Patents
Literature
Patsnap Eureka AI that helps you search prior art, draft patents, and assess FTO risks, powered by patent and scientific literature data.

50results about How to "Shorten completion time" patented technology

Charging pile power regulation and control method and system

The invention discloses a charging pile power regulation and control method and system, and particularly relates to the technical field of charging pile management, data is collected from a multi-dimensional data source and summarized to form a preliminary information set, and a standard information set is obtained through format standardization processing; feature data of the charging pile are extracted based on the standard information set, and charging task analysis and resource occupation analysis are carried out respectively; inputting the analysis result into a pre-trained machine learning model to generate a charging priority coefficient; constructing a priority matrix, and dividing the charging piles into high priorities and low priorities according to the charging priority coefficients; and finally, dynamically adjusting a power distribution strategy according to the priority type, optimizing low-priority task power distribution while meeting high-priority task requirements, and improving the operation efficiency and the resource utilization rate of the charging station. According to the method, the real-time operation state and the charging task characteristics of the charging piles are combined, the power distribution strategy is dynamically adjusted, the priority division and power distribution proportion of the charging piles are adjusted in real time, and the self-adaptive capacity of operation of the charging station is improved.
Owner:江西驴充充物联网科技有限公司

Optimization method for flexible job shop scheduling problem considering adjustment time

The invention relates to the technical field of flexible job shop scheduling in intelligent manufacturing, in particular to an optimization method for a flexible job shop scheduling problem considering adjustment time, and the method comprises the steps: initializing parameters of a collaborative variable neighborhood search algorithm; initializing a population; jittering operation is executed, and two new solutions are generated through the jittering operation; executing a variable neighborhood descent operation; updating the current algebra; judging whether a restart strategy is executed or not; and selecting a solution with shorter maximum completion time in the improved solutions generated by the collaborative variable neighborhood search algorithm as an initial solution of the constrained programming model, and further optimizing the quality of the solution through the constrained programming model. The invention provides an effective method for solving the problem of flexible job shop scheduling considering the adjustment time of sequence dependence, and makes substantial progress in reducing the maximum completion time of the shop.
Owner:YANCHENG FENGMANG TECHNOLOGY CO LTD

Intelligent recommendation method and system based on shared robot

The invention belongs to the field of intelligent manufacturing, and provides an intelligent recommendation method and system based on a shared robot, and the method comprises the steps: receiving a picture and a processing demand provided by a user; determining the processing attribute of the target workpiece according to the picture; determining a processing sub-task according to the processing attribute; determining an optimal shared robot combination according to the processing subtasks and the processing requirements; and displaying the optimal shared robot combination to the user.
Owner:DONGGUAN XINBAIREN ROBOT TECH CO LTD

Rocket tank assembly process irrelevant parallel machine scheduling method considering variable resources

The invention discloses a variable resource-considered uncorrelated parallel machine scheduling method for a rocket tank assembly process. According to the method, for multi-variety and small-batch rocket tank assembly tasks, the influence of skill constraints and skill proficiency of workers and cooperation efficiency of worker combinations on scheduling optimization is comprehensively considered, and an irrelevant parallel machine scheduling model with the minimum completion time as the target is constructed. By solving the scheduling optimization model, an optimal scheduling scheme is obtained, reasonable matching among the tasks, the work center and the workers is ensured, the utilization rate of the work center and the cooperation efficiency of the workers are improved, and therefore the completion time of the assembly tasks is effectively shortened. And on the premise that various constraint conditions are met, the optimal scheduling scheme is applied to the rocket tank assembly process, so that the assembly efficiency is improved.
Owner:NANJING UNIV OF AERONAUTICS & ASTRONAUTICS

Flexible job shop energy-saving batch flow scheduling optimization method considering machine tool deterioration and preventive maintenance

The invention relates to a flexible job shop energy-saving batch flow scheduling optimization method considering machine tool deterioration and preventive maintenance, and belongs to the technical field of shop scheduling. Comprising the following steps: constructing a flexible job shop energy-saving batch flow scheduling mathematical model taking completion time, total energy consumption and total production cost as optimization targets and considering machine tool deterioration and preventive maintenance; a model solving method is put forward, a strategy is introduced based on population diversity of historical populations, and the coping capacity of the algorithm to environment changes is enhanced; developing a self-adaptive crossover mutation operator, and balancing the diversity and convergence of the algorithm; and proposing a neighborhood search strategy, and exploring a local optimal solution of the energy-saving batch flow scheduling problem. According to the method, the problem of flexible job shop energy-saving batch flow scheduling multi-objective optimization considering machine tool deterioration and preventive maintenance can be efficiently solved, under-maintenance or excessive maintenance of the machine tool is effectively avoided, production interruption and maintenance cost are reduced, and a scheduling scheme with excellent completion time, total energy consumption and total production cost is obtained.
Owner:SHENYANG UNIVERSITY OF TECHNOLOGY

Multi-substream parallel forwarding scheduling method for novel intelligent computing center network

The invention discloses a multi-substream parallel forwarding scheduling method for a novel intelligent computing center network. The whole communication forwarding process is divided into three stages of data stream segmentation and distribution, parallel transmission and receiving end sorting and recombination. The cooperation of a data stream segmentation distribution strategy of a sending end and an out-of-order data packet sorting and recombination operation of a receiving end is comprehensively considered, an optimal distribution ratio and a corresponding optimal forwarding path are determined, and the optimal distribution ratio and the corresponding optimal forwarding path are mapped to a physical network to complete parallel transmission of multiple substreams. According to the method, the problem of failure of a traditional network scheduling method caused by new characteristics of small flow quantity and large single flow bandwidth of distributed training service data flow of an AI large model in a novel intelligent calculation center is solved.
Owner:NANJING UNIV OF AERONAUTICS & ASTRONAUTICS

Heterogeneous multi-core task scheduling method based on whale optimization algorithm

The invention discloses a heterogeneous multi-core task scheduling method based on a whale optimization algorithm. The heterogeneous multi-core task scheduling method comprises the following steps: step 1, task information acquisition and model establishment; firstly, to-be-scheduled task information needs to be obtained, and the task information comprises task types, calculation amount, memory requirements and communication overhead; then, according to the architecture characteristics of the heterogeneous multi-core processor, establishing a task scheduling system model; 2, the number of white whale individuals is initialized and corresponds to a task scheduling scheme; 3, solving an individual with the highest fitness based on a white whale optimization algorithm, and taking the individual as an optimal solution; and 4, determining a final task scheduling scheme according to the optimal solution, wherein the final task scheduling scheme comprises a processor to which each task is allocated and an execution sequence. 5, the scheduling scheme is applied to the heterogeneous multi-core system, and task scheduling is executed. According to the method, the task scheduling efficiency is remarkably improved, and the overall task completion time is shortened.
Owner:TIANJIN JINHANG COMP TECH RES INST

GPU (Graphics Processing Unit) sharing scheduling method and device supporting heterogeneous operation scale

The invention discloses a GPU sharing scheduling method and device supporting heterogeneous job scale, and the method comprises the steps: carrying out the interference portrait analysis of a newly submitted job, obtaining the interference feature representation of the job, and adding the interference feature representation into a job queue; collecting cluster operation information and job information; determining a job scheduling interval; on the basis of the job runtime information, the job priority is updated, and the jobs are sorted through a maximum sharing efficiency priority method; traversing jobs in the job queue, and determining jobs to be run and jobs to be preempted; calling an optimal packaging strategy and job co-located interference prediction model to determine a placement position of a to-be-operated job; placing jobs and updating a cluster state; and monitoring a new job submission condition and updating a scheduling interval, if a new job is submitted, skipping to carry out interference portrait analysis on the newly submitted job, otherwise, skipping to the step of determining the job scheduling interval. According to the method, the job completion time and the queuing time can be effectively reduced, and the resource scheduling efficiency of the GPU cluster is remarkably improved.
Owner:SOUTH CHINA UNIV OF TECH

Navigation planning method and system based on detailed terrain information and vehicle characteristics

The invention discloses a navigation planning method and system based on detailed terrain information and vehicle characteristics. The method comprises the following steps: acquiring latest terrain, weather and road condition information in real time; carrying out data preprocessing on the vehicle performance parameters and the terrain, weather and road condition information; extracting a topographic feature vector and a meteorological feature vector from the preprocessed topographic and meteorological information; calculating the driving feasibility of the vehicle on the specific terrain by using the vehicle performance parameters and the terrain feature vectors, and establishing an evaluation model; and calculating the driving risk degree of the vehicle under the specific terrain and weather conditions by using the vehicle performance parameters and the terrain and weather feature vectors, and establishing a risk model. And on the basis of the evaluation model and the risk model in the step S300, generating one or more possible navigation routes by applying a path planning algorithm to generate an optimal or more alternative navigation paths. The navigation precision and safety of the special off-road vehicle in a complex and severe environment are remarkably improved, and related risks are reduced.
Owner:DONGFENG OFF ROAD VEHICLE CO LTD

Optimization method for edge computing unloading decision and resource allocation of Internet of Vehicles

The invention relates to an optimization method for edge computing unloading decision and resource allocation of the Internet of Vehicles, and belongs to the technical field of edge computing of the Internet of Vehicles. According to the method, under the scene that the task generated by the vehicle can be partially uploaded to the edge server to be processed, a communication model of the vehicle and the edge server and a task unloading model of the vehicle are established, and the task unloading proportion, the channel bandwidth and the resources allocated by the edge server for processing the unloading task are combined, so that the total completion time of the system is minimized as the target; the method comprises the following steps: establishing a system total overhead optimization function model, solving the optimization function model by using an improved genetic algorithm under the condition that wireless transmission resources and computing resources are limited, and obtaining optimal solutions of a task unloading proportion, a channel bandwidth and resource allocation, thereby effectively reducing the vehicle-mounted task completion time and improving the system performance. And meanwhile, the method is excellent in the aspect of reserving the optimal solution set, so that the exploration capability of the algorithm in a wide search space is improved, and the risk that the algorithm falls into local optimum and cannot jump out is remarkably reduced.
Owner:CHANGCHUN UNIV OF SCI & TECH

Operation and maintenance task scheduling method based on operation and maintenance data

The invention discloses an operation and maintenance task scheduling method based on operation and maintenance data, and relates to the technical field of operation and maintenance management. The method comprises the steps of initializing an operation and maintenance task set, an operation and maintenance resource set, a first time matrix and a second time matrix; a population is initialized, the population comprises a plurality of individuals in one-to-one correspondence with the operation and maintenance tasks, and an operation and maintenance resource sequence occupied by the corresponding operation and maintenance task is recorded in each individual; based on the first time matrix and the second time matrix, according to a preset task execution logic, solving a fitness value, inversely proportional to the spent time of the operation and maintenance task, of each individual; performing crossover and mutation operation on individuals with fitness values lower than a first preset threshold value in the population so as to iteratively update the population; and when an iteration termination condition is reached or the number of iterations reaches a preset maximum number of iterations, determining an operation and maintenance task scheduling scheme based on the operation and maintenance resource sequence occupied by the operation and maintenance task corresponding to each individual in the latest population. The method disclosed by the invention can improve the utilization rate of operation and maintenance resources and reduce the operation and maintenance cost.
Owner:QIMO TECH (GUANGZHOU) CO LTD

Optimization method for solving flexible job-shop scheduling with resource constraints

The present application relates to the technical field of flexible job shop scheduling in intelligent manufacturing and production scheduling, and particularly relates to an optimization method for solving flexible job shop scheduling with adjusted resource constraints. The method comprises the following steps: initializing parameters and randomly generating an initial population; using a crossover operator and a mutation operator in sequence to evolve the current population; executing a hybrid decoding strategy on the current population; sorting the individuals in the current population in ascending order of maximum completion time to form an elite population, and updating the elite population through problem-specific local search; judging whether the evolution condition is met, and if yes, executing mathematical evolution based on CP, and outputting a final solution when the running time reaches the total running time. The present application has the positive effects of reducing resource waiting time, improving machine utilization, and improving the resource utilization efficiency and scheduling performance of the entire workshop production.
Owner:LIAOCHENG UNIV

A flow group scheduling method in optoelectronic hybrid data center network

The present invention discloses a flow group scheduling method in an optoelectronic hybrid data center network, and belongs to the field of network transmission technology. The method first divides the input traffic into flow groups according to business semantics, and then inputs the network topology structure. Secondly, based on the flow group information, the order of the flow groups is solved in combination with the characteristics of the optical circuit switch and the electrical packet switch port bandwidth to obtain an ordered flow group sequence. Finally, the flow groups and the sub-flows contained therein are scheduled according to the flow group sequence to optimize the transmission of the traffic in the optoelectronic hybrid network structure, reduce the completion time of the flow group transmission, and achieve the purpose of comprehensive optimization of the overall transmission. The present invention can optimize the scheduling and transmission process of the traffic in the optoelectronic hybrid structure, and can be applied to the scenarios of integrated services in data centers, etc.
Owner:NANJING UNIV OF AERONAUTICS & ASTRONAUTICS

A unified communication agent supporting concurrent flow scheduling in cluster computing

The present invention provides a unified communication agent that supports concurrent flow scheduling in cluster computing. By deploying the unified communication agent on machines participating in cluster computing, the number of flows between two pairs of servers is reduced to one. Specific invention highlights include: ① When a cluster computing framework on a server has a communication demand, the framework forwards the communication demand to the agent of the present invention deployed on the machine, rather than directly initiating a data flow; ② After collecting the communication demand, it interacts with the currently deployed co-flow scheduling algorithm to calculate the rate or order of the unified flow; ③ Initiate a data flow based on the result calculated in ②, and distribute the transmitted data to the cluster computing framework on the server respectively; the present invention can be applied to a data center that deploys a cluster computing framework; by reducing the number of concurrent flows, the network performance is optimized to accelerate the completion speed of cluster computing tasks and improve computing efficiency.
Owner:JIANGSU FUTURE NETWORKS INNOVATION

A resource management method and system in a cloud computing environment

A resource management method and system in a cloud computing environment, relating to the technical field of cloud computing, is used to solve the problem that the existing static resource management of virtual nodes in a cloud computing environment cannot effectively improve their operating performance. The technical points of the present invention include: collecting the CPU, memory and bandwidth resource information of the virtual node, proposing a DDPG-based resource allocation algorithm to calculate the target resource configuration amount of each virtual node according to the above resource information to obtain the result of the resource configuration decision, and dynamically adjust the resources of each virtual node according to the result, thereby improving the operating performance of the virtual node on the basis of improving the overall resource utilization; further, in the DDPG-based resource allocation algorithm, the Actor network is used to generate the corresponding action output, and the ratio-based output is converted into the target resource configuration size through an action decoding process. The present invention can shorten the completion time of different applications as a whole, and will not generate a large amount of system overhead.
Owner:HARBIN INST OF TECH

Distributed computing cooperative flow routing method and system for satellite time-varying network topology and storage medium

The application provides a distributed computing cooperative flow routing method and system for a satellite time-varying network topology and a storage medium, and the distributed computing cooperative flow routing method comprises the following steps: step one, constructing a cooperative flow transmission time-varying graph model under a satellite inter-satellite link dynamic network; step two, a heuristic routing algorithm based on the cooperative flow transmission time-varying graph; solving the available path set through path calculation on the cooperative flow time-varying graph model, setting a scheduling rule based on traffic priority, and constantly selecting the optimal path under the current situation through the evaluation of the paths in the feasible path set. The application has the beneficial effects that: 1. the time used in satellite distributed computing is effectively reduced; and 2. while effectively reducing the cooperative flow completion time, the efficiency of satellite distributed computing is effectively improved.
Owner:HARBIN INSTITUTE OF TECHNOLOGY (SHENZHEN) (INSTITUTE OF SCIENCE AND TECHNOLOGY INNOVATION HARBIN INSTITUTE OF TECHNOLOGY SHENZHEN)

Task list scheduling method based on vehicle cooperation cluster

The invention provides a task list scheduling method based on a vehicle cooperation cluster, and the method comprises the steps: firstly, designing a vehicle task model and an Internet of Vehicles communication model, and considering the link occupation condition of V2V communication; then, a task list scheduling algorithm based on communication occupation and task priority prediction is designed, the algorithm can predict the calculation priority of each task, executed tasks and vehicle calculation equipment are selected according to priority values, and then the optimal path is selected to transmit data of the tasks. Finally, through experimental comparison between the algorithm and other algorithm schemes, simulation results show that the algorithm reduces the average task completion time by about 27% and improves the efficiency of vehicle computing equipment by about 18%.
Owner:NANJING UNIV OF POSTS & TELECOMM

Overwater side propeller

The utility model discloses an overwater side propeller which comprises a driving mechanism and a connecting device which are detachably connected, the driving mechanism comprises a propeller and a hydraulic motor, the propeller is provided with a connecting shaft, and the output end of the hydraulic motor is connected to the connecting shaft of the propeller so as to be used for driving the propeller; the connecting device comprises a connecting rod and a sliding rail, the top end of the connecting rod is connected with the sliding rail, and the position of the propeller in the vertical direction is adjusted through vertical movement of the sliding rail. According to the overwater propeller, the propellers, the hydraulic motors, the connecting shafts, the connecting rods, the sliding rails, the hydraulic oil cylinders and other structures are combined, the operability of a gate bridge in the training process can be improved, collision of the gate bridge in the training process is reduced, the completion time of each subject of the gate bridge is shortened, and the training efficiency of the gate bridge is improved. And the task execution efficiency of the gate bridge is greatly improved. The structure is widely applied to the fields such as flood fighting and disaster relief.
Owner:李伟邦

Dynamic multi-product integrated scheduling method with tight delivery deadlines

The present invention proposes a dynamic multi-product comprehensive scheduling method with tight delivery deadlines. The method first adopts the "layer priority + short time + equipment priority" strategy to improve the parallel processing capability of the equipment; secondly, a product contribution value strategy is proposed to obtain the product contribution value according to the number of urgent processes, the maximum delivery date and the total processing time of the product, thereby determining the processing order of the product, effectively making multiple products processed on time within the delivery period; finally, a game process dynamic adjustment strategy is proposed to dynamically adjust the process according to different situations, so that the process is compactly processed, the idle time of the equipment is reduced, and the overall utilization rate of the equipment is improved.
Owner:JILIN NORMAL UNIV

Power transmission line iron tower detection and quality management system

The invention relates to the technical field of power equipment quality detection, and particularly discloses a power transmission line iron tower detection and quality management system. The system comprises a user authentication module, an authority management module, a detection plan management module, a sample information management module, a physical and chemical detection management module and a report generation module. Differentiated operation authorities are distributed to different user groups through an authority management module; creation, execution and state management of a detection plan are realized through a detection plan management module, and a sampling list and a detection entrance are automatically generated; complete sample data is forcibly recorded through the sample information management module so as to ensure data traceability; the detection data filling unit supports communication with external detection equipment so as to automatically collect and calculate data, supports multi-user collaborative editing and displays an editing state in real time; and finally, the report generation module automatically synthesizes various types of detection reports according to the template, so that the detection efficiency, the data accuracy and the management standardization are effectively improved.
Owner:ANSHAN ANTA QUALITY INSPECTION CO LTD

A task distribution method and device based on Manage and Worker

The present invention provides a task distribution method and device based on Manage and Worker, which relate to the technical field of electronic digital data processing. A genetic algorithm is used to optimize the dynamic allocation of tasks to nodes, with the goal of minimizing the total task execution time and data transmission time. A real-time monitoring and feedback adjustment mechanism is deployed in a Docker Swarm environment, and a scheduling strategy can be dynamically adjusted according to actual feedback from system operation. The scheduling result can be further optimized by fine-tuning an index, so that the system always maintains the best operating efficiency when facing constantly changing workloads and resource states. A graph theory model is introduced to accurately analyze and optimize the dependencies between tasks, so that tasks with close dependencies can be scheduled to nodes with similar physical locations for execution, thereby significantly reducing communication delays and data transmission time.
Owner:BEIJING BAILONG MAYUN TECH CO LTD

Multi-agent communication information prediction method under ACML model based on neural network

The invention provides a multi-agent communication information prediction method under an ACML model based on a neural network, and relates to the field of multi-agent communication. Establishing a small ball environment; defining actions, local observation and agent team joint rewards of the agents in the small ball environment; an ACML model is established; the communication budget is not limited, and an Actor network, a Critic network, an information coordination network and an information generation network in the ACML model are trained; communication budget is not limited, a plurality of multi-agent communication records are obtained, and an information prediction network data set related to communication budget limitation is obtained; training an information prediction network in the ACML model by using the information prediction network data set related to the communication budget limitation; and applying a communication budget limit, and executing the ACML model to obtain an agent moving trajectory and a team reward. By predicting the information of the agents at the moment of forbidding communication, the task completion time of the multiple agents in the small ball environment is shortened, and team rewards are obviously improved.
Owner:NORTHEASTERN UNIV AT QINHUANGDAO

Quick-change mechanism for replacing the breaking fixture of the die-casting island breaking machine

The invention discloses a quick-change mechanism for replacing a breaking fixture of a die-casting island breaking machine. The breaking machine is vertically provided with a breaking cylinder, and the lower end of the breaking cylinder is connected to a fixture mounting plate through a breaking cylinder shaft. The breaking fixture is fixedly mounted on the fixture mounting plate and moves up and down with the cylinder. The quick-change mechanism comprises a U-shaped rack support and a rack guide rail fixed on the upper part of the rack support, the rack guide rail is symmetrically arranged on the upper part of the rack support, and alignment axes are arranged at the front and rear ends of the rack guide rail. Auxiliary guide rails are symmetrically arranged at the left and right ends of the fixture mounting plate, and guide rail grooves are arranged on the auxiliary guide rails along the front and rear directions. The auxiliary guide rails are used in conjunction with the rack guide rails as support when the breaking fixture is replaced, and the breaking fixture can slide back and forth in the guide rail grooves. Alignment holes are also arranged on the auxiliary guide rails along the front and rear directions. The alignment holes cooperate with the alignment axes on the rack guide rails for alignment and positioning. The invention improves the replacement efficiency of the breaking fixture, and the completion time is shortened from 30 minutes to 7 minutes.
Owner:重庆建设·雅马哈摩托车有限公司

Method, device, equipment, medium and program product for multi-objective computing power scheduling

PendingCN122593987ANarrow down the candidatesforward-looking
The application discloses a multi-target computing power scheduling method, device, equipment, medium and program product, and relates to the technical field of computing power networks. The scheduling device of the method receives task carrying willingness scores and resource state information of a plurality of resource pool systems, and splits a computing power task sent by a task party to obtain a plurality of sub-computing power tasks; the task carrying willingness scores, the resource state information and the sub-computing power tasks are matched to obtain candidate resource pool systems, and scheduling information is sent to the candidate resource pool systems. The candidate resource pool systems determine predicted load evaluation results according to the scheduling information and real-time running data and send the predicted load evaluation results to the scheduling device; the scheduling device determines a target resource pool system of each sub-computing power task according to the predicted load evaluation results and the sub-computing power tasks, and sends a corresponding sub-computing power task to the target resource pool system, so that the target resource pool system executes the sub-computing power task. Therefore, the effect of computing power scheduling is improved.
Owner:CHINA MOBILE GRP HENAN CO LTD +1

A RA-IDNC network coding method based on partial blind transmission

The present invention provides a RA‑IDNC network coding method based on partial blind transmission, comprising: (1) obtaining a forward link packet loss rate and a feedback link packet loss rate of a user node and a channel capacity of the user; (2) the target user feeds back a receiving state to a base station through a feedback link, performs ML judgment on lost feedback, processes nodes according to the judgment result, and updates a feedback matrix; (3) constructs a RA‑IDNC coding relationship graph; (4) determines whether an active node set in the RA‑IDNC coding relationship graph is empty; if the active node set is empty but there are hidden nodes, resets all the hidden nodes to active nodes; (5) if the active node set is not empty, constructs a maximum weight group; (6) performs RA‑IDNC according to data packets corresponding to nodes in the maximum weight group, and goes to (2) for iterative operation until all node sets in the RA‑IDNC coding relationship graph are empty.
Owner:NANJING UNIV OF POSTS & TELECOMM

Conveying belt with following pressing function

The utility model relates to the field of plastic bottle processing and production, in particular to a conveyor belt with a following pressing function, which comprises a pressing frame and a conveying frame, at least one pressing frame is arranged on one side of the conveying frame, each pressing frame comprises a shell, two groups of pressing arms are arranged on the shell, and driving modules for driving the pressing arms to move left and right are arranged in the shell; the driving module comprises a driving screw rod; the pressing arms are pushed left and right through the arranged driving lead screw, so that when the conveying frame conveys the bottle bodies, the single pressing arm can move along with the bottle bodies, after the pressing action is completed, resetting or acceleration can be conducted, and handle assembling can be conducted on the previous bottle body or the next bottle body; therefore, the assembling efficiency of a single pressing arm in unit time is improved, two sets of pressing arms are arranged on one independent pressing frame, the pressing work can be completed in two times, the completion time of the pressing action of the single pressing arm is shortened, and the assembling efficiency of the whole pressing frame is further improved.
Owner:孙浩

Resource scheduling method, system and device for AI training tasks on cloud and medium

The embodiment of the invention provides an on-cloud AI training task resource scheduling method, system and device and a medium, and belongs to the field of cloud computing and artificial intelligence. The method comprises the following steps: deploying a monitoring agent in a cloud environment to collect multi-dimensional resource utilization characteristics of an AI training task in real time; the multi-dimensional resource utilization features are input into a pre-constructed time series prediction model, a short-term resource demand prediction result and a long-term resource demand prediction result are output to cover the whole life cycle of the AI training task, and the time series prediction model is constructed based on a long and short-term memory network; formulating a scheduling strategy and generating a scheduling instruction according to the short-term resource demand prediction result, the long-term resource demand prediction result and the current resource state; and carrying out resource capacity expansion and contraction and task migration operation based on the scheduling instruction. Through multi-dimensional feature prediction and elastic scheduling, the cloud resource utilization rate is remarkably improved, the AI training task completion time and the operation cost are reduced, and meanwhile the system self-adaptive capacity is enhanced.
Owner:SHANDONG LANGCHAO YUNTOU INFORMATION TECH CO LTD

Flow control method and device, equipment and storage medium

Provided in an embodiment of the present application are a flow control method, apparatus and device, and a storage medium, the method comprising: a network device receiving a plurality of data packets, the plurality of data packets comprising a first data packet and a second data packet, the first data packet carrying a first preset tag, the second data packet carrying a second preset tag, the first preset label indicates that the sender does not obtain the credit value allocated by the receiver, and the second preset label indicates that the sender has obtained the credit value allocated by the receiver; if the available length of the cache queue in the network equipment is greater than a first threshold value, writing the first data packet and the second data packet into the cache queue; and sending the data packet in the cache queue. According to the scheme, the network bandwidth can be fully utilized in an AI large model training scene, the flow completion time is shortened, and the training efficiency of the AI large model is improved.
Owner:NEW H3C TECH CO LTD