Zhenzhen Xu

Cyber-Physical Control Over Wireless Sensor and Actuator Networks with Packet Loss

There is a growing interest in design and implementation of cyber-physical control systems over wireless sensor and actuator networks (WSANs). Thanks to the use of wireless communications and distributed architectures, these systems encompass many advantages as compared to traditional networked control systems using hard wirelines. While WSANs are enabling a new generation of control systems, they also introduce considerable challenges to quality-of-service (QoS) provisioning. In this chapter, we examine some of the major QoS challenges raised by WSANs, including resource constraints, platform heterogeneity, dynamic network topology, and mixed traffic. These challenges make it difficult to fulfill the requirements of cyber-physical control in terms of reliability and real time. The focus of this chapter is on addressing the problem of network reliability. Specifically, we analyze the behavior of wireless channels via simulations based on a realistic link-layer model. Packet loss rate (PLR) is taken as a major metric for the analysis. The results confirm the unreliability of wireless communications and the uncertainty of packet loss over WSANs. To tackle packet loss, we present a simple solution that can take advantage of existing prediction algorithms. Simulations are conducted to evaluate the performance of several classical prediction algorithms used for loss compensation. The results give some insights into how to deal with packet loss in cyber-physical control systems over unreliable WSANs.

Exploiting Publication Contents and Collaboration Networks for Collaborator Recommendation

Thanks to the proliferation of online social networks, it has become conventional for researchers to communicate and collaborate with each other. Meanwhile, one critical challenge arises, that is, how to find the most relevant and potential collaborators for each researcher? In this work, we propose a novel collaborator recommendation model called CCRec, which combines the information on researchers’ publications and collaboration network to generate better recommendation. In order to effectively identify the most potential collaborators for researchers, we adopt a topic clustering model to identify the academic domains, as well as a random walk model to compute researchers’ feature vectors. Using DBLP datasets, we conduct benchmarking experiments to examine the performance of CCRec. The experimental results show that CCRec outperforms other state-of-the-art methods in terms of precision, recall and F1 score.

Mining Advisor-Advisee Relationships in Scholarly Big Data: A Deep Learning Approach

Mining advisor-advisee relationships can benefit many interesting applications such as advisor recommendation and protege performance analysis. Based on the hypothesis that, advisor-advisee relationships among researchers are hidden in scholarly big data, we propose in this work a deep learning based advisor-advisee relationship identification method which considers the personal properties and network characteristics with a stacked autoencoder model. To the best of our knowledge, this is the first time that a deep learning model is utilized to represent coauthor network features for relationships identification. Moreover, experiments demonstrate that the proposed method has better performance compared with other state-of-the-art methods.

Exploiting Trust and Usage Context for Cross-Domain Recommendation

Cross-domain recommender systems are usually able to suggest items, which are not in the same domain, where users provided ratings. For this reason, cross-domain recommendation has attracted more and more attention in recent years. However, most studies propose to make cross-domain recommendation in the scenario, where there are common ratings between different domains. The scenario without common ratings is seldom considered. In this paper, we propose a novel method to solve the cross-domain recommendation problem in such a scenario. We first apply trust relations to the cross-domain scenario for predicting coarse ratings pertaining to cross-domain items. Then, we build a new rating matrix, including known ratings and predicted ratings of items from different domains, and transform a user-item matrix into an item-item association matrix. Finally, we compute the similarities of items belonging to different domains and use item-based collaborative filtering to generate recommendations. Through relevant experiments on a real-world data set, we compare our method to a trust-aware recommendation method and demonstrate its effectiveness in terms of prediction accuracy, recall, and coverage.

Bio-inspired packet dropping for ad-hoc social networks

SUMMARY Ad-hoc social networks (ASNETs) explore social properties of nodes in communications. The usage of various social applications in a resource-scarce environment and the dynamic nature of the network create unnecessary congestion that might degrade the quality of service dramatically. Traditional approaches use drop-tail or random-early discard techniques to drop data packets from the intermediate node queue. Nonetheless, because of the unavailability of the social properties, these techniques are not suitable for ASNETs. In this paper, we propose a Bio-inspired packet dropping (BPD) algorithm for ASNETS. BPD imitates the matching procedure of receptors and epitopes in immune systems to detect congestions. The drop probability settings depend on the selection of data packets, which is based on node popularity level. BPD selects the most prioritized node through social properties, which is inspired by the B-cell stimulation in immune systems. To fairly prioritize data packets, two social properties are used: (1) similarity and (2) closeness centrality between nodes. Extensive simulations are carried out to evaluate and compare BPD to other existing schemes in terms of mean goodput, mean loss rate, throughput, delay, attained bandwidth, and overhead ratio. The results show that the proposed scheme outperforms these existing schemes. Copyright

Meta-heuristic algorithms for parallel identical machines scheduling problem with weighted late work criterion and common due date

To our knowledge, this paper investigates the first application of meta-heuristic algorithms to tackle the parallel machines scheduling problem with weighted late work criterion and common due date (

Multi-Robot Dynamic Task Allocation Using Modified Ant Colony System

P|{{d}_{j}}=d|{{Y}_{w}}

Urban traffic congestion estimation and prediction based on floating car trajectory data

P|dj=d|Yw). Late work criterion is one of the performance measures of scheduling problems which considers the length of late parts of particular jobs when evaluating the quality of scheduling. Since this problem is known to be NP-hard, three meta-heuristic algorithms, namely ant colony system, genetic algorithm, and simulated annealing are designed and implemented, respectively. We also propose a novel algorithm named LDF (largest density first) which is improved from LPT (longest processing time first). The computational experiments compared these meta-heuristic algorithms with LDF, LPT and LS (list scheduling), and the experimental results show that SA performs the best in most cases. However, LDF is better than SA in some conditions, moreover, the running time of LDF is much shorter than SA.