Publications

An aggregative game with local constraint sets is studied in this article, where each player's cost function is dependent on the aggregation function that is unavailable to all players. To compute the Nash equilibrium (NE) point of the game in a distributed manner, the players are endowed with several auxiliary state variables that are used to estimate the aggregation function by exchanging their estimates with local neighbors on a directed graph. In the two cases with strongly connected weight-balanced and weight-unbalanced directed graphs, respectively, NE seeking strategies are proposed by the interconnection of projected gradient-play with average consensus dynamics. The proposed algorithms are proved to be able to reach the NE point by using tools from variational inequality theory and Lyapunov stability theory. Finally, an example is simulated to demonstrate the theoretical results.

We study the phenomenon of spatial coherence resonance (SCR) on Hodgkin–Huxley (HH) neuronal networks that are characterized with information transmission delay. In particular, we examine the ability of additive Gaussian noise to optimally extract a particular spatial frequency of excitatory waves in diffusive and small-world networks on which information transmission amongst directly connected neurons is not instantaneous. On diffusively coupled HH networks, we find that for short delay lengths, there always exists an intermediate noise level by which the noise-induced spatial dynamics is maximally ordered, hence implying the possibility of SCR in the system. Importantly thereby, the noise level warranting optimally ordered excitatory waves increases linearly with the increasing delay time, suggesting that extremely long delays might nevertheless preclude the observation of SCR on diffusive networks. Moreover, we find that the small-world topology introduces another obstacle for the emergence of ordered spatial dynamics out of noise because the magnitude of SCR fades progressively as the fraction of rewired links increases, hence evidencing decoherence of noise-induced spatial dynamics on delayed small-world HH networks. Presented results thus provide insights that could facilitate the understanding of the joint impact of noise and information transmission delay on realistic neuronal networks.

Diffusion behavior of Nb in elemental powder metallurgy high Nb containing TiAl alloys was investigated. The results show that Nb element dissolves into the matrix by diffusion. Pore nests are formed in situ after Nb diffusion. With the increase in hot pressing temperature, the diffusion of Nb will be more sufficient, and the microstructure is more homogeneous. Nb element diffuses completely at 1400°C. Meanwhile, compression deformation and agglomeration phenomena of pores are observed in some pore nests. Hot isostatic pressing (HIP) treatment can only efficiently decrease but not eliminate porosity completely.

Fatty acid composition and content in seed oil of plants in Chungbaishan region was studied by gas chromatography and reported.

Laser coagulation is a minimally invasive therapy that utilizes laser energy to thermally kill benign and malignant lesions such as cancers, at the temperature range of 55–85°C. It is of clinical importance to control the laser deposition into the tissue in such a way that the lesion will be destroyed while the surrounding healthy tissue will remain intact. However, a primary technical difficulty in achieving this goal lies in the fact that the relationship between the delivered laser energy and the tissue damage is nonlinear and time-varying, which cannot be accurately predicted or rigorously modeled due to the significant difference in various physical properties of even similar tissues. In this paper, we present a novel real-time ultrasound-guided fuzzy laser control system for coagulation. Current status of tissue coagulation depth, noninvasively measured by an innovative ultrasound system that we recently developed, was fed into a fuzzy proportional-derivative (PD) controller, which periodically adjusted output power of a 1064 nm Nd:YAG laser. The ultrasound-guided system was tested in 21 in vitro experiments in which fresh sheep liver samples were irradiated by the laser with a coagulation setpoint ranging from 4 to 14 mm with a 2 mm increment. We provide analytical analysis and design of the fuzzy controller, which turns out to be an inherently nonlinear PD controller with self-tuning variable gains. We also present the hardware and software implementation of the entire measurement and control system. Our control system is unique, and it is the first laser control system that is guided by noninvasive ultrasonic measurement in real-time.

Graphs are naturally used to describe the structures of various real-world systems in biology, society, computer science etc., where subgraphs or motifs as basic blocks play an important role in function expression and information processing. However, existing research focuses on the basic statistics of certain motifs, largely ignoring the connection patterns among them. Recently, a subgraph network (SGN) model is proposed to study the potential structure among motifs, and it was found that the integration of SGN can enhance a series of graph classification methods. However, SGN model lacks diversity and is of quite high time complexity, making it difficult to widely apply in practice. In this paper, we introduce sampling strategies into SGN, and design a novel sampling subgraph network model, which is scale-controllable and of higher diversity. We also present a structural feature fusion framework to integrate the structural features of diverse sampling SGNs, so as to improve the performance of graph classification. Extensive experiments demonstrate that, by comparing with the SGN model, our new model indeed has much lower time complexity (reduced by two orders of magnitude) and can enhance a series of graph classification methods.