior resting state and three types of BCI performance as low-, intermediate-, and high-performance groups that were statistically discriminated by the subsequent BCI performance. As a result, we found that the frontal theta rhythm ranging from 4 to 8 Hz during the eyes open condition was highly associated with the subsequent BCI performance. The RQA findings that higher determinism and lower mean recurrent time were mainly observed in higher-performance groups indicate that more regular and stable properties in the EEG signals over the frontal regions during the prior resting state would provide a critical clue to assess an individual BCI ability in the following motor imagery task.

A socially aware routing based on local contact information in delay-tolerant networks.

In delay-tolerant networks, network topology changes dynamically and there is no guarantee of continuous connectivity between any two nodes. These features make DTN routing one of important research issues, and the application of social network metrics has led to the design of recent DTN routing schemes. In this paper, we propose an efficient routing scheme by using a node's local contact history and social network metrics. Each node first chooses a proper relay node based on the closeness to the destination node. A locally computed betweenness centrality is additionally utilized to enhance the routing efficiency. Through intensive simulation, we finally demonstrate that our algorithm performs efficiently compared to the existing epidemic or friendship routing scheme.