13 results about "Priority function" patented technology

The invention discloses a value maximization calculation unloading method and system based on dynamic priority scheduling and reinforcement learning. According to the method, an optimization framework with maximization of the'long-term average task value 'of the system as a core target is constructed, and the task value is defined by introducing an exponential value function, so that the problem of'value neglect' existing in a traditional'average delay or energy consumption optimization 'strategy is solved. According to the value-oriented optimization mechanism, the multi-agent system can autonomously learn a strategy which preferentially ensures high-value task execution when resources are limited in the training process. In addition, the invention further provides a dynamic priority function which can accurately measure the actual emergency level of the task at different moments, so that the defect that a static priority model is difficult to deal with the real-time change of the emergency state is overcome. Particularly, the dynamic priority method is not only used for task selection in a waiting queue, but also innovatively introduced into a resource dynamic allocation mechanism of a running queue.

The invention provides an electric power operation and maintenance big data monitoring method, and relates to the technical field of data monitoring management, and the method comprises the steps: carrying out the joint expression of electric power data through a multi-dimensional data fusion scheme; embedding the physical topological relation and the power flow information of the power system into the graph neural network structure, and identifying spatial correlation and load imbalance characteristics between devices; embedding a neighborhood load change guide item and a periodic stability constraint based on a gated loop network, and carrying out precise modeling on an equipment state change rule on different time scales; the prediction probability and the state offset score are fused, a hierarchical early warning mechanism oriented to regional risks is established, a scheduling priority function is constructed in combination with the importance weight and the operation load of the equipment, and intelligent scheduling for different equipment is realized; through dynamic load fluctuation analysis and scheduling optimization of the computing nodes, high-risk equipment preferentially obtains computing resource support, and it is ensured that the system has efficient response and computing capacity in a high-concurrency and high-load power grid environment.

The invention discloses a resource dynamic scheduling method and system for a computing network cluster in a power grid information system, and the method comprises the steps: enabling a data collection module to collect the historical resource use data of each task in the cluster in real time after a user submits a computing network calculation task through a task submission interface; the resource demand prediction module extracts spatial and temporal features by means of a one-dimensional convolution and stacking LSTM structure, and accurately predicts task resource demands in a plurality of time slots in the future; the multi-level queue scheduling decision module constructs a comprehensive priority function according to task waiting time, predicted resource requirements and user-defined priorities, realizes task layering and dynamic sorting, and generates a resource allocation scheme with the lowest total scheduling cost based on a rolling horizon optimization method; and the task migration module converts the scheduling decision into actual operation, and completes containerization deployment and resource dynamic adjustment through a Kubernetes interface, thereby realizing real-time optimization and efficient utilization of computing network cluster resources. According to the method, the computing network cluster resource scheduling efficiency can be improved to the greatest extent.

The invention relates to the technical field of software security, and provides a source code high-order vulnerability detection method and system based on execution path analysis, and the method comprises the steps: obtaining a to-be-analyzed source code, and constructing a function execution graph for each function in the source code; on the function execution graph, starting from a function entry node, generating a function execution path according to an execution sequence edge, and applying bounded constraint to a function execution path generation process; calculating a priority score for any function execution path by adopting a path priority function, and sorting and screening the function execution paths based on the priority scores to obtain candidate execution paths; and obtaining a vulnerability type and vulnerability location through a path coding model, a statement encoder, an attention mechanism and a vulnerability detection model. And the vulnerability detection accuracy and the positioning interpretability are improved under the controllable calculation overhead.

The application relates to a V2G-VPP resource aggregation method fusing isomerism and dynamic priority, which comprises the following steps: through protocol non-inductive access, communication data of various types of devices involved in V2G-VPP are processed, automatic analysis and conversion of cross-protocol device instructions are realized; based on the analyzed and converted data, line chain characteristics and social attributes are extracted, a user elastic quantization model is constructed, and an elastic coefficient is calculated; a dynamic coupling priority function driven by power grid demand and resource state is constructed by combining power grid frequency modulation demand, predicted moving probability and the elastic coefficient, resource priority is dynamically evaluated, each electric vehicle is sorted and screened in descending order based on the resource priority, a resource priority scheduling sequence is generated and optimized. Compared with the prior art, the application has the advantages of being capable of improving protocol conversion efficiency, improving high-value resource identification accuracy, greatly reducing V2G schedulable capacity evaluation error and the like.

The present application belongs to the technical field of computer, the present application discloses a heterogeneous SoC task adaptive allocation system;Including: migration determination module, identifies the load imbalance state between each core in the running process, and filters migration candidate task;Migration preparation module, the running state of migration candidate task generates baseline snapshot, and carries out hierarchical division;Semantic annotation module, the memory page occupied by migration candidate task is carried out semantic annotation, and the regularized semantic score is calculated, and the memory page data is allocated level according to regularized semantic score;Compression migration module, the layered state data is carried out mixed compression and adaptive restoration in the migration process;Preheat loading module, on the target core after migration, according to the preheat priority sequence, the preheat loading of the code and data of high priority function is executed to the cache of target core;Through semantic annotation and hierarchical compression, the transmission data volume is effectively reduced, and the migration pause time is reduced.

To provide a method and a system for constructing a customized knowledge graph.SOLUTION: The method generates a seed graph by extracting entities and relationships between the entities from the user dataset based on a specific rule. The large knowledge graph is then traversed using a finite state machine to identify candidate entities and / or relationships to add to the seed graph. A priority function is used to select entities and / or relations from the candidate entities and / or relations. The selected entities and / or relationships are then added to the seed graph to generate the customized knowledge graph.SELECTED DRAWING: Figure 5

The invention relates to a multi-sensor adaptive task scheduling method for hoisting operation monitoring, and belongs to the technical field of multi-sensor task scheduling. The method comprises the steps of establishing a task scheduling basic model; calculating scheduling time delay and communication bandwidth occupation parameters; calculating a system data communication rate supporting scheduling of all the sensors at the minimum frequency under the condition that the communication bandwidth is insufficient; establishing a multi-sensor task optimization function; and the gateway automatically generates an optimal scheduling scheme of each sensor task. According to the method, the task priority function is constructed, so that the task processing sequence of each sensor can be adaptively decided, and the method has relatively good environmental adaptability to the real-time change of a hoisting operation scene state; the relation between the task execution efficiency of the safety monitoring multi-sensor in the hoisting operation scene and the limited bandwidth resource of the system network is effectively balanced, and the optimal execution of the safety monitoring task under the limited resource condition is ensured.

The invention relates to the technical field of video conference operation and maintenance management, in particular to an automatic operation and maintenance management method and system for a video conference, and the method comprises the steps: obtaining current conference task data, constructing a state matrix based on the conference task data, calculating a state score vector and abnormal strength of each terminal, and comparing the state score vector and abnormal strength with a preset threshold value; generating a terminal exception indication vector according to a comparison result, generating a to-be-executed set according to the terminal exception indication vector, dividing the to-be-executed set into corresponding action queues according to recommended action types, generating an execution sequence of each terminal based on a priority function, and executing the recommended action types based on the execution sequence, generating an execution result vector and an execution record set; the terminal sets which are not successfully confirmed after the automatic operation and maintenance are finished are reserved, priority weights are constructed for the terminal sets which are not successfully confirmed, the terminal sets are sorted, and prompt information is generated for the terminal sets in combination with the sorting result and the execution record set.

The application discloses a delay-tolerant network cache management method, system, device, medium and program product based on connected block sociality, and belongs to the technical field of wireless communication and computer network. The management method comprises the following steps: each node in the delay-tolerant network constructs a connected block society with a limited hop number centering on the node itself, extracts the social characteristics of the node and the business message in the connected block through a consensus message broadcast and multi-hop propagation mechanism, and performs real-time synchronous updating; based on the social characteristics, a multi-dimensional cache priority function is constructed by fusing the properties of the business message, the cache value of each business message is quantified comprehensively, and low-value business messages are discarded when the cache is full; the system, the device and the medium are used for realizing the mechanism; the program product comprises a computer program for realizing the mechanism; and accurate business message screening when the cache is full is realized, so that the arrival rate of the business message is improved.

The application relates to the technical field of software security, and provides a source code high-order vulnerability detection method and system based on execution path analysis, which comprises the following steps: obtaining source code to be analyzed, and constructing a function execution graph for each function in the source code; starting from a function entry node on the function execution graph, generating a function execution path according to an execution order edge, and imposing a bounded constraint on the function execution path generation process; calculating a priority score for any function execution path by using a path priority function, sorting and screening the function execution paths based on the priority score, and obtaining a candidate execution path; and obtaining a vulnerability type and vulnerability positioning by using a path encoding model, a statement encoder, an attention mechanism and a vulnerability detection model. The vulnerability detection accuracy and positioning interpretability are improved under controllable computing overhead.

This invention discloses a value maximization computational offloading method and system based on dynamic priority scheduling and reinforcement learning. The invention constructs an optimization framework with the core objective of maximizing the system's "long-term average task value." By introducing an exponential value function to define task value, it addresses the "value neglect" problem inherent in traditional "average latency or energy consumption optimization" strategies. This value-oriented optimization mechanism enables multi-agent systems to autonomously learn strategies to prioritize high-value tasks during training when resources are limited. Furthermore, this invention proposes a dynamic priority function that accurately measures the actual urgency level of tasks at different times, thus overcoming the shortcomings of "static priority" models in handling real-time changes in urgency. Notably, this dynamic priority method is not only used for task selection in the waiting queue but is also innovatively introduced into the dynamic resource allocation mechanism of the running queue.

The application discloses an RTC transmission optimization technology based on Priority NACK, which comprises the following steps: firstly, a data structure PrioritySeqBuffer with a packet sequence number as a storage object is designed at a receiving end; secondly, a NACK request protocol with a priority function is designed at the receiving end; and finally, a protocol processing module PriorityNackModule is designed. According to the application, all lost data packets have the same priority by using a general NACK request protocol realized internally in WebRTC, and the rule is to perform packet loss retransmission once upon receiving a NACK request. Compared with the general NACK request protocol realized internally in WebRTC, the Priority NACK request can greatly increase the packet integrity and continuity in the scenario that only one lost data packet needs to be supplemented to make a large number of subsequent data packets continuous, thereby improving the decoding and rendering frame rate of a video frame and reducing lag.