17 results about "Resource description" patented technology

The application discloses a kind of general task scheduling framework systems and methods supporting dynamic arrangement, belong to computer software system architecture technical field, including framework engine core, field adaptation layer and functional plug-in;The framework engine core is the general scheduling execution engine irrelevant to specific business field, including resource abstract module, task abstract module, dynamic arrangement and execution engine and plug-in warehouse and manager;The resource abstract module defines uniform resource description interface, and converts original resource data into resource image.The application can dynamically retrieve, load plug-in and assemble into exclusive plug-in workflow according to the unique needs of individual task from plug-in warehouse during scheduling operation by dynamic arrangement mechanism, realize the "thousand people thousand faces" and "one task one strategy" of scheduling strategy, so that scheduling strategy can dynamically adapt to task demand, completely solve the defect that static strategy cannot match individualized needs of task.

The invention relates to the technical field of unmanned collaboration and resource virtualization, and provides a task collaboration-oriented unmanned system virtual resource representation and organization method and device and a storage medium. The method comprises the following steps: firstly, determining physical resource virtualization granularity (platform level, equipment level and function unit level) according to unmanned collaboration styles (task level, resource level and element level); secondly, establishing a resource tree model comprising a resource base class, an attribute resource, a data table resource and a resource sub-tree based on granularity; resource storage is carried out by utilizing an organization structure constructed by a hash table, a list and a set; and finally, forming a resource pool through standardized packaging, providing management interfaces (resource registration, cancellation and state maintenance) and scheduling interfaces (demand response, resource allocation and release), and realizing resource addressing by adopting URI coding. According to the method, the problems of low task execution efficiency, insufficient resource utilization rate and complex management caused by poor coordination task and resource adaptability, non-uniform heterogeneous resource description and non-standard packaging are solved.

This invention discloses a method and system for pooling description and scheduling adaptation of heterogeneous hardware resources on the device side for embodied tasks. The method includes: identifying heterogeneous hardware resources on the device side; performing pooling modeling on the resources and collecting capability parameters, real-time status, and interface protocol types; generating standardized resource description files and uploading them to the cloud; and matching and adapting resources according to the requirements of the embodied task and issuing scheduling instructions from the cloud. This invention achieves a unified expression of device-side resource capabilities, improving resource utilization and scheduling efficiency.

The application discloses a data transmission system resource scheduling method based on a declarative API and a storage medium, belongs to the technical field of resource scheduling under a Kubernetes environment, and aims to solve the technical problem of how to utilize Kubernetes resource arrangement capability to realize convenient and quick scheduling of data transmission system resources.The technical scheme is as follows: the method is as follows: custom API resource type CRD: defining the resource type CRD of the collection end resource, the warehousing end resource and the message middleware resource of the data transmission system respectively; wherein, the group of the resource type CRD is dts; the resource description file corresponding to the resource type CRD is used to schedule the collection end resource, the warehousing end resource and the message middleware resource, and the corresponding resource scheduling work is realized through the resource file of the application captured by the custom controller; the definition of the resource file is used to complete the abstraction of the underlying physical resource and the business logic resource, and the scheduling of the business resource is realized through the custom controller.

This application relates to the field of unmanned aerial vehicle (UAV) technology and discloses a method for generating a modeling language for UAV swarm application scenarios. The method includes: constructing a task model to determine the three-level hierarchical relationship between the UAV swarm's scenario tasks, tasks, and sub-tasks, the collaborative relationships between tasks, and the collaborative relationships between sub-tasks; constructing a resource model to analyze the resources required by the task model and provide a unified resource description; constructing a constraint model to analyze the constraints of each UAV regarding resources, motion, environment, and time, and determining the constraints of the UAV swarm based on the constraints corresponding to each UAV; determining the BNF paradigm of the task model, resource model, and constraint model, and constructing syntax rules; constructing XML conforming to the BNF paradigm based on the syntax rules, and combining the results to generate a modeling language for UAV swarm application scenarios applicable to different application scenarios, providing a good underlying foundation for multi-UAV collaborative task planning.

Embodiments of the present application disclose a method and device for fault diagnosis of a charging gun. In the embodiments of the present application, real-time charging order data is acquired; at least two groups of resource description information features and real-time charging data features are determined according to the real-time charging order data; the at least two groups of resource description information features are respectively input into respective corresponding resource description information decision tree models in a resource description information decision tree model group, to generate at least two types of resource description information data, wherein the resource description information data is numerical data, and the resource description information decision tree model group includes at least two resource description information decision tree models; the at least two types of resource description information data and the real-time charging data features are input into a pre-trained fault diagnosis model, to output a fault type. Through the above method, the fault type of the charging pile can be quickly and accurately determined without the need for on-site detection by an operation and maintenance personnel.

This application discloses a container creation method, apparatus, and electronic device. The method includes: receiving a container creation request, the creation request carrying a state set, the state set including first container resource description information of a first container to be created and second container resource description information of a second container; responding to the creation request, scheduling a first topology domain from multiple topology domains according to the first container resource description information, and creating a first container in the first topology domain according to the first container resource description information, scheduling a second topology domain from multiple topology domains according to the second container resource description information, and creating a second container in the second topology domain according to the second container resource description information, wherein the first topology domain and the second topology domain are different; forming a target cluster according to the first topology domain and the second topology domain, assigning the first container to a first role in the target cluster, and assigning the second container to a second role in the target cluster.

The application discloses a global AI heterogeneous computing power dynamic scheduling and unified management method and system, and belongs to the technical field of artificial intelligence computing power scheduling and heterogeneous resource management. The method comprises the following steps: acquiring resource capability information and running state information of heterogeneous computing power nodes in multiple computing power domains, and constructing a unified resource description object; analyzing an AI task to be executed, and generating a task stage portrait; constructing a heterogeneous compatibility matrix based on the unified resource description object and the task stage portrait; performing domain-level scheduling and node-level scheduling based on the heterogeneous compatibility matrix; generating a unified execution state package and an incremental state block and migrating and restoring when a rescheduling condition is met; and performing prediction and suppression control based on a video memory fragment state. The application can improve the scheduling efficiency, recovery efficiency and resource utilization rate in a heterogeneous computing power environment.

This application discloses a heterogeneous computing power resource aggregation and intelligent task matching and distribution system and method, belonging to the field of computing power resource distribution technology. The heterogeneous computing power resource aggregation and intelligent task matching and distribution system includes a computing power cluster registration module, a task publishing module, and a computing power market coordination module. The computing power cluster registration module is used for clusters to declare their heterogeneous hardware and software environment capabilities; the task publishing module allows users to define task requirements; the computing power market coordination module has a built-in intelligent scheduling engine, which achieves optimal matching between tasks and computing power clusters based on multiple dimensions such as cluster capabilities, real-time load, pricing strategy, user budget, task priority, and cluster reputation score, and issues task instructions to the proxy service of the matched cluster. This invention aims to solve the problem of difficult unified management and efficient utilization of heterogeneous computing power resources, and improves resource allocation efficiency and user task execution satisfaction through unified resource description and multi-dimensional intelligent matching.

The application discloses a kind of multi-dimensional resource perception edge flow task processing methods based on distributed hash table, it is applied to the edge network environment of the heterogeneous edge computing node group composition.To each node establishes resource description model, flow processing application is abstracted as directed acyclic graph topology;Three-tier decentralized flow processing architecture including edge physical network layer, structured overlay network layer of fusion multi-dimensional resource perception routing algorithm and dynamic flow data abstraction layer is built;According to topological sorting, each operator calculates independent hash routing key and encapsulates deployment message injection overlay network;In the process of hop-by-hop routing of deployment message, node computing load, available memory, physical communication delay and logical hash distance are included in comprehensive cost evaluation;Through instant interception mechanism, intermediate node actively carries operator instance, realize the efficient adaptation of logical topology and physical resource, reduce flow processing application deployment time and end-to-end query delay.

The application provides a pointer recycling method, system and device, the method comprising: scanning source code or an abstract syntax tree through a compile-time automatic instrumentation module, identifying a resource reference entry and a target variable to be managed, inserting a cleanup attribute and ownership transfer semantics for the target variable, and generating a resource description table; the resource reference entry comprises a naked pointer; the naked pointer is wrapped using a lightweight proxy structure, and metadata is associated through the lightweight proxy structure; a resource in the resource description table is registered, reference count management is performed, cross-function transfer migration is performed, and a unified cleaning operation is performed through a runtime resource manager, to obtain a pointer life cycle state; system memory occupation information is collected in real time at runtime, memory recycling frequency and strategy are dynamically adjusted according to the memory occupation information, and resource recycling is performed according to the pointer life cycle state. The scheme improves the stability of long-term system operation, and improves the stability, reliability and development efficiency of C language software development.

The present disclosure provides a resource processing method and device, an agent and an electronic device, relates to the technical field of artificial intelligence, in particular to the technical field of deep learning, cloud computing, natural language processing, intelligent search, large model and the like. The specific implementation scheme is: obtaining a target question, a word segmentation result of the target question and a resource description index library; the resource description index library includes candidate resource description information of at least one candidate resource; querying the resource description index library according to the word segmentation result to obtain a target description parameter in resource description information of a target resource; generating a resource processing instruction according to the word segmentation result and the target description parameter, and then obtaining a resource processing result; wherein, the resource description index library is queried according to the word segmentation result of the target question to select the target resource, which can avoid the situation that the selection accuracy is poor due to poor understanding ability of the large model on the resource content and the truncation processing, thereby improving the selection accuracy and the resource processing efficiency.

The application discloses a resource recommendation method and device based on model training, and a medium product. The method comprises the following steps: acquiring a resource description text of a sample media resource and a resource semantic identifier of the sample media resource; calling a recommendation model to reconstruct the resource semantic identifier of the sample media resource according to the resource description text of the sample media resource, and obtaining an identifier reconstruction result; calling the recommendation model to reconstruct the resource description text of the sample media resource according to the resource semantic identifier of the sample media resource, and obtaining a text reconstruction result; and based on the identifier reconstruction result and the text reconstruction result, training the recommendation model in resource semantic identifier and resource description text alignment, so that the trained recommendation model is used for media resource recommendation according to an input resource semantic identifier, thereby improving the accuracy of media resource recommendation.

This application relates to a cross-platform data processing method, belonging to the field of data security technology. The method includes: displaying data resource description information on a data resource selection interface at the data user end, describing data resources published by the data provider end to a converged computing platform compatible with at least two privacy computing components; each privacy computing component has a corresponding computing component interface; determining target data resource description information from the data resource description information in response to a data resource selection operation; the target data resource pointed to by the target data resource description information being provided by the target data provider end; triggering authorization for the target data resource in response to an authorization request operation; after the target data resource is authorized, instructing the user end to jointly train a business model based on the user end's sample data and the target data resource, during which the data user end and the target data provider end communicate with the converged computing platform based on their respective matching computing component interfaces. This method can meet the needs of more business scenarios.

An Internet of Things (IoT) monitoring method, a monitoring module generation method, and an electronic apparatus are provided. First, a monitoring module of an IoT device is generated, including: receiving a resource description file of the IoT device; based on a resource message of each monitorable device in the resource description file, obtaining a visualization object and a control code corresponding to each monitorable device and then generating a monitoring interface; generating a control logic based on a connection message in the resource description file; generating a monitoring module based on the control logic and the monitoring interface. Then, in response to the monitoring module running in a client device, the monitoring module may be applied to monitor the IoT device via a cloud server.

The application provides a media file transmission method, a packaging method and an unpackaging method. The method relates to the technical field of file transmission in the multimedia field, and comprises the following steps: acquiring a media resource description file, wherein the media resource description file comprises description files of a plurality of media files; selecting a description file of a target media file from the description files of the plurality of media files; the target media file comprises extension information of a target image in the target media file; sending a resource request based on the description file of the target media file, wherein the resource request is used for requesting the target media file; receiving the target media file; and presenting the extension information or presenting a control used for a user to perform an operation on the extension information in response to a selection operation performed on a target region of the target image. The media file transmission method provided by the application can improve the viewing experience.