14results about How to "Implement dynamic scheduling" patented technology

A method, medium, and electronic device for model training are disclosed. The method includes: acquiring multiple samples and using these samples as candidate samples to obtain multiple candidate samples; determining the current training value of each candidate sample based on its influence on the current dynamic parameters of the model to be trained; selecting a portion of the samples from the multiple candidate samples as current training samples based on their current training values; providing the current training samples as input to the model to be trained; calculating loss based on the output of the model to be trained; and updating the model parameters of the model to be trained based on the result of the loss calculation. The technical solution provided by this disclosure helps to concentrate computational resources during the training process on candidate samples with higher current training value, thereby improving the training efficiency of the model and enhancing its performance in terms of sparse sample recognition and generalization capabilities.

The application discloses a scheduling method and device for encrypted card virtual resources, equipment and medium, and relates to the technical field of cloud computing. The method is applied to an encrypted card system based on SR-IOV technology, and comprises the following steps: grouping a plurality of VFs divided from a physical encrypted card to construct at least one VF group; mapping each VF group to a corresponding virtual machine and providing each virtual machine with a unified encrypted device interface; monitoring the running state indicators of the VF groups, and adjusting the number of VFs in the VF groups according to the running state indicators through a reinforcement learning algorithm. Through the technical scheme, the intelligent scheduling and elastic scaling of the resources of the encrypted card can be realized without deleting or rebuilding the virtual machine or modifying the upper-layer application program when the user business demand changes, so that the system flexibility and resource utilization rate are improved.

This invention provides a multi-agent-based software pilot testing method and apparatus. The method includes: matching verification agents corresponding to the software type of the software to be verified; constructing a verification agent cluster for the software to be verified based on the verification agents; verifying the software to be verified based on the verification agent cluster and pilot testing requests to obtain agent verification results; and determining the pilot testing result of the software to be verified based on the agent verification results. The method provided by this invention automatically acquires and parses verification requests, accurately matches and constructs a verification agent cluster, efficiently executes verification tasks and generates results, and ultimately objectively determines the pilot testing conclusion. This constitutes a complete, end-to-end intelligent pilot testing solution, significantly improving the efficiency and accuracy of software pilot testing, achieving precise matching and dynamic scheduling of verification resources, and enhancing the traceability and objectivity of verification results.

The invention relates to a water transmission and distribution dynamic scheduling optimization method, which comprises the following steps that a water transmission and distribution test area is established, the water transmission and distribution test area comprises a test branch pipe, a gas pumping unit and a test user side, and two ends of the test branch pipe are respectively connected with a target water plant and the test user side; performing water hammer disturbance effect analysis on the test branch pipe to obtain a water hammer focus of the test branch pipe; air is pumped into the test branch pipe based on the air pumping unit, and vibration parameters and cutoff parameters of the water hammer focus are obtained through strain vibration monitoring and underwater sound cutoff monitoring; establishing an interference factor modulation unit, and performing interference parameter correction on the vibration parameter and the cutoff parameter based on the interference factor modulation unit to obtain a training parameter; circularly executing the previous step to obtain multiple groups of training data, and constructing and training based on the obtained training data to obtain a filtering matching unit; establishing a water transmission and distribution optimization network in the target pipe network area; and managing and optimizing water supply scheduling in the target pipe network domain based on the water transmission and distribution optimization network.

The invention provides a task scheduling method and device, electronic equipment, a storage medium and a program product, and relates to the technical field of parallel computing and task scheduling, the method comprises the following steps: in response to a received to-be-scheduled computing task, determining a demand credit value and a resource demand quantity corresponding to the computing task, determining the residual quantity of resources in each MPU (Microprocessor Unit) and the residual credit value of each MPU; determining a target MPU from each MPU based on the demand credit value, the demand quantity of the resources, the residual quantity of the resources in each MPU and the residual credit value of each MPU; and scheduling the calculation task to the target MPU. According to the method, the task dynamic scheduling based on resource awareness can be realized, so that the task load balance among MPUs is ensured while the task execution efficiency is improved, and the resource utilization rate can be improved.

The invention provides a multi-user resource management method and system in a wireless communication system. According to the method, the dynamic energy consumption characteristic is generated based on the equipment state information and the real-time power consumption information of the Internet of Things equipment, the resource demand combination is determined by combining the dynamic energy consumption characteristic and the service data characteristic of the service data to be transmitted, and the uplink channel condition is obtained by performing channel detection on the uplink of the Internet of Things equipment. And generating a resource allocation strategy according to the resource demand combination and the uplink channel condition, and dynamically allocating a transmission resource unit to the Internet of Things equipment based on the resource allocation strategy. According to the technical scheme provided by the invention, the accuracy, the energy efficiency and the overall adaptability of system resource allocation are improved, and the coordination and the high efficiency of multi-user service transmission are also ensured.

The application provides a digital twin power supply dynamic allocation method and system based on an unmanned aerial vehicle, relates to the technical field of power supply and intelligent scheduling, and comprises the following steps: collecting the operation parameters of a battery pack of the unmanned aerial vehicle, and preprocessing the operation parameters to obtain power supply operation state information; a digital twin model of the power supply system of the unmanned aerial vehicle is established based on the power supply operation state information; and virtual energy consumption data is generated in real time by synchronizing the digital twin model with the operation parameters. The digital twin power supply dynamic allocation method and system of the unmanned aerial vehicle collects the operation parameters of the battery pack of the unmanned aerial vehicle in real time, establishes a digital twin model, and dynamically optimizes and manages the power supply system. Through adaptive Kalman filtering and a multi-objective optimization algorithm, the power demand and the residual energy supply capacity can be accurately predicted, so that the dynamic scheduling and optimized allocation of the power supply of the unmanned aerial vehicle are realized, and the accuracy and reliability of power supply management are effectively improved.

The application discloses a resource coordination control method and device, electronic equipment and a storage medium. The method comprises the following steps: in the case that the power consumption of a central processor is greater than a safety power consumption threshold, a first resource usage rate is calculated by an internal integrated algorithm; if the first resource usage rate is greater than or equal to a usage rate threshold, a second resource usage rate is obtained from an operating system through an in-band channel; the first resource usage rate and the second resource usage rate are verified to determine the resource occupation state of the central processor; in the case that the resource occupation state is a first occupation state, an abnormal proportion component of the central processor is identified and processed, and the first resource usage rate calculated by the internal integrated algorithm and the second group of resource usage rate data obtained from the operating system through the in-band channel are cross-verified to eliminate the misjudgment risk of single-source data, so that the humanoid robot can maintain an efficient and stable operating state in different use environments.

This invention relates to the field of autonomous driving technology and discloses an autonomous driving simulation testing system and equipment. The system allows end-users to configure an interface that sends simulation tasks with target scenarios to a central scheduler and visualizes the process. The central scheduler receives the simulation tasks and, via a message middleware, sends them to distributed simulation computing nodes, building a global task scheduling engine to automate the entire testing process. The message middleware uses an asynchronous communication mechanism to provide message transmission channels. The terminal scene database writes the entire simulation task process data and encrypts scene resources from a cloud data management platform. After confirming receipt of the simulation task via a vehicle-to-everything (V2X) interaction module, the distributed simulation computing nodes retrieve the corresponding target scenario data from the terminal scene database, call the simulation engine to load the target scenario, and execute it. This invention solves the technical problems of insufficient flexibility and scalability, chaotic scene management, low automation, and weak security in existing systems.

This disclosure discloses embodiments of a transport route adjustment method, apparatus, system, electronic device, and readable medium. One specific implementation of the method includes: in response to receiving a cargo replenishment instruction, determining whether there are goods conforming to the cargo replenishment instruction among goods transported using a second transport mode; in response to determining that there are goods conforming to the instruction, analyzing a first transport attribute value generated by the goods being transported using a first transport mode; and determining whether to adjust the transport route of the goods based on a comparison result between the first transport attribute value and a second transport attribute value. This implementation relates to multimodal transport technology, determining whether to adjust the transport route of goods conforming to the instruction by comparing and analyzing the first transport attribute value generated by the first transport mode with the second transport attribute value generated by the second transport mode. This allows for dynamic allocation of resources based on actual load factors, thereby reducing resource waste.

The invention relates to the technical field of elevator maintenance, in particular to an elevator maintenance method and system, and double improvement of elevator maintenance efficiency and safety is achieved through multi-dimensional data fusion, intelligent algorithm optimization and full-process closed-loop management. The level of the method comprises five core steps of real-time data acquisition, dynamic anomaly detection, fault prediction and early warning, intelligent task allocation and quality tracing. The system layer is integrated with five functional units including a data acquisition module, an anomaly detection module, a fault early warning module, a task distribution module and a quality tracing module. The specific embodiment comprises five embodiments of a cloud intelligent platform, an edge computing node, fault early warning model training, task dynamic scheduling and block chain evidence storage.

The invention relates to the technical field of data communication, and discloses a router hardware acceleration processing method and system based on a heterogeneous architecture, and the method comprises the steps: generating a hardware acceleration instruction containing a physical memory address pointer of a to-be-processed data packet; writing the hardware acceleration instruction into a hardware task queue mapped to a special accelerator address space, and triggering a doorbell interrupt signal after writing is completed; in response to the doorbell interrupt signal, reading a hardware acceleration instruction by the special accelerator according to an instruction sequence in the hardware task queue to obtain a physical memory address pointer, and directly capturing a to-be-processed data packet from a host memory through the physical memory address pointer for acceleration processing to generate a processing result; and the general processor obtains a processing result, updates a local routing table item according to a routing update field in the processing result, and executes a final forwarding action of the to-be-processed data packet. According to the invention, the data processing efficiency of the network in an ultra-low time delay and ultra-high throughput scene can be improved.

This invention relates to a port data sharing method based on intelligent scheduling, applicable to data demander systems, data provider systems, and data sharing platforms. The demander submits a sharing ticket, which the platform sends to the provider. The provider returns a commitment result based on pruning rules: when the sharing policy corresponding to the purpose identifier is "only allow judgment" or "prohibit judgment," only the judgment result field is delivered; when the time limit requirement exceeds the committable range, the committable time limit and delivery window are returned; when the range identifier meets the splitting conditions, sub-tickets are obtained and delivery fields and delivery windows are determined separately; the provider forms a delivery package; the demander initiates a delivery request within the validity period calculated from the time the delivery package is published; the platform verifies the request and delivers it; if the time limit is exceeded or the scope is crossed, the request is rejected and a new ticket is required; the demander sends back the signature information, and the platform updates the subsequent ticket scheduling parameters accordingly, lowering the priority of tickets without signatures and restricting delivery methods.

The invention discloses a charging system and a scheduling method thereof, and relates to the technical field of charging. The charging system comprises at least two power modules, at least one charging interface, a main controller and a power distribution device, and the method comprises the following steps: obtaining state data of the charging system, the state data being used for representing current states of each power module, each charging interface and the power distribution device; an environment vector is generated based on the state data, a topology constraint matrix and an effective charging strategy, and the topology constraint matrix is obtained based on the switch topology of the power distribution device; and inputting the environment vector into a pre-trained scheduling model to obtain an action vector which is used for indicating a main controller to schedule each power module. Therefore, dynamic and intelligent power distribution is realized by sensing the full state of the charging system, combining hardware topology constraints and fusing a multi-dimensional optimization target, the efficiency and the adaptability of the charging system are improved, and the service life of equipment is prolonged.