Guoyong Wang’s research while affiliated with Nanjing University of Information Science and Technology and other places

Flying ad hoc networks (FANETs) have highly dynamic and energy-limited characteristics. Compared with traditional mobile ad hoc networks, their nodes move faster and their topology changes more frequently. Therefore, the design of routing protocols faces greater challenges. The existing routing schemes rely on frequent and fixed-interval Hello transmissions, which exacerbates network load and leads to high communication energy consumption and outdated location information. MP-QGRD combined with the extended Kalman filter (EKF) is used for node position prediction, and the Hello packet transmission interval is dynamically adjusted to optimize neighbor discovery. At the same time, reinforcement learning methods are used to comprehensively consider link stability, energy consumption, and communication distance for routing decisions. The simulation results show that compared to QMR, QGeo, and GPSR, MP-QGRD has an increased packet delivery rate, end-to-end latency, and communication energy consumption by 10%, 30%, and 15%, respectively.

Multi-objective optimization (MOO) has become an important method in machine learning, which involves solving multiple competing objective problems simultaneously. Nowadays, many MOO algorithms assume that gradient information is easily available and use this information to optimize functions. However, when encountering situations where gradients are not available, such as black-box functions or non-differentiable functions, these algorithms become ineffective. In this paper, we propose a zeroth-order MOO algorithm named SZMG (stochastic zeroth-order multi-gradient algorithm), which approximates the gradient of functions by finite difference methods. Meanwhile, to avoid conflicting gradients between functions and reduce stochastic multi-gradient direction bias caused by stochastic gradients, an SGD-type method is adopted to acquire weight parameters. Under the non-convex setting and mild assumptions, the convergence rate is established for the SZMG algorithm. Simulation results demonstrate the effectiveness of the SZMG algorithm.

Although deep reinforcement learning has achieved notable practical achievements, its theoretical foundations have been scarcely explored until recent times. Nonetheless, the rate of convergence for current neural temporal-difference (TD) learning algorithms is constrained, largely due to their high sensitivity to stepsize choices. In order to mitigate this issue, we propose an adaptive neural TD algorithm (AdaBNTD) inspired by the superior performance of adaptive gradient techniques in training deep neural networks. Simultaneously, we derive non-asymptotic bounds for AdaBNTD within the Markovian observation framework. In particular, AdaBNTD is capable of converging to the global optimum of the mean square projection Bellman error (MSPBE) with a convergence rate of ${{\mathcal {O}}}(1/\sqrt{K})$, where K denotes the iteration count. Besides, the effectiveness AdaBNTD is also verified through several reinforcement learning benchmark domains.

This paper investigates the simultaneous fault detection and control(SFDC) problem for a class of uncertain nonlinear networked control systems with limited communication. In order to improve the utilization of network resources, a novel dynamic event-triggered mechanism is employed affected by the Bernoulli package dropouts phenomenon, which is usually appearing in practical systems. Then, a fault detection filter and dynamic output feedback controller module is applied to achieve the required fault detection and control objectives simultaneously. The linear matrix inequality-based approach is provided to obtain the designed filter and controller parameters so that the closed-loop system is robustly stochastically stable and meets the required performance. Finally, two practical examples are adopted to show the effectiveness of the developed scheme.

This paper considers a multi-depot heterogeneous vehicle routing problem (MDHVRP) with time windows, which is very crucial for hazardous materials transportation. For this reason, we formalize this problem as a multi-objective MDHVRP optimization model, where the actual load dependent risk of hazardous materials transportation is considered. To solve the optimization problem, we propose a hybrid multi-objective evolutionary algorithm (HMOEA) and a two-stage algorithm (TSA). In addition, we verify the performance of the proposed algorithms by experiments on the modified Solomon’s VRPTW examples. In the experiment, it can be seen from the distribution of Pareto solution sets and the convergence distribution of IGD values that HMOEA is significantly superior to the other three algorithms in searching for Pareto solutions, as well as in the convergence and diversity of the algorithm. HMOEA and TSA were compared, and the minimum cost obtained by TSA was 13.38% lower than HMOEA, while the minimum risk was 81.69% higher than HMOEA. The advantages of each algorithm in finding solutions in reality were analyzed. A comparison was made between multi-depots heterogeneous VRP and multi-depots homogeneous VRP in the C101 instance, and the results showed that scheduling heterogeneous vehicles would reduce risk and cost.

Abstract Zero‐shot learning (ZSL) is to identify target categories without labeled data, in which semantic information is used to transfer knowledge from some seen categories. In the existing Generalized Zero‐Shot Learning (GZSL) methods, domains shift problem always appeared during generating feature stage. In order to solve this problem, a new method to Auto‐Encode the Synthesis Pseudo Features for the GZSL task (AESPF‐GZSL) is proposed in this manuscript. Specifically, the AESPF‐GZSL method trains the generated features under the semantic auto‐encoder framework and exploits attention mechanism to train the generated features again. Then, the generated features are input to the classifier. The proposed method is performed on three benchmark data sets referred as to AWA, CUB and SUN. The experimental results show that the proposed method achieves the state‐of‐the art classifier accuracy both in ZSL and GZSL settings. In ZSL setting, the classification accuracy of our method is superior to the compared algorithms, improved by 0.40% in AWA and 0.30% in SUN, respectively. And in GZSL setting, the classification accuracy of the method is superior to the comparison algorithm 0.41% in Harmonic mean on AWA, and 1.01%, 0.62%, and 1.05% in training data set, testing data set, and harmonic average on SUN.

Named Data Networking (NDN) can effectively deal with the rapid development of mobile video services. For NDN, selecting a suitable forwarding interface according to the current network status can improve the efficiency of mobile video communication and can also avoid attacks to improve communication security. For this reason, we propose a stochastic adaptive forwarding strategy based on deep reinforcement learning (SAF-DRL) for secure mobile video communications in NDN. For each available forwarding interface, we introduce the twin delayed deep deterministic policy gradient algorithm to obtain a more robust forwarding strategy. Moreover, we conduct various numerical experiments to validate the performance of SAF-DRL. Compared with BR, RFA, SAF, and AFSndn forwarding strategies, the results show that SAF-DRL can reduce the delivery time and the average number of lost packets to improve the performance of NDN.