Multi-agent reinforcement learning provides promising prospect for task scheduling in cloud-edge computing environment in recent years. However, there remains a formidable challenge due to partial observation and the rigid coupling between action spaces and schedulable devices. These limit the ability of agent to perceive global communication patterns and adapt to dynamic environments, resulting in unsatisfactory scheduling decisions. To address these issues, this work proposes SPAD, a novel spatial perception and action decoupling empowered distributed multi-agent AI task scheduling framework. By constructing a global spatial feature distillation mechanism, SPAD can approximate the implicit heterogeneous connection patterns and communication dynamics between devices and tasks under constrained observability, enhancing its ability to make robust decisions in dynamic environments with limited observations. Additionally, SPAD employs a Lyapunov-based action decoupling module to alleviate scalability challenges from rigid action-device coupling, while a novel intrinsic penalty mechanism augments the agent’s advantage function with the instantaneous Lyapunov cost, thereby aligning the policy optimization process with the decoupling module’s underlying stability constraints. Through a comprehensive empirical evaluation spanning synthetic, bursty, and real-world trace-driven workloads, we show that SPAD consistently outperforms state-of-the-art benchmarks in reducing task completion latency and improving resource utilization, while maintaining remarkable resilience and scalability across diverse network topologies and under non-stationary load conditions.

With the rapid development of artificial intelligence technology, generative AI has shown broad application prospects and potential in various fields. Frequent emergencies have put forward higher requirements for traditional Emergency Response Decision Support Systems (ERDSS). This paper proposes a theoretical framework of ERDSS based on generative AI (ERDSS-GAI), which deeply integrates generative AI with three stages of emergency response decision-making. The framework aims to leverage the advantages of generative AI in massive data processing, knowledge mining, strategy optimization, and other aspects, thereby enhancing the intelligence level and emergency response capability of ERDSS. The key components and implementation path of ERDSS-GAI are systematically explained from a theoretical perspective, and its application value is analyzed through the case of rainstorm and flood disaster in Shenzhen. This research demonstrates that generative AI can improve the scientific and refined level of emergency response decision-making, providing a theoretical framework and practical insights for its widespread adoption in emergency management.

Studies of resource provision in cloud computing have drawn extensive attention, since effective task scheduling solutions promise an energy-efficient way of utilizing resources while meeting diverse requirements of users. Deep reinforcement learning (DRL) has demonstrated its outstanding capability in tackling this issue with the ability of online self-learning, however, it is still prevented by the low sampling efficiency, poor sample validity, and slow convergence speed especially for deadline constrained applications. To address these challenges, an Imitation Learning Enabled Fast and Adaptive Task Scheduling (ILETS) framework based on DRL is proposed in this paper. First, we introduce behavior cloning to provide a well-behaved and robust model through Offline Initial Network Parameters Training (OINPT) so as to guarantee the initial decision-making quality of DRL. Next, we design a novel Online Asynchronous Imitation Learning (OAIL)-based method to assist the DRL agent to re-optimize its policy and to against the oscillations caused by the high dynamic of the cloud, which promises DRL agent moving towards the optimal policy with a fast and stable process. Extensive experiments on the real-world dataset have demonstrated that the proposed ILETS can consistently produce shorter response time , lower energy consumption and higher success rate than the baselines and other state-of-the-art methods at the accelerated convergence speed.

Strong nonlinearities under extreme conditions pose intractable challenges for the motion control of Automated Vehicles (AVs). Incapable or inaccurate modeling of nonlinearities, coupled with enormous cost of nonlinear controls, severely limit stability and performance enhancements in these scenarios. This paper proposes a novel modeling and robust control framework to address these issues. First, a novel hybrid modeling approach for trajectory tracking of AVs, combining a prior nominal model and a data-driven uncertain model based on Koopman theory, is proposed to enhance model predictive ability effectively. The finite approximation of Koopman operators captures the intrinsic characteristics of the nonlinear AV system via linear evolution in lifted observable space. Second, a Koopman-based Tube Robust MPC (K-TRMPC) is developed based on the hybrid model and zonotopic set theory. Koopman modeling error raised by the finite operators is considered a disturbance of the perturbed system. Tube-based design for constraint-tightening is developed for the nominal and lifted systems to guarantee closed-loop robustness. A reachability analysis on the future evolution of the perturbed system proves its convergence. Finally, the proposed framework is validated on real-time experiments and simulations, confirming the improved tracking performance on various surface conditions and vehicle stability in combined-slip scenarios.

Advances in single-cell sequencing and data analysis have made it possible to infer biological trajectories spanning heterogeneous cell populations based on transcriptome variation. These trajectories yield a wealth of novel insights into dynamic processes such as development and differentiation. However, trajectory analysis relies on an assumption of trajectory continuity, and experimental limitations preclude some real-world scenarios from meeting this condition. The current lack of assessment metrics makes it difficult to ascertain if/when a given trajectory deviates from continuity, and what impact such a divergence would have on inference accuracy is unclear. By analyzing simulated breaks introduced into in silico and real single-cell data, we found that discontinuity caused precipitous drops in the accuracy of trajectory inference. We then generate a simple scoring algorithm for assessing trajectory continuity, and found that continuity assessments in real-world cases of intestinal stem cell development and CD8 + T cells differentiation efficiently identifies trajectories consistent with empirical knowledge. This assessment approach can also be used in cases where a priori knowledge is lacking to screen a pool of inferred lineages for their adherence to presumed continuity, and serve as a means for weighing higher likelihood trajectories for validation via empirical studies, as exemplified by our case studies in psoriatic arthritis and acute kidney injury. This tool is freely available through github at qingshanni/scEGRET.

Mobile edge computing (MEC) has become a research trend that solves effectively computationally intensive and latency-sensitive tasks. MEC environments in the real world are dynamic and uncertain and then the changes of the environments bring challenges to the generalization and robustness of offloading algorithms. In order to solve the above problem, we propose a meta-reinforcement learning task offloading algorithm GASTO based on Graph Neural Network and seq2seq network. Meta-learning can learn the optimal initialization parameter through several gradient descent steps and samples to adapt to new environments more quickly. The task generated in the user equipment is composed of multiple subtasks rather than a single task, and there are dependencies between the subtasks. Therefore, the task on the user equipment is modeled as a Directed Acyclic Graph (DAG). The connection relationship between the subtasks in DAG plays an important role. Drawing on the idea of message passing, Graph Neural Network is applied in DAG to extract the intrinsic correlation between subtasks in GASTO. In addition, Seq2Seq network can reduce the dimension of action space effectively, and the scheduling decisions of all subtasks can be generated simultaneously. Besides, in order to enhance the sampling efficiency of tasks and the robustness of GASTO, the priority of sampling tasks is adjusted dynamically during the training process. The experimental results of four algorithms in different environments show that the proposed algorithm GASTO can quickly adapt to the new environment.

In the research of autonomous vehicles, most existing studies treat the decision/planning and control as two separate problems. This idea originates from robotics. But since there are essential differences between robot and autonomous vehicle, the structure in Robotics may not be suitable for autonomous vehicles. Considering decision/planning and control separately may affect the performance of autonomous vehicle under complex driving conditions. To fill in the research gap, this paper proposes a novel scheme which considers the local motion planning and control in a combined manner. Firstly, the local motion planning is transformed into the longitudinal control problem based on the proposed scenario adaptive MPC, by which the motion behavior (driving along the global path, car-following, lane-change) can be automatically decided. Then, the lateral MPC controller is designed to track the global path and conduct the local motion commands. To ensure the performance of the path tracking control and a smooth lane-change process simultaneously, an adaptive weight mechanism is introduced in the lateral controller. Comprehensive case studies including both straight and curve road are conducted based on Carsim-Simulink co-simulation platform. The results show that the proposed algorithm can not only ensure the vehicle safety in complex driving conditions, but also ensure that the vehicle can drive at its desired velocity as much as possible by intelligently judging the most proper motion behaviors.

In the background of autonomous driving at level 3 to level 4, an automated vehicle should own smarter driving brain to face complicated transportation situations. In order to construct a safe automated driving brain under highway conditions, this paper focused on driving motion decision in order to generate the control target parameters in the time domain. The coordinate transformation was proposed to convert the complicated curving road to local straight coordinate or inverse, then a receding horizon programming based on mixed logic dynamic constraints was established to formulate a safe-guaranteed optimization model, where the objectives were assigned by driver's steering wheel and speed control, as well as the lateral lane tracking performance. Based on the motion optimization model, the links to the driver characteristics were analyzed, and the weight for each objective in optimization model was tuned by driver statistical features, in which the entropy weights, variance weights, and unique weights are compared. The simulation based on the simulating driving scenario was developed and the optimization results validated the safety and feasibility of motion decision and with the help of k-nearest neighbors (KNN) classifier, the clustering prediction results qualitatively revealed the proposed weights tuning methods for objectives in optimization model could better determine a human-like driving decision, furthermore, this paper gave a basis to compromise multi-objectives in driving decision.

The high-spatial-resolution aerosol retrieval algorithm using Chinese High-Resolution Earth Observation Satellite I (GF-1) wide-field images is developed, which retrieves the aerosol optical depth (AOD) over China for studying the impact of aerosol on climatic and environmental change. The algorithm is based on the red/blue surface reflectance correlations and the lookup table method. To reduce the enormous relative error caused by the constant surface reflectance relationship in the retrieval algorithm, the correlation is parameterized as a function of low, medium, and high values of normalized difference vegetation index (NDVI). Three linear relationships are simulated using MODIS BRDF-adjusted reflectance products (MCD43A4), and MODIS NDVI products are used to ascertain the value of NDVI. By applying the present algorithm to GF-1 images, two different aerosol cases of clear and turbid are analyzed to test the algorithm. Compared with the 10-km MODIS aerosol properties productions, the GF-1 retrieved AOD by our algorithm revealed a significant correlation coefficient with MODIS Dark Target AOD (R = 0.912) and Deep Blue AOD (R = 0.895). Otherwise, the retrieved AOD results are found to be highly correlated with Aerosol Robotic Network (AERONET) sunphotometer observations (R = 0.931). Compared with the results relying on the MODIS surface reflectance model, preliminary validation is encouraging that the method based on our updated surface reflectance assumptions successfully improved the accuracy, particularly under the clear sky background and over bright surface.

P2P SIP is proposed to leverage Peer-to-Peer computing to control multimedia sessions in a decentralized manner. The deployment and maintenance cost of P2PSIP is reduced compared to conventional SIP. In this paper, we propose a hierarchical P2PSIP system to address the connectivity and overhead problems which haven't been solved in the P2PSIP literature. The hierarchical P2PSIP system is implemented under Linux, which demonstrates the feasibility of the proposed scheme. Finally, exhaustive simulations are performed to evaluate the performance of various P2PSIP schemes. Results indicate that the hierarchical approach not only solves the connectivity problem caused by heterogeneous overlays, but also performs more efficiently than the flat scheme when the percentage of nodes in the upper level overlay is less than 10%.

This paper presents a context-aware infrastructure for the easy creation and flexible deployment of the context aware application. We introduce the concept of plane in this infrastructure to manage the sensors which are distributed in the communication environment and used to collect the specific context information. A layered structure of the context information is designed based on this infrastructure. Especially, the Demspter-Shafer evidence theory is applied in the design of the reasoning mechanism in this infrastructure for its superiority over Bayesian method in handling the uncertainty problem, and the inferencer is constructed based on the combination rule of this theory. We also choose the rough set theory and genetic algorithm to design the aggregating mechanism and construct the aggregator.

The cluster-based wireless sensor network (WSN) can enhance the whole network lifetime. In each cluster, the cluster head (CH) plays an important role in aggregating and forwarding data sensed by other common nodes. A major challenge in the WSN is the appropriate cluster head selection approach. In this paper, we propose a centralized cluster head selection approach using analytical hierarchy process (AHP). Three factors contributing to the network lifetime are considered and they are energy, mobility and the distance to the involved cluster centroid respectively. Simulation results demonstrate that the proposed approach is effective in prolonging the network lifetime

Our objective is to build a distributed terminal system to provide smart, context-aware, rich-experienced applications upon personal environment networking technologies, such as WLAN, IEEE 802.15.3 series, ZigBee, Bluetooth, etc. For the proliferation of smart devices with autonomous applications, the users can get much more service experiences than before. While smart devices facilitate human operation, the coordination of devices through networks may provide applications proactively by gathering much more service context, which indicates the emergence of the pervasive computing age. Hence, we proposed the distributed terminal system for the cooperating of smart devices. In this paper, we analyzed the organization architecture of universal service terminal (UST), the distributed terminal system proposed by us before. In UST project, we have abstracted and encapsulated the capabilities of devices as servers for remote invocation by applications, moreover, the framework functionalities have been introduced for the organization of the distributed system. Though the architecture of UST has been validated feasible in a demonstration, the centralized control mechanisms in the heterogeneous environment are inefficient and unreliable. Thus, we propose an evolved scheme by introducing decentralized mechanisms in this paper. The devices around the user are organized in an overlay peer-to-peer network, and some powerful nodes of them provide the decentralized mechanisms for resource management, service discovery, etc