Search

Abstract The sample data for reinforcement learning algorithms often exhibit sparsity and instability, making the training results susceptible to falling into local optima. Mini-Max-Multi-agent Deep Deterministic Policy Gradient (M3DDPG) algorithm is a multi-agent reinforcement learning algorithm, which introduces the minimax theorem into Multi-Agent Deep Deterministic Policy Gradient (MADDPG) algorithm. It also has unstable convergence caused by sparse sample data and randomization. However, the Particle Swarm Optimisation (PSO) algorithm, unlike traditional reinforcement learning methods, involves the construction of independent populations of policy networks to generate sample data, followed by training the reinforcement learning algorithm. PSO optimizes and updates the policy population based on a fitness function, aiming to enhance the efficiency and convergence speed of the algorithm in learning from the sample data. In order to address the multi-agent pursuit-evasion problem, we propose the PSO-M3DDPG algorithm, which combines the PSO algorithm with the M3DDPG algorithm. Through experimental simulations, the improved algorithm demonstrates superior training results and faster convergence speeds, thus validating its effectiveness.

Leptin has been widely studied and found to have a significant impact on the development of osteoarthritis (OA). However, there are conflicting findings regarding the impact of leptin on chondrocytes. The study aimed to examine the impact of leptin on human chondrocytes and rats with OA. In the in vitro experiment, cartilage tissue obtained from patients hospitalized for knee replacement due to OA was collected for primary culture of chondrocytes. The proliferation and apoptosis of chondrocytes were assessed using cell counting kit-8 and flow cytometry. Autophagy levels were evaluated through monodansylcadaverine staining, mRFP-GFP-LC3 fluorescence, and transmission electron microscopy. Additionally, the expression of autophagy-related genes and proteins was analyzed using qRT-PCR and western blotting. In the in vivo experiment, an OA rat model was established. Following treatment with leptin and leptin antagonists, the cartilage tissues were examined using histology analysis (hematoxylin-eosin and Safranin O/fast green staining) and immunohistochemical. Mankin's score was utilized to assess the severity of OA, while qRT-PCR and western blotting were employed to detect the expression of autophagy-related genes and proteins in the cartilage. The ability of leptin to protect chondrocytes is achieved through the inhibition of autophagy via phosphoinositide 3 kinase/protein kinase B/mammalian target of rapamycin signaling pathway.

Abnormality detection of power generation equipment is of great significance in enhancing equipment reliability. In the era of Industry 4.0, with the rapid development and extensive application of IoT and sensor technology, large-scale sensing devices are deployed in power plants, resulting in a vast amount of sensor streaming data. The emergence of edge computing enables the streaming data processing and computation on edge devices. It reduces the latency of streaming data processing and improves throughput. However, those edge nodes are heterogeneous, decentralized, and capacity-constrained edge nodes. Moreover, the conventional cloud-edge model exhibits some characteristics such as tight coupling of data and computing, hard to deal with temporally varying continuity, and lack of flexibility. It poses significant challenges to detect normality and provide anomaly warning of power generation equipment. In order to deal with those challenges, we propose a service-oriented approach for fusing cloud-edge capabilities. Firstly, we encapsulated the streaming data and its processing into suitable granular service, serving as fundamental units for basic streaming data processing. These services can be deployed independently and flexibly scheduled in cloud-edge environment to facilitate the development of IoT application systems by developers. The services help to decouple the streaming data and its computing. Secondly, we proposed an event-driven mechanism to enable dynamic collaboration among services, allowing proactive response to events and adaptive adjustment of logic for streaming data processing to cope with the dynamics of IoT and time-varying logic of streaming data processing. This enhances service deployment flexibility, reduces latency in streaming data processing and improved the efficiency. Finally, based on the actual scenario of the fire power plants, we validated the feasibility and effectiveness in detecting equipment abnormalities by using our proposed cloud-edge fusion approach. We compared our proposed approach with three typical streaming data processing architectures including Cloud, iFogSim and PureEdgeSim from processing latency and system throughput. It reduced the latency remarkably with an average reduction of about 78%, 22%, and 16%. It improved system throughput of about 57%, 17%, and 14%.

Abstract With the continuously increasing penetration of electric vehicles (EVs), the mutual match between the distribution of charging resources and the spatial–temporal distribution of EV charging demands is becoming increasingly important. To address this, this paper proposes a novel two‐stage customized EV charging–navigation strategy. Building on previous research on the real‐time information from dynamic traffic networks, a personalized dynamic road impedance (PDRI) model is built to transform three main criteria (distance, time, and finance) affecting charging–navigation into comprehensive road impedance. In the first navigation stage, fast‐charging stations (FCSs) with the lowest overall objective are selected. In the second navigation stage, an improved Floyd–Warshall algorithm is utilized to identify the routes with the lowest personalized weight to the selected FCS in the PDRI model. Notably, the personalized preferences of EV drivers for the three primary criteria are considered in both stages of the navigation process. Finally, simulation results demonstrate a significant improvement in the degree of matching between charging navigation plans and drivers' personalized requirements, and a more balanced spatial–temporal distribution of EV charging demands among FCSs, which verifies the effectiveness of the proposed strategy.

Recently, developing multi-UAVs to cooperatively pursue a fast-moving target has become a research hotspot in the current world. Although deep reinforcement learning (DRL) has made a lot of achievements in the UAV pursuit game, there are still some problems such as high-dimensional parameter space, the ease of falling into local optimization, the long training time, and the low task success rate. To solve the above-mentioned issues, we propose an improved twin delayed deep deterministic policy gradient algorithm combining the genetic algorithm and maximum mean discrepancy method (GM-TD3) for multi-UAV cooperative pursuit of high-speed targets. Firstly, this paper combines GA-based evolutionary strategies with TD3 to generate action networks. Then, in order to avoid local optimization in the algorithm training process, the maximum mean difference (MMD) method is used to increase the diversity of the policy population in the updating process of the population parameters. Finally, by setting the sensitivity weights of the genetic memory buffer of UAV individuals, the mutation operator is improved to enhance the stability of the algorithm. In addition, this paper designs a hybrid reward function to accelerate the convergence speed of training. Through simulation experiments, we have verified that the training efficiency of the improved algorithm has been greatly improved, which can achieve faster convergence; the successful rate of the task has reached 95%, and further validated UAVs can better cooperate to complete the pursuit game task.

Background: The development of biomaterials capable of accelerating bone wound repair is a critical focus in bone tissue engineering. This study aims to evaluate the osteointegration and bone regeneration potential of a novel multilayer gelatin-supported Bone Morphogenetic Protein 9 (BMP-9) coated nano-calcium-deficient hydroxyapatite/poly-amino acid (n-CDHA/PAA) composite biomaterials, focusing on the material-bone interface, and putting forward a new direction for the research on the interface between the coating material and bone. Methods: The BMP-9 recombinant adenovirus (Adenovirus (Ad)-BMP-9/Bone Marrow Mesenchymal Stem Cells (BMSc)) was produced by transfecting BMSc and supported using gelatin (Ad-BMP-9/BMSc/Gelatin (GT). Multilayer Ad-BMP-9/BMSc/GT coated nano-calcium deficient hydroxyapatite/polyamino acid (n-CDHA/PAA) composite biomaterials were then prepared and co-cultured with MG63 cells for 10 days, with biocompatibility assessed through microscopy, Cell Counting Kit-8 (CCK-8), and alkaline phosphatase (ALP) assays. Subsequently, multilayer Ad-BMP-9/BMSc/GT coated n-CDHA/PAA composite biomaterial screws were fabricated, and the adhesion of the coating to the substrate was observed using scanning electron microscopy (SEM). In vivo studies were conducted using a New Zealand White rabbit intercondylar femoral fracture model. The experimental group was fixed with screws featuring multilayer Ad-BMP-9/BMSc/GT coatings, while the control groups used medical metal screws and n-CDHA/PAA composite biomaterial screws. Fracture healing was monitored at 1, 4, 12, and 24 weeks, respectively, using X-ray observation, Micro-CT imaging, and SEM. Integration at the material-bone interface and the condition of neo-tissue were assessed through these imaging techniques. Results: The Ad-BMP-9/GT coating significantly enhanced MG63 cell adhesion, proliferation, and differentiation, while increasing BMP-9 expression in vitro. In vivo studies using a rabbit femoral fracture model confirmed the biocompatibility and osteointegration potential of the multilayer Ad-BMP-9/BMSc/GT coated n-CDHA/PAA composite biomaterial screws. Compared to control groups (medical metal screws and n-CDHA/PAA composite biomaterial screws), this material demonstrated faster fracture healing, stronger osteointegration, and facilitated new bone tissue formation with increased calcium deposition at the material-bone interface. Conclusion: The multilayer GT-supported BMP-9 coated n-CDHA/PAA composite biomaterials have demonstrated favorable osteogenic cell interface performance, both in vitro and in vivo. This study provides a foundation for developing innovative bone repair materials, holding promise for significant advancements in clinical applications.

Aiming at the difficulty of measuring various costs and time-consuming elements in multimodal transport, this paper constructs a green vehicle comprehensive multimodal transport model which incorporates transportation, transit, quality damage, fuel consumption, and carbon emission costs and proposes a hybrid embedded time window to calculate the time penalty cost in order to reflect the actual transport characteristics. Furthermore, in order to better solve the model, a hybrid sand cat swarm optimization (HSCSO) algorithm is proposed by introducing Logistic–Tent chaotic mapping and an adaptive lens opposition-based learning strategy to enhance the global search capability, and inspired by the swarm intelligence scheme, a momentum–bellicose strategy and an equilibrium crossover pool are introduced to improve the search efficiency and convergence ability. Through testing nine benchmark functions, the HSCSO algorithm exhibits superior convergence accuracy and speed in dealing with complex multi-dimensional problems. Based on the excellent global performance, the HSCSO algorithm was utilized for multimodal vehicle transportation in East China, and a path with a lower comprehensive cost was successfully planned, which proved the effectiveness of the HSCSO algorithm in green intermodal transport path planning.

Using dual polarization multiplexing alternate mark inversion (AMI) downlink signals, a novel radio over fiber (RoF) system integrating optical fiber and FSO channel is designed to adapt to applications in mountainous areas and other complex terrain areas. Optical heterodyne technology and self-mixing homodyne detection method are used to realize high sensitivity detection of the received signals after 25.1 km channel (including 1 km single-mode fiber and 100 m free space link) transmission. Moreover, polarization multiplexing technology is introduced to exponentially increase the transmission capacity of downlink signals. This scheme not only can be compatible with traditional optical fiber transmission systems, but also support the wireless optical access application of millimeter wave signals in RoF systems.

ObjectiveTo explore the characteristics of spontaneous brain activity changes in patients with lumbar disc herniation (LDH), and help reconcile the contradictory findings in the literature and enhance the understanding of LDH-related pain.Materials and methodsPubMed, Web of Science, Embase, Chinese National Knowledge Infrastructure (CNKI), SinoMed, and Wanfang databases were searched for literature that studies the changes of brain basal activity in patients with LDH using regional homogeneity (ReHo) and amplitude of low-frequency fluctuation/fraction amplitude of low-frequency fluctuation (ALFF/fALFF) analysis methods. Activation likelihood estimation (ALE) was used to perform a meta-analysis of the brain regions with spontaneous brain activity changes in LDH patients compared with healthy controls (HCs).ResultsA total of 11 studies were included, including 7ALFF, 2fALFF, and 2ReHo studies, with a total of 269 LDH patients and 277 HCs. Combined with the data from the ALFF/fALFF and ReHo studies, the meta-analysis results showed that compared with HCs, LDH patients had increased spontaneous brain activity in the right middle frontal gyrus (MFG), left anterior cingulate cortex (ACC) and the right anterior lobe of the cerebellum, while they had decreased spontaneous brain activity in the left superior frontal gyrus (SFG). Meta-analysis using ALFF/fALFF data alone showed that compared with HCs, LDH patients had increased spontaneous brain activity in the right MFG and left ACC, but no decrease in spontaneous brain activity was found.ConclusionIn this paper, through the ALE Meta-analysis method, based on the data of reported rs-fMRI whole brain studies, we found that LDH patients had spontaneous brain activity changes in the right middle frontal gyrus, left anterior cingulate gyrus, right anterior cerebellar lobe and left superior frontal gyrus. However, it is still difficult to assess whether these results are specific and unique to patients with LDH. Further neuroimaging studies are needed to compare the effects of LDH and other chronic pain diseases on the spontaneous brain activity of patients. Furthermore, the lateralization results presented in our study also require further LDH-related pain side-specific grouping study to clarify this causation.Systematic review registrationPROSPERO, identifier CRD42022375513.

Abstract Temporal network link prediction is an important task in the field of network science, and has a wide range of applications in practical scenarios. Revealing the evolutionary mechanism of the network is essential for link prediction, and how to effectively utilize the historical information for temporal links and efficiently extract the high-order patterns of network structure remains a vital challenge. To address these issues, in this paper, we propose a novel temporal link prediction model with adjusted sigmoid function and 2-simplex structure (TLPSS). The adjusted sigmoid decay mode takes the active, decay and stable states of edges into account, which properly fits the life cycle of information. Moreover, the latent matrix sequence is introduced, which is composed of simplex high-order structure, to enhance the performance of link prediction method since it is highly feasible in sparse network. Combining the life cycle of information and simplex high-order structure, the overall performance of TLPSS is achieved by satisfying the consistency of temporal and structural information in dynamic networks. Experimental results on six real-world datasets demonstrate the effectiveness of TLPSS, and our proposed model improves the performance of link prediction by an average of 15% compared to other baseline methods.

Underwater robots often encounter the influence of confined underwater environments during underwater exploration. These environments include underwater caves, sunken ships, submerged houses, and pipeline structures. Robot positioning in these environments is strongly disturbed, leading not only to the failure of some commonly used positioning methods but also to an increase in errors in positioning systems that normally function well in open water. In order to overcome the limitations of positioning methods in confined underwater environments, researchers have studied different underwater positioning methods and have selected suitable methods for positioning in such environments. These methods can achieve high-precision positioning without relying on assistance from other platforms and are referred to as autonomous positioning methods. Autonomous positioning methods for underwater robots mainly include SINS/DR positioning and SLAM positioning. In addition, in recent years, researchers have developed some bio-inspired autonomous positioning methods. This article introduces applicable robot positioning methods and sensors in confined underwater environments and discusses the research directions of robot positioning methods in such environments.

An evolutionary perspective enhances our understanding of biological mechanisms. Comparison of sex determination and X-chromosome dosage compensation mechanisms between the closely related nematode species Caenorhabditis briggsae (Cbr) and Caenorhabditis elegans (Cel) revealed that the genetic regulatory hierarchy controlling both processes is conserved, but the X-chromosome target specificity and mode of binding for the specialized condensin dosage compensation complex (DCC) controlling X expression have diverged. We identified two motifs within Cbr DCC recruitment sites that are highly enriched on X: 13 bp MEX and 30 bp MEX II. Mutating either MEX or MEX II in an endogenous recruitment site with multiple copies of one or both motifs reduced binding, but only removing all motifs eliminated binding in vivo. Hence, DCC binding to Cbr recruitment sites appears additive. In contrast, DCC binding to Cel recruitment sites is synergistic: mutating even one motif in vivo eliminated binding. Although all X-chromosome motifs share the sequence CAGGG, they have otherwise diverged so that a motif from one species cannot function in the other. Functional divergence was demonstrated in vivo and in vitro. A single nucleotide position in Cbr MEX can determine whether Cel DCC binds. This rapid divergence of DCC target specificity could have been an important factor in establishing reproductive isolation between nematode species and contrasts dramatically with the conservation of target specificity for X-chromosome dosage compensation across Drosophila species and for transcription factors controlling developmental processes such as body-plan specification from fruit flies to mice.

The vibrating wire strain sensors have been widely used in the field of stress and strain monitoring of oil and gas pipelines. However, different types of vibrating wire sensors vary greatly in service performance due to the influence of pipeline operating conditions and installation processes. In order to understand the actual service life of this type of sensors in strain monitoring of oil and gas pipelines and compare the error of monitoring data, the spot-weldable and arc-weldable vibrating wire strain sensors were studied and continuous monitoring was conducted to the stress state of gas pipelines in the goaf area of Qufu Coal Mine. Through comparative analysis to the 12-year continuous monitoring data, it is found that the survival rate of arc-weldable sensors decreases linearly with time, while the error of monitoring data increases linearly, resulting in a large cumulative measuring error. This is because the arc-weldable sensors are installed on the anticorrosive coating, cannot obtain the real stress of pipeline directly, and may fall off due to the influence of the external factors. Compared with the arc-weldable sensors, the spot-weldable sensors can truly reflect the stress-strain state of pipelines with the measurement data, have a longer service life, and meet the requirements of continuous monitoring for the minimum survival rate of sensors.

Magnesium (Mg) and its alloys have attracted attention as potential biodegradable materials in orthopedics due to their mechanical and physical properties, which are compatible with those of human bone. However, the effect of the mismatch between the rapid material degradation and fracture healing caused by the adverse effect of hydrogen (H2), which is generated during degradation, on surrounding bone tissue has severely restricted the application of Mg and its alloys. Thus, the development of new Mg alloys to achieve ideal degradation rates, H2 evolution and mechanical properties is necessary. Herein, a novel Mg–1Zn–1Sn-xSr (x = 0, 0.2, 0.4, and 0.6 wt%) quaternary alloy was developed, and the microstructure, mechanical properties, corrosion behavior and biocompatibility in vitro/vivo were investigated. The results demonstrated that a minor amount of strontium (Sr) (0.2 wt %) enhanced the corrosion resistance and mechanical properties of Mg–1Zn–1Sn alloy through grain refinement and second phase strengthening. Simultaneously, due to the high hydrogen overpotential of tin (Sn), the H2 release of the alloys was significantly reduced. Furthermore, Sr-containing Mg–1Zn–1Sn-based alloys significantly enhanced the viability, adhesion and spreading of MC3T3-E1 cells in vitro due to their unique biological activity and the ability to spontaneously form a network structure layer with micro/nanotopography. A low corrosion rate and improved biocompatibility were also maintained in a rat subcutaneous implantation model. However, excessive Sr (>0.2 wt %) led to a microgalvanic reaction and accelerated corrosion and H2 evolution. Considering the corrosion resistance, H2 evolution, mechanical properties and biocompatibility in vitro and in vivo, Mg–1Zn–1Sn-0.2Sr alloy has tremendous potential for clinical applications.

As for the characteristics of various types of retinal diseases and uncertainty of the location of the lesions, a retinal multi-disease screening and recognition method based on deep convolutional ensemble network was proposed.Firstly, the black borders on both sides of the retinal fundus image were cut off, and the noise in the image was removed to reduce the interference to the retinal image and increase the clarity of the image.After that, data augmentation methods such as cropping and rotating were performed to process retinal fundus image to amplify the dataset.Then, a model based on deep convolutional neural network was built for feature extraction, and the network model was fine-tuned to complete the task of screening and identifying retinal diseases.Finally, the results of multiple models were ensembled.The experimental results show that this method has achieved good results for the screening and recognition of retinal diseases, the accuracy of retinal disease screening is 96.05%, and the accuracy of retinal disease recognition is 72.55%.

In recent years, Web of Science has published nearly one hundred reports per year on quinoxalin-2(1H)-ones, which have attracted great interest due to their wide applications in pharmaceutical and materials fields, especially in recyclable heterogeneous catalytic reactions for direct C–H functionalisation. This review summarises for the first time the methods and reaction mechanisms of heterogeneous catalytic reactions of quinoxalin-2(1H)-ones, including six major types of heterogeneous catalysts involved. The heterogeneous reactions of quinoxalin-2(1H)-ones are summarised by classifying different types of catalytic materials (graphitic phase carbon nitride, MOF, COF, ion exchange resin, piezoelectric materials, and microsphere catalysis). In addition, this review discusses the future development of heterogeneous catalytic reactions of quinoxalin-2(1H)-ones, including the construction of C-B/Si/P/RF/X/Se bonds by heterogeneous catalytic reactions, the enrichment of heterogeneous catalysts such as metal oxides, graphene-based composites, doped metal nanoparticles, and molecular sieve-based porous materials, asymmetric synthesis, and other areas. The aim of this review is to contribute to the development of green and sustainable heterogeneous reaction methods for quinoxalin-2(1H)-ones with applications in materials chemistry and pharmacology.

The direct C–H multifunctionalization of quinoxalin-2(1H)-ones via multicomponent reactions has attracted considerable interest due to their diverse biological activities and chemical profile. This review will focus on recent achievements. It mainly covers reaction methods for the simultaneous introduction of C–C bonds and C–RF/C/O/N/Cl/S/D bonds into quinoxalin-2(1H)-ones and their reaction mechanisms. Meanwhile, future developments of multi-component reactions of quinoxalin-2(1H)-ones are envisaged, such as the simultaneous construction of C–C and C–B/SI/P/F/I/SE bonds through multi-component reactions; the construction of fused ring and macrocyclic compounds; asymmetric synthesis; green chemistry; bionic structures and other fields. The aim is to enrich the methods for the reaction of quinoxalin-2(1H)-ones at the C3 position, which have rich applications in materials chemistry and pharmaceutical pharmacology.

The growing penetration of renewable energy resources (RESs) increases uncertainties in active distribution networks (ADNs), leading the networks to suffer low inertia, bidirectional power flows, and rapid power changes. The frequency overshoots, overvoltage, and unbalance power sharing are prone to happen in ADNs. To address those issues, this paper proposes a novel unified modeling and a hierarchical control strategy for different distributed generators (DGs). First, to unify the control states of different DGs and separate control states of different control layers, a unified modeling method for grid-following and grid-forming DGs is proposed. Then, based on the unified model, a two-layer control strategy is designed. The robust controllers are locally configured to achieve rapid recovery control, avoiding the impact of communication delays. The collaborative power controller is centralized designed to prevent power circulation caused by the interaction of local controllers of DGs. Compared with other hierarchical control strategies, our strategy can make further use of the residual capacities of different DGs and can more rapidly suppress the unpredictable changes of RESs. The effectiveness of the proposed strategy is validated by several cases in MATLAB/Simulink.

Abstract Background Diffusion tensor imaging (DTI) is an effective method to identify subtle changes to normal-appearing white matter (WM). Here we analyzed the DTI data with other examinations, including motor evoked potentials (MEPs), histopathological images, and behavioral results, to reflect the lesion development in different degrees of spinal cord injury (SCI) in acute and subacute stages. Method Except for 2 Sprague -Dawley rats which died from the anesthesia accident, the rest 42 female rats were randomized into 3 groups: control group (n = 6), moderate group (n = 18), and severe group (n = 18). Moderate (a 50-g aneurysm clip with 0.4-mm thickness spacer) or severe (a 50-g aneurysm clip with no spacer) contusion SCI at T8 vertebrae was induced. Then the electrophysiological assessments via MEPs, behavioral deterioration via the Basso, Beattie, and Bresnaha (BBB) scores, DTI data, and histopathology examination were analyzed. Results In this study, we found that the damage of WM myelin, MEPs amplitude, BBB scores and the decreases in the values of fractional anisotropy (FA) and axial diffusivity (AD) were more obvious in the severe injury group than those of the moderate group. Additionally, the FA and AD values could identify the extent of SCI in subacute and early acute SCI respectively, which was reflected in a robust correlations with MEPs and BBB scores. While the values of radial diffusivity (RD) showed no significant changes. Conclusions Our data confirmed that DTI was a valuable in ex vivo imaging tool to identify damaged white matter tracts after graded SCI in rat, which may provide useful information for the early identification of the severity of SCI.

The high penetration rate of electric vehicles (EVs) will have impacts on the operation of power system in the near future. One of the most significant impacts is the extreme load valley caused by “consistency” in off-grid actions of EVs automatically finishing their charging, which normally occurs at night. The dramatic load reduction at limited period not only leads to security problem of power system due to the deficiency of downward reserve in generation side, but also forces the thermal units to operate in the extremely uneconomical deep peak regulation (DPR) state. To address this issue, a coordinated optimization strategy considering nighttime DPR state of units and uncertain EV demand response (DR) is proposed in this paper. Firstly, unlike traditional studies, a detailed unit commitment (UC) model considering life loss and oil cost is established in this paper. After that, on the basis of elastic pricing method and day-ahead contract, price-based demand response (PDR) and incentive-based demand response (IDR) model are proposed to avoid units operating in the DPR state. The uncertainty of both IDR and PDR progress are considered as a whole in the form of opportunity constrain, and then transformed into deterministic linear constraints for solving by adopting sequence operations. Finally, various cases are simulated in the improved IEEE 30-bus system, the results indicate that the presented coordinated optimization strategy can reduce the units' operating cost effectively, as well as bring benefits to EV owners.

With the development of artificial intelligence and integrated sensor technologies, unmanned aerial vehicles (UAVs) are more and more applied in the air combats. A bottleneck that constrains the capability of UAVs against manned vehicles is the autonomous maneuver decision, which is a very challenging problem in the short-range air combat undergoing highly dynamic and uncertain maneuvers of enemies. In this paper, an autonomous maneuver decision model is proposed for the UAV short-range air combat based on reinforcement learning, which mainly includes the aircraft motion model, one-to-one short-range air combat evaluation model and the maneuver decision model based on deep Q network (DQN). However, such model includes a high dimensional state and action space which requires huge computation load for DQN training using traditional methods. Then, a phased training method, called “basic-confrontation”, which is based on the idea that human beings gradually learn from simple to complex is proposed to help reduce the training time while getting suboptimal but efficient results. Finally, one-to-one short-range air combats are simulated under different target maneuver policies. Simulation results show that the proposed maneuver decision model and training method can help the UAV achieve autonomous decision in the air combats and obtain an effective decision policy to defeat the opponent.

The automatic standardization of nomenclature for anatomical structures in radiotherapy (RT) clinical data is a critical prerequisite for data curation and data-driven research in the era of big data and artificial intelligence, but it is currently an unmet need. Existing methods either cannot handle cross-institutional datasets or suffer from heavy imbalance and poor-quality delineation in clinical RT datasets. To solve these problems, we propose an automated structure nomenclature standardization framework, 3D Non-local Network with Voting (3DNNV). This framework consists of an improved data processing strategy, namely, adaptive sampling and adaptive cropping (ASAC) with voting, and an optimized feature extraction module. The framework simulates clinicians' domain knowledge and recognition mechanisms to identify small-volume organs at risk (OARs) with heavily imbalanced data better than other methods. We used partial data from an open-source head-and-neck cancer dataset to train the model, then tested the model on three cross-institutional datasets to demonstrate its generalizability. 3DNNV outperformed the baseline model, achieving higher average true positive rates (TPR) over all categories on the three test datasets (+8.27%, +2.39%, and +5.53%, respectively). More importantly, the 3DNNV outperformed the baseline on the test dataset, 28.63% to 91.17%, in terms of F1 score for a small-volume OAR with only 9 training samples. The results show that 3DNNV can be applied to identify OARs, even error-prone ones. Furthermore, we discussed the limitations and applicability of the framework in practical scenarios. The framework we developed can assist in standardizing structure nomenclature to facilitate data-driven clinical research in cancer radiotherapy.

White-light OLED devices play an important application in information display fields. Optical interference of the microcavity structure has an important effect on device performances. According to the design of the band structure, ITO/MoO3 composite films were used as the anode, and Mg : Ag (1%) composite films were prepared by coevaporation as the translucent cathode; CuPc was used as the hole injection layer and anode passivation layer, NPB as the hole transmission layer and yellow light main material, rubrene as yellow dopant material, ADN as blue light main material, DSA-Ph as blue dopant material, and TPBi and Alq3 as the electron transport layers. We realized the change of the microcavity structure by adjusting the thickness of each organic functional layer film and simulated and calculated the optimized thickness of each organic film layer and influence on OLED device performances using the SimOLED software system. The optimized OLED microdisplay structure is Si(CMOS)/ITO (35 nm)/MoO3 (2 nm)/CuPc (5 nm)/2-TNATA (20 nm)/NPB (10 nm)/NPB : rubrene (1.5%)ADN : DSA-Ph (5%) (25 nm)/TPBi (15 nm)/Alq3 (1.2 nm)/Mg (13 nm) : Ag (1%). The optimized OLED microdisplay was prepared by the vacuum coating system, and the photoelectric performances of the OLED device were characterized by a spectral testing system consisting of the Photo Research PR655 spectrometer and Keithley 2400 program-controlled power supply. The effect of the microcavity structure on OLED device performances was studied. The results show that the variation of the film thickness of each organic functional layer has an important effect on the performances of OLED microdisplay, such as brightness and color coordinate, and the OLED microdisplay reaches a higher brightness of 3342 cd/m2 under the normal working voltage at 5.0 V after the structure is optimized, with CIE coordinate (0.28, 0.37), which is closer to the energy point of standard white light.

Link prediction tasks have an extremely high research value in both academic and commercial fields. As a special case, link prediction in bipartite graphs has been receiving more and more attention thanks to the great success of the recommender system in the application field, such as product recommendation in E-commerce and movie recommendation in video sites. However, the difference between bipartite and unipartite graphs makes some methods designed for the latter inapplicable to the former, so it is quite important to study link prediction methods specifically for bipartite graphs. In this paper, with the aim of better measuring the similarity between two nodes in a bipartite graph and improving link prediction performance based on that, we propose a motif-based similarity index specifically for application on bipartite graphs. Our index can be regarded as a high-order evaluation of a graph’s local structure, which concerns mainly two kinds of typical 4-motifs related to bipartite graphs. After constructing our index, we integrate it into a commonly used method to measure the connection potential between every unconnected node pair. Some of the node pairs are originally unconnected, and the others are those we select deliberately to delete their edges for subsequent testing. We make experiments on six public network datasets and the results imply that the mixture of our index with the traditional method can obtain better prediction performance w.r.t. precision, recall and AUC in most cases. This is a strong proof of the effectiveness of our exploration on motifs structure. Also, our work points out an interesting direction for key graph structure exploration in the field of link prediction.

Mg-Zn-Sn-based alloys are widely used in the industrial field because of their low-cost, high-strength and heat-resistant characteristics. However, their application in the biomedical field has been rarely reported. In the present study, biodegradable Mg-1Zn-1Sn and Mg-1Zn-1Sn-0.2Sr alloys were fabricated. Their microstructure, surface characteristics, mechanical properties and bio-corrosion properties were carried out using an optical microscope (OM), X-ray diffraction (XRD), electron microscopy (SEM), mechanical testing, electrochemical and immersion test. The cell viability and morphology were studied by cell counting kit-8 (CCK-8) assay, live/dead cell assay, confocal laser scanning microscopy (CLSM) and SEM. The osteogenic activity was systematically investigated by alkaline phosphatase (ALP) assay, Alizarin Red S (ARS) staining, immunofluorescence staining and quantitative real time-polymerase chain reaction (qRT-PCR). The results showed that a small amount of strontium (Sr) (0.2 wt.%) significantly enhanced the corrosion resistance of the Mg-1Zn-1Sn alloy by grain refinement and decreasing the corrosion current density. Meanwhile, the mechanical properties were also improved via the second phase strengthening. Both Mg-1Zn-1Sn and Mg-1Zn-1Sn-0.2Sr alloys showed excellent biocompatibility, significantly promoted cell proliferation, adhesion and spreading. Particularly, significant increases in ALP activity, ARS staining, type I collagen (COL-I) expression as well as the expressions of three osteogenesis-related genes (runt-related transcription factor 2 (Runx2), osteopontin (OPN), and osteocalcin (Bglap)) were observed for the Mg-1Zn-1Sn-0.2Sr group. In summary, this study demonstrated that Mg-Zn-Sn-based alloy has great application potential in orthopedics and Sr is an ideal alloying element of Mg-Zn-Sn-based alloy, which optimizes its corrosion resistance, mechanical properties and osteoinductive activity.

Dye-sensitized solar cells are regarded as promising candidates to resolve the energy and environmental issues in recent years, arising from their solution-processable fabrication technology and high power conversion efficiency. However, there are still several problems regarding how to accelerate the development of this type of photovoltaics, including the limited light-harvesting ability and high-production cost of molecular dye. In the current work, we have systematically studied the role of nitrogen-doped carbon quantum dots (N-CQDs) as co-sensitizers in traditional dye sensitized solar cells. A series of N-CQDs have been prepared by employing chitosan as a precursor via one-pot hydrothermal technology for various times, demonstrating a maximized efficiency as high as 0.089% for an only N-CQDs-based device. Moreover, the co-sensitized solar cell based on N719 dye (C58H86N8O8RuS2) and optimized N-CQDs shows significantly enhanced performance, yielding a solar-to-electric conversion efficiency of up to 9.15% under one standard sun (AM 1.5G) irradiation, which is much higher than the 8.5%-efficiency of the controlled device without N-CQDs. The matched characteristics of energy level, excellent up-convention, and FRET (Förster resonance energy transfer) abilities of N-CQDs are responsible for their improved power conversion efficiency.

Stick-slip vibration presents one of the main problems in the quality of drilling performance limiting tool life and productivity. This type of vibration can be suppressed by means of many approaches, such as varying parameters and use of control tools. Although tremendous improvements have been made in overcoming this dysfunction, stick-slip vibration suppression remains a large problem in the drilling industry. This paper provides an up-to-date review of stick-slip vibration behavior in drillstrings. First, the phenomena and the modeling methods of stick-slip vibrations are reviewed. Then an overview of the approaches for stick-slip suppression in oilwell drillstrings is presented, grouping the references under the categories of passive vibration control and active vibration control. Literature related to passive control is grouped under the categories of optimization of bottom hole assembly (BHA) configurations, bit selection and bit redesign, and use of downhole tools. The contributions related to the active control approaches for stick-slip mitigation are grouped under the categories of drilling parameters optimization based on real-time measurement and use of active control systems. Finally, related discussions and recommendations are conducted.