Aiming at the Multi-Crane Scheduling Problem (MCSP) in the transit warehouse, a scheduling scheme with dynamic priority is proposed in this paper. The problem was modelled with the goal of respectively minimizing the completion time of all tasks and the frequency of crane avoidance. A Simulation-based Genetic Algorithm (SbGA) is designed to solve the model and generate feasible scheduling schemes. Besides, a novel updating strategy for task priority is designed to implement dynamic scheduling. The proposed scheme is applied to an actual industrial case in an iron and steel enterprise. Numerous experiments demonstrate the efficiency of the proposed model and scheme.

In this paper, a novel combined controller is proposed for an industrial heavy-duty 3D overhead crane system with load hoisting or lowering. It is of great significance to develop the controller of industrial heavy-duty bridge crane to improve production efficiency and reduce safety accidents. Many existing research works have not fully considered many practical factors in the actual industrial overhead crane system, such as actuators, speed limitation, motor current and work efficiency, which has caused difficulties in practical application. The anti-sway module of the proposed controller combines the commonly used the ZVD shaper, the first-order inertial filter and the saturation element to effectively suppress the residual swing of the time-varying rope length, while alleviating the motor current surge and preventing the speed from exceeding the upper limit. The positioning module of the proposed controller optimizes the traditional PID algorithm based on the Multi-dimensional Taylor Network (MTN) to increase efficiency, which saves about 30% of the settling time when the maximum allowable error is 5cm. The overhead crane, vector drive and AC induction motor are modeled on the SIMULINK platform, the proposed algorithm is given in discrete form, and some simulations are performed to verify the effectiveness of the proposed combined controller.

Precipitation nowcasting is quite important and fundamental. It underlies various public services ranging from rainstorm warnings to flight safety. In order to further improve the prediction accuracy for the spatiotemporal sequence forecasting problem, we propose an encoder–decoder deep residual attention prediction network, which adaptively rescales the multiscale sequence- and spatial-wise features and achieves very deep trainable residual prediction by integrating global residual learning and local deep residual sequence and spatial attention blocks (RSSABs). Experiments in a real-world radar echo map dataset of South China show that compared with the ingenious PredRNN++, TrajGRU methods, and newly proposed Unet-based methods, our ED-DRAP network performs better on the precipitation nowcasting metrics, as well as occupies small GPU memory.

In this paper, we propose an event-driven output feedback control strategy for a class of nonlinear systems subject to disturbances. Different from the time-driven control, the event-driven control can be regarded as a more reactive approach where the control actions are taken only when an event is triggered, thus the event-driven control has a better balance between the control performance and other system aspects (such as processor load, communication load, and system cost price). Based on the extended state observer (ESO), we propose a composite event-driven controller, which is asynchronously updated only when an intolerable effect on the closed-loop performance is produced. It is proved that the closed-loop system is globally uniformly bounded, and has a good robustness against disturbances. Meanwhile, the closed-loop system considered in this paper is a hybrid system, thus we need to consider the problem of Zeno behavior, which is a phenomenon unique to hybrid systems, and describes the situation where a hybrid system undergoes an unbounded number of discrete transitions in a finite and bounded length of time. Fortunately, it is proved that the system under the event-driven controller can avoids the Zeno behavior of the sampling, and has the significantly reduced sampling frequency compared with the time-driven controller. Finally, a simulation of DC-DC buck converter is conducted to demonstrate the efficiency of the new scheme.