35results about How to "Efficient scheduling" patented technology

The invention discloses an MAC unit supporting input activation and weight double-end sparsity, and relates to the technical field of neural network accelerator hardware design. Aiming at the technical defects existing in a sparse matrix acceleration scheme, the scheme is adopted, an activation value and non-zero elements of a weight matrix are stored through a special compressed data register, corresponding zero value distribution is recorded through a matched sparse bitmap register, and a joint sparse bitmap marking a double-non-zero effective calculation position is generated through dual-port joint operation; and the control and address generation subunit scans the joint sparse bitmap and outputs an address and an enable signal, the MAC calculation subunit is driven to read non-zero data to complete multiply-accumulate operation, and finally, an operation result is compressed and coded through an output end sparse encoder, so that efficient transmission of interlayer sparse data streams is realized. According to the method, input activation and weight double-end sparsity can be utilized at the same time, end-to-end sparse data flow is supported, zero value calculation can be dynamically skipped, and the energy efficiency ratio and the calculation throughput of neural network reasoning are improved.

This utility model discloses an intelligent three-dimensional storage rack that is easy to schedule, including a support frame and multi-layer placement plates arranged between the support frames. Side frames are respectively arranged on the front sides of the support frame, and a top plate is arranged between the tops of the side frames. A lifting screw is arranged inside the side frame, and a lifting gear is arranged at the top of the lifting screw. A belt is arranged between the two lifting gears. A mounting plate is arranged on the top of the side frame, and a lifting motor is arranged on the mounting plate. A pair of slide rail grooves are arranged on the moving base. The traverse motor is powered by an external power supply. A moving gear and a fixed gear are arranged inside the moving base. A support plate is arranged above the moving base. This storage rack achieves efficient positioning and scheduling of goods through precise mechanical linkage design. This layered positioning mechanism significantly improves the accuracy and efficiency of goods scheduling, facilitates subsequent goods picking operations, and reduces the risk of manual goods handling.

The invention discloses a marketing content generation method based on multi-agent cooperation and related equipment, and the method comprises the steps: determining a target business scene according to content generation demand information and content generation reference information, and obtaining a corresponding target scene data package from a marketing scene library; determining a target agent list and a task list according to the target scene data packet, generating a task directed acyclic graph in combination with the material information and the constraint information, and further generating an arrangement plan; and calling a target agent plug-in to execute a generation task according to the target agent list and the arrangement plan, obtaining a stage content generation result, and performing acceptance check to obtain a final content generation result. According to the embodiment of the invention, full-process automation from scene recognition, task disassembly, task arrangement, multi-agent task generation execution to content acceptance can be realized, and the marketing content generation efficiency, quality and stability can be effectively improved. The method can be widely applied to the technical field of generative artificial intelligence.

The present application belongs to the technical field of intelligent manufacturing and logistics automation, and specifically relates to a multi-task scheduling optimization method and system for RGV shuttle vehicles, which comprises the following steps: firstly, the system is digitally modeled by constructing a time-space resource network and a task association graph; then, a joint prediction model is trained to predict congestion risk and energy consumption cost; further, the prediction results are fused with task value and power state into dynamic heuristic information to guide the self-adaptive ant colony algorithm; finally, online correction and closed-loop self-learning of the model are realized by analyzing the execution deviation vector. The present application continuously improves the actual execution efficiency, robustness and economy of the scheduling scheme by calculating the deviation between the actual execution result and the prediction, correcting the prediction model online, and forming a closed-loop self-learning mechanism.

The invention relates to the field of oil field energy supply, and discloses an oil field united station energy supply system and method based on carbon dioxide energy storage and waste water hydrogen production. The carbon dioxide energy storage device is configured to selectively store or release electric energy by using carbon dioxide generated by the oil field united station; the production load comprises a hydrogen production device, and the hydrogen production device is configured to produce hydrogen by using the production wastewater of the oil field united station; the power distribution device is electrically connected with the clean energy power generation device, the carbon dioxide energy storage device and the hydrogen production device; and the energy management device monitors the power generation power of the clean energy power generation device and the electricity utilization power of the production load, and controls the operation of the carbon dioxide energy storage device according to the power generation power and the electricity utilization power. The safe, efficient and stable energy supply system can be provided, the discharge amount of carbon dioxide and waste water can be reduced, and the utilization rate of clean energy is increased.

The application discloses a lossless and efficient data processing method and network interface card based on RMDA, and belongs to the field of data communication of a data center. When a large-scale QP is concurrent in the RMDA, the efficient scheduling processing problem of different QP requests is solved, the lossless request signal, the out-of-order QP request and the waste-free scheduling period are ensured, the scheduling fairness problem of different QPs is effectively solved in combination with congestion control, the blocking of other QP processing by a large message and the excessive number of requests needing to be processed in a single scheduling are avoided, and the head blocking problem is solved.

The utility model provides a waste heat recovery device for a steam pipeline, belongs to the technical field of waste heat recovery, and aims to solve the problems that high-temperature and high-pressure steam is transmitted for a long time in a conveying pipeline with too many bends, the consumption is high, the efficiency of converting the high-temperature steam into electric energy is not high, a pipeline system is complicated, and the conveying process needs to be detected. The heat utilization mechanism is mounted in the steam conveying pipeline; one end of the conveying mechanism is fixedly communicated with the heat utilization mechanism; the liquid control mechanism is communicated with the conveying mechanism; the steam conveying pipeline is arranged in the steam boiler and is resistant to high temperature and corrosion, the heat preservation performance is high, and a steam pipe network is monitored in real time; the heat utilization mechanism is inserted into the steam conveying pipeline to improve the steam utilization rate; high-temperature and low-temperature liquid flows in or out in a one-way mode, and auxiliary absorption of waste heat is rapidly completed.

The invention discloses a computing power resource dynamic slicing and isolation method for an artificial intelligence target range, and the method comprises the following modules: S1, business feature perception and dynamic resource slicing: carrying out the analysis and modeling of an artificial intelligence attack and defense task through a mechanism, and mapping a task demand into a corresponding computing power slicing specification; s2, a driving-level video memory firewall, which monitors the video memory usage behavior of the virtual instance in real time on a driving layer of an execution node, and prevents attack behaviors such as abnormal occupation of the video memory; and S3, carrying out atomized rollback based on the model state of the shared memory. According to the mechanism, rapid recovery of the environment is realized through page table resetting. The method is oriented to an artificial intelligence target range, and the problems of static low efficiency, insufficient safety and isolation, lack of a quick recovery mechanism and the like of resource scheduling are solved through the mechanisms.

The invention discloses a regional multi-microgrid cluster-oriented collaborative operation system, which relates to the technical field of power system dispatching and control, and comprises an asynchronous sensing fusion module, an asynchronous receiving and time axis reconstruction module, a time axis reconstruction module and a time axis reconstruction module, which are used for carrying out asynchronous receiving and time axis reconstruction on operation state data reported by a plurality of microgrid nodes; the credibility level is determined based on the communication link state and the data integrity, and weighted merging is carried out in the same time window according to the credibility level to form a unified state view; and the prediction difference modeling module is used for extracting state change characteristics among micro-grid nodes based on the state view, determining state offset information and a historical change boundary thereof, and taking the state offset information and the historical change boundary as numerical constraints of scheduling calculation. According to the system provided by the invention, through the asynchronous sensing fusion module, time axis reconstruction is realized on the basis of asynchronous data receiving, the credibility level of each piece of data is dynamically evaluated in combination with a communication link state and data integrity, and then weighted fusion is carried out in a unified time window.

The application relates to a dynamic distributed short-term scheduling method and system for short-term precipitation prediction, which comprises the following steps: inquiring whether new radar echo data is received in a preset storage folder; if the new radar echo data is received in the storage folder, the new radar echo data is segmented, and a plurality of calculation units are obtained after the segmentation; the segmented calculation units are stored, and the stored calculation units are dynamically allocated to each precipitation prediction processing machine; precipitation prediction is performed on the calculation units based on a preset space-time prediction model, and precipitation prediction results of the calculation units are respectively generated; the precipitation prediction results are combined and precipitation prediction data is generated. The application realizes efficient scheduling of multiple machines for dynamic distributed short-term scheduling.

The utility model provides a bank outlet delivery machine control system. The bank outlet delivery machine control system comprises a client, a delivery total device, a dispatching center and an outlet terminal. Data communication connection is established between the client and the delivery total device; and data communication connection is established between the delivery total device and the dispatching center and the network terminal. Wherein the delivery total device comprises a delivery middleware and a delivery transaction cloud, and the delivery middleware adopts an intelligent picking-up cabinet.

This application relates to the technical field of resource scheduling, and in particular to a task scheduling method, apparatus, and computer device. The method is applied to a scheduler and includes: in response to a task acquisition request reported by a computing container, determining a target task instruction from an instruction queue; wherein the instruction queue contains task flows of received tasks to be executed, and each task flow includes at least one task instruction; the computing container includes multiple threads, and the task acquisition request is generated by the computing container in response to the determination of an idle thread, where the idle thread is a thread in an idle state; generating an allocation instruction containing the target task instruction, and feeding the allocation instruction back to the computing container; the allocation instruction is used to instruct the computing container to call an idle thread to execute the target task instruction. The solution of this application can improve hardware utilization.

The application discloses a transmission line, wherein the transmission line comprises a first outer linear stator, a first bifurcated stator, a first inner linear stator, a cross stator and a first power line; the first outer linear stator transmits a mover in a first direction; the first bifurcated stator comprises a first confluence end, a first and a second bifurcation end, the first confluence end is spliced with the first outer linear stator; the first inner linear stator is spliced with the first bifurcation end and transmits a mover in a second direction perpendicular to the first direction; the cross stator comprises a first splicing end, a second splicing end, a third splicing end and a fourth splicing end arranged in sequence and perpendicularly, the first splicing end is spliced with the first inner linear stator; the first power line connects the first outer linear stator, the first bifurcated stator, the first inner linear stator and the cross stator in sequence. The technical scheme of the application can simplify the wiring mode of the transmission line.

The invention provides a real-time collaborative Web test capturing method and system based on multi-feature scoring, and relates to the technical field of Web automatic testing. According to the method, the interactive coordinates clicked by the user are obtained, cyclic penetration detection is carried out, and a candidate element sequence is output; then five-dimensional feature quantitative scoring is executed, a scoring result is output to select a target element, a corresponding unique identifier is generated, and a capture event is packaged; a change observer interface is used for monitoring change, and an anti-shake mechanism is combined to perceive the page steady state of the browser, so that when the page is structurally changed or redrawn, a related monitor is awakened, computing power is distributed, a target element is reselected, and a capture event is packaged, and when the page is stable, the capture event is sent to a back-end server. And broadcasting to other cooperation terminals in the same room is facilitated. According to the method, the problems of misalignment of element positioning and low cooperation efficiency can be effectively solved, and efficient real-time cooperation and financial-level data security protection of distributed crowdsourcing testing are realized while system resource loss is reduced.

The invention relates to the technical field of network transmission optimization, and provides a data transmission method and device. The method comprises the steps of receiving credit authorization information sent by a receiving end, wherein the credit authorization information is used for authorizing a sending end to transmit to-be-transmitted data based on a credit authorization amount; determining a transmission performance index of each data transmission path; determining a credit allocation weight of each data transmission path according to the transmission performance index; determining the target credit of each data transmission path according to the credit authorization amount and the credit allocation weight of each data transmission path; and based on the target credit of each data transmission path, distributing the to-be-transmitted data to the corresponding data transmission path for transmission. According to the data transmission method provided by the invention, multi-path data transmission under active control is realized, and the robustness of data transmission and the resource utilization rate are improved.

Embodiments of the present disclosure disclose a data processing method based on reinforcement learning, an electronic device and a readable medium. The method is applied to an execution node, a policy node and a training node included in a reinforcement learning system, and a specific implementation of the method comprises: performing environment simulation by the execution node to obtain an observation result; performing policy updating by the training node according to a training sample and a model to be trained, wherein the training sample is transmitted from the execution node to the training node through a sample transmission stream; the policy node performs policy inference according to the observation result and updated policy information, wherein in a non-in-line inference mode, the execution node and the policy node perform data transmission on the observation result and an action to be executed through an inference transmission stream; and the execution node executes the action to be executed to generate an updated observation result. The implementation realizes effective scheduling of large-scale computer hardware resources, meets the training requirements for complex reinforcement learning tasks, and improves the training speed.

The invention belongs to the field of multi-energy power system dispatching, and relates to a water-wind-light power source base power generation dispatching method considering alternating current and direct current peak regulation requirements. The method comprises the following steps: firstly, constructing an optimal scheduling model containing time and space double dimensions; in the aspect of time, a differentiated objective function is adopted in a dry period, the dry period minimizes the sum of squares of residual loads of an alternating current receiving end to smooth the loads, and the flood period minimizes the system abandoned electric quantity to guarantee clean energy consumption. In space, the model overall plans multiple types of actual constraints such as cascade hydraulic connection, upstream and downstream collaboration, water and electricity distribution proportion, direct-current stepped power and standby, and differential peak regulation driving is implemented on alternating-current and direct-current receiving ends. And through model solution, optimal distribution of base output among multiple receiving ends and cascade power station cooperative scheduling are realized. According to the method, the resource utilization efficiency and the power grid operation safety are remarkably improved, and a reliable method is provided for efficient scheduling of the alternating-current and direct-current hybrid delivery clean energy base.

This invention proposes a method for monitoring network signal operation. By deploying a centralized management service in a cloud computing environment, it performs threshold judgments on aggregated signal data, dynamically acquires additional cloud resources to support monitoring, and simultaneously constructs a cloud-based signal monitoring dashboard to update remote signals in real time, identify abnormal patterns, and integrate them into the network monitoring system. For anomalies related to resource scheduling, it obtains optimized paths, enhances dashboard functionality, and ultimately adjusts management services based on communication feedback data. This method achieves efficient aggregation of signal data, accurate identification of anomalies, and intelligent resource scheduling, significantly improving the reliability and response speed of cross-regional communication monitoring and providing stable support for multi-regional network operation.

The application provides a three-dimensional resource dynamic allocation method and system based on grain storage, and belongs to the technical field of digital twinning. The three-dimensional resource dynamic allocation method comprises the following steps: deploying a multi-level management architecture of cloud rendering hosts; wherein the multi-level management architecture is used for hierarchical management of the cloud rendering hosts and dynamic scheduling of three-dimensional resources; controlling the multi-level management architecture to obtain an access request of a user terminal to the three-dimensional resources, find and schedule the three-dimensional resources corresponding to the access request to a local warehouse area; judging whether the cloud rendering hosts in the local warehouse area are sufficient in task processing resources; if the task processing resources are sufficient, allocating the three-dimensional resources to the cloud rendering hosts in the local warehouse area for cloud rendering through a resource distributed deployment mode; if the task processing resources are insufficient, scheduling and allocating the three-dimensional resources to a warehouse area or a region above in stages; and forwarding the rendered video stream to the user terminal. The application can solve the problems of insufficient flexibility of resource allocation and insufficient cross-region scheduling capability in the prior art.

This application discloses a method, apparatus, and electronic device for predicting user electricity usage data. The method, applied in the field of data processing, includes: collecting target electricity data from a target user; inputting the target electricity data into a target prediction model to calculate the target probability distribution of the target user's electricity usage data after responding to a power dispatch request; and predicting the target user's electricity usage data based on the target probability distribution. This application solves the problem in related technologies where, due to high sample acquisition costs and limited training sample capacity, prediction accuracy is low when using neural network models to predict a user's ability to regulate electricity in response to a power dispatch request.

This invention relates to the field of task scheduling technology, specifically to a method for unloading and scheduling interactive tasks and resources in intelligent terminals based on edge-side state awareness. The method includes the following steps: collecting CPU utilization, GPU utilization, and communication latency data; smoothing communication latency to form an edge-side state indicator set; calculating a suitability index based on the edge-side state indicator set and task characteristics; generating a priority ranking table by scoring and fusing according to rules; selecting the first-ranked path; analyzing computing power and latency to construct an execution strategy; invoking dynamic resource scheduling; monitoring latency and providing feedback to adjust weights to generate resource scheduling instructions. In this invention, by collecting CPU and GPU utilization and communication latency in real time, using a sliding window to smooth and optimize scheduling, weighting the maximum tolerable latency with data size to match edge-side state and generate priorities, and dynamically adjusting resource allocation to match execution paths and network latency, this method avoids resource waste and task delays caused by static rules, thereby improving performance and utilization.

This application applies to the field of mining technology, providing a method, apparatus, and computer device for multi-objective optimization scheduling of mining equipment based on digital twins. The method includes: acquiring mining operation parameters and mining equipment parameters, and inputting these parameters into a multi-objective optimization model to obtain a set of candidate scheduling schemes. For each candidate scheduling scheme, the scheme is input into a target simulation model constructed based on digital twin technology to obtain equipment operation data. Based on the equipment operation data corresponding to the multiple candidate scheduling schemes, a target scheduling scheme is determined from the set of candidate schemes. The target scheduling scheme is used for optimal scheduling of mining equipment during the mining operation. This application achieves scientific, efficient, and highly robust scheduling of mining equipment through a closed-loop combination of multi-objective optimization and digital twin simulation.

The application discloses a dynamic robust feasible region solving method for integrated energy system scheduling robust optimization, comprising the following steps: S1, constructing an integrated energy system model according to the integrated energy system; S2, converting the integrated energy system model into a compact form to obtain a compact model, and constructing an uncertainty set according to the compact model; S3, defining the correlation and continuity of decision variables in a preset number of time periods according to the uncertainty set to obtain a first set; the first set comprises robust feasible regions of the preset number of time periods; S4, rolling updating the robust feasible regions of the preset number of time periods in the first set; optimizing first constraints according to the first set to obtain second constraints; S5, solving the dynamic robust feasible region according to the second constraints and the first set to obtain a second set; the second set comprises robust feasible regions of the preset number of time periods. The application further discloses a dynamic robust feasible region solving system for integrated energy system scheduling robust optimization.

The invention particularly relates to an unmanned aerial vehicle emergency disaster rescue cluster system and method based on a large visual language model, and the method comprises the steps: carrying out the space-time alignment of collected multi-modal environment observation data through a cross-modal space-time fusion mechanism; a multi-mode encoder and an online contrast learning mechanism are combined, a large visual language model is used for generating structured situation description containing disaster objects, positions and states, and an optimal path is dynamically planned through a Monte Carlo tree search algorithm; on the basis, the flight state of the unmanned aerial vehicle is monitored in real time through a time sequence anomaly detection network, and task migration and topology recombination are triggered when the unmanned aerial vehicle is abnormal. According to the method, the disaster situation identification accuracy can be improved, the task planning response time can be shortened, the rescue efficiency can be remarkably improved, the method is suitable for high-dynamic disaster scenes, a reliable technical framework is provided for multi-source heterogeneous data fusion and swarm intelligent decision making, and then the problems that a traditional unmanned aerial vehicle cluster rescue system is poor in flexibility and low in adaptive capacity are solved.

This application provides a method and system for scheduling nest resources for emergency inspection tasks. The method includes: obtaining the target location of the emergency inspection task; retrieving nest resources based on the target location to obtain a candidate nest set; obtaining the status information of each candidate nest in the candidate nest set, the status information indicating whether the nest is in a planned occupancy state; for each candidate nest in the candidate nest set, if the status information indicates that the candidate nest is in a planned occupancy state, obtaining the task attribute of the task to be executed corresponding to the candidate nest; evaluating the releaseability of the candidate nest based on the task attribute to obtain a releaseability score; and determining the target nest from the candidate nest set based on the status information of each candidate nest and all releaseability scores. This method improves scheduling accuracy, enhances emergency response capabilities, optimizes resource utilization, and improves scheduling stability, achieving efficient and precise scheduling of nest resources.

The application discloses a multi-agent collaborative scheduling method and system for scientific and technological information analysis, relates to the technical field of scientific and technological information analysis, and comprises the following steps: a task publishing layer outputs a standardized task description by performing task demand element analysis; an intelligent coordination and scheduling layer performs closed-loop collaborative decision after receiving the standardized task description and outputs a task allocation instruction; and a professional agent execution layer drives a data acquisition agent group to acquire scientific and technological information data through a data access adapter of a data resource layer and execute a scientific and technological information analysis task, and outputs a structured processing result, which is visually presented to a user through the task publishing layer. The application solves the technical problem that a scientific and technological information processing system in the prior art lacks intelligent task allocation, leading to low information processing efficiency, achieves four-layer architecture closed-loop collaboration and efficient scheduling of a professional agent group for scientific and technological information analysis tasks, and improves the technical effects of task allocation accuracy and information processing efficiency.

The invention discloses an asset management method and device, electronic equipment and a storage medium, and can be applied to the technical field of communication. According to the method, the global asset information is acquired, and asset filing is performed according to the global asset information, so that the global assets can be effectively managed; after asset borrowing request information is obtained, borrowing the archived assets according to the asset borrowing request information; after asset allocation request information is obtained, allocating the archived assets according to the asset allocation request information; after the asset maintenance request information is obtained, the archived assets are maintained according to the asset maintenance request information, so that the global assets can be effectively scheduled, the workload of managers is reduced, and the timeliness of asset information is improved.

This invention discloses an adaptive router and its routing method based on a three-dimensional on-chip network with irregular inter-layer interconnections. It utilizes the vertically connected topology of the three-dimensional on-chip network, a healthy TSV (Through Silicon Via) location mechanism, and a congestion propagation network to adaptively reconstruct routing paths based on congestion level, injection rate, and TSV availability. By designing healthy TSV information tables, healthy TSV distance information tables, healthy TSV congestion information tables, a routing decision module, and healthy TSV propagation and congestion awareness mechanisms, it achieves collaborative operation between distance-based and congestion-based routing algorithms to reduce performance losses caused by congestion and failures. This invention uses relatively low hardware overhead, enhances the reliability and adaptability of three-dimensional on-chip network routing, and effectively reduces the average packet latency in the presence of irregular vertical interconnections and dynamic network traffic changes, significantly improving the overall performance and robustness of the system under fault and congestion environments.

This invention relates to a custom time-delay message processing method based on a distributed message queue, belonging to the field of distributed message queue technology. The method includes: a client receiving an arbitrary delay duration specified by the user, calculating a target Unix timestamp based on the client's current system time, and sending a message carrying that timestamp to a message broker layer; the message broker layer storing the message in a delay queue and inserting it into the corresponding time slot of a hierarchical circular time wheel according to the target Unix timestamp; the hierarchical circular time wheel consists of multiple levels of circular arrays, connected by an overflow mechanism; the message broker layer advances a global pointer at fixed time intervals, scans the messages to be scheduled in the time slots, and transfers expired messages from the delay queue to a regular queue for consumer consumption. This achieves flexible support for arbitrary delay durations and achieves task scheduling with O(1) time complexity through the hierarchical circular time wheel.

The invention discloses a wind power prediction method and device, equipment and a storage medium. The method comprises the steps of obtaining wind power data; decomposing the wind power data into a plurality of data subsequences by using a variational mode decomposition technology; predicting each data sub-sequence in combination with a long-short-term memory network model to obtain a prediction result; and performing weighted superposition on the prediction results of the data subsequences to obtain a final wind power prediction power value. The method comprises the following steps: collecting and sorting wind power data of a given wind power plant, and preprocessing the data; decomposing the wind power data by using a variational mode decomposition technology, and predicting each data sub-sequence by using a long short-term memory network model; and performing weighted stacking on the prediction results of the data subsequences to obtain a final wind power prediction power value. The method can effectively overcome the defects of a traditional prediction method in the aspect of processing wind power fluctuation and randomness, and enables a wind power plant to participate in operation and scheduling of a power system more effectively.