38results about How to "Improve completion rate" patented technology

The invention discloses a storage, calculation and training AI server load fluctuation self-balancing method based on a recursive game network. The method comprises the following steps: S1, recording real-time load data of each task; s2, constructing a macroscopic hierarchical recursive game network model based on the overall resource pool of the AI server; s3, generating an overall resource allocation strategy through multi-agent competition calculation of the macroscopic hierarchical recursive game network model; s4, constructing a micro-hierarchy recursive game network model based on the resource allocation strategy generated by the macro-hierarchy; s5, sub-task resources are optimally allocated; s6, issuing the adjusted resource quota to the microscopic hierarchical recursive game network model in real time; s7, carrying out resource pre-allocation on the upcoming high-load task; and S8, enabling resource allocation to gradually tend to an optimal state. According to the method, the resource utilization efficiency, the system response capability and the task completion stability of the AI server in a multi-task load fluctuation scene are remarkably improved.

The invention discloses a system and a method based on laboratory process intelligent management, and relates to the technical field of intelligent management. The system comprises an experiment task classification and sorting module, a historical data analysis and task duration prediction module, an overlapping time period analysis and intelligent scheduling module and a real-time monitoring and task progress adjustment module. The experiment task classification and sorting module is responsible for acquiring an initial experiment task list and carrying out task classification and sorting; the historical data analysis and task duration prediction module predicts the total duration of the experiment task by using historical data and provides accurate task duration prediction; time overlapping between the overlapping time period analysis and intelligent scheduling module processing tasks is carried out, and the task execution sequence and the equipment use time are intelligently adjusted; and the real-time monitoring and task progress adjusting module tracks the task execution condition in real time, analyzes the actual and predicted deviation and carries out necessary progress adjustment.

The invention provides a distributed fault detection and dynamic task scheduling system for an unmanned ship cluster, and the system achieves the efficient cooperation of the unmanned ship cluster in a complex marine environment through the cooperation of a sensor data collection module, a data processing and feature extraction module, a fault state classification module, an information sharing module, a weighted voting decision module, a task scheduling module and the like. The system firstly collects state information of each unmanned ship and adjacent ships, performs feature extraction and classification on data, and judges whether a fault exists or not. Fault information is shared to each boat in a cluster through a multi-path redundancy communication mechanism, and weighted decision is carried out by a voting mechanism, so that the accuracy of fault detection is improved. After a fault is confirmed, the system dynamically adjusts task allocation according to task priorities, key tasks are transferred to the unmanned surface vehicle in a good state, and continuity and stability of cluster tasks are guaranteed. Different from traditional centralized control, the distributed control method disclosed by the invention enhances the damage resistance and the self-adaptability of the system.

The invention provides a communication network resource efficiency evaluation method and system, and relates to the technical field of communication networks. The method comprises the following steps: establishing a multi-level index system model of communication network resources; obtaining a weight coefficient of each index element according to the importance degree of the index element in the business demand; establishing an index membership calculation model to obtain the index membership of each end index element in the index system model; performing weighted summation on the index membership degree and the weight coefficient of the end index element to obtain an evaluation value of an upper-level index element; performing weighted summation on the evaluation value and the weight coefficient of the upper-level index element to obtain the evaluation values of the upper-level index element and the lower-level index element; and by parity of reasoning, iterating step by step from bottom to top to obtain an evaluation value of a first-level index element, and performing weighted summation on the evaluation value of the first-level index and the weight coefficient to obtain an efficiency evaluation value of the current communication network resource. According to the method, the problems of blindness, subjectivity and prejudice of existing efficiency evaluation are solved.

The invention discloses a cross-platform content generation and distribution method based on a multi-modal AI, and belongs to the technical field of cross-platform content generation and distribution, and the method comprises the steps: carrying out the content analysis and feature extraction of an original material based on a multi-modal AI model, and generating a structured content label and a semantic vector. By combining a vector matching degree formula of target portrait features and platform features, an adaptation strategy is dynamically generated, it is ensured that content not only conforms to platform rules, but also can accurately reach a target group, the conversion rate and user viscosity are finally improved, visual, text and semantic vectors are aligned through a Transform multi-modal fusion model, a cross-modal joint representation vector is generated, and the user experience is improved. And in combination with an adversarial generative network, differentiated variants of the same theme are generated in batches, and a content diversity score mechanism ensures that generated contents are balanced between creativity and compliance.

The invention relates to a layered cathode structure based on functional layer-buffer layer alternate distribution and a rapid trial-manufacturing method thereof, and belongs to the technical field of machining, the layered cathode structure comprises a functional structure surface layer and a matrix, and the functional structure surface layer is further formed by stacking and alternately arranging a plurality of sub functional layers and buffer layers. In actual service, a refractory oxide layer is generated in situ on the basis of the functional layer made of the thermochemical material, self-protection is completed, and electrons are emitted to support arc discharge; and the buffer layer is made of a nanocrystalline material with high plastic deformation capacity, so that the thermal stress of the functional layer can be released through proper deformation of the buffer layer, and meanwhile, cathode spot migration can be accelerated when the buffer layer is in contact with a cathode arc so as to slow down cathode burning loss. Through the synergistic effect of the functional layer and the buffer layer, the problems of cracking, splitting and coating falling caused by plasma discharge shock heating and shock cooling of the layered cathode in application scenes such as electric spark machining hole making and electric arc welding are solved, meanwhile, the ablation uniformity is improved, and the ablation rate is reduced.

The invention discloses an industrial waste salt rapid decomposition treatment device and method, and relates to the technical field of industrial waste salt treatment.The industrial waste salt rapid decomposition treatment device comprises a fixing unit which comprises a fixing steel frame, a supporting frame arranged at the bottom of the fixing steel frame and a ladder arranged on the supporting frame; the reaction unit comprises a reaction assembly arranged on the fixed steel frame; the catalytic unit comprises a catalyst adding part arranged on the reaction assembly and a catalyst liquid supplementing part arranged on the catalyst adding part, and the catalyst liquid supplementing part is connected with the fixed steel frame. Through cooperation of the electric push rod and the quantitative barrel, accurate control of the adding amount of a catalyst is achieved, so that the reaction efficiency and the treatment speed are improved, automatic operation is formed through cooperation with the reaction sensor, the operation process is simplified, the operation complexity is reduced, the achievement rate of a treatment target is increased, and resource waste is reduced.

The present disclosure relates to an activity participation method, a resource sending method, a device, a terminal and a storage medium, and relates to the field of Internet technology. The method includes: displaying an entry control of a target activity on the live broadcast interface of the anchor; in response to a trigger operation on the entry control, displaying multiple types of target activities, and the flow volume is used to indicate the number of second virtual resources received by the anchor; in response to a selection operation on any type of target activity, and the anchor's flow volume meets the admission threshold of the flow volume corresponding to any type of target activity, displaying a participation interface of any type of target activity, and a participation control is displayed on the participation interface; in response to a trigger operation on the participation control, participating in any type of target activity. The method improves the diversity of target activities. Moreover, since the target activity of the type in which the anchor participates matches the anchor's flow volume, the probability of the anchor completing the task in the participating target activity can be increased, thereby improving the user experience.

The present application relates to a content processing method, device, electronic device and storage medium, including: obtaining relevant data of multiple contents to be selected; displaying multiple first controls in a content selection page according to a pre-set first animation effect, wherein each of the first controls is used to display relevant data of one content to be selected, and the relevant data of the contents to be selected displayed by any two of the first controls are different; receiving a first input of a user to the first control, and obtaining a target first control from the multiple first controls according to the first input; and displaying a pre-set second animation effect when the content corresponding to the target first control is successfully selected.

The application discloses a semantic perception routing scheduling method and system for computing power network, and belongs to the technical field of computing power network and communication. The application aims at the problems that the traditional host-centered routing scheduling mechanism cannot perceive service semantics and computing power state, resource supply and demand are not matched, and multi-computing power instance optimization capability is insufficient, and proposes an OSAF-SORS paradigm: by introducing a service tag (STag) decoupled from location and home as a new type of addressing routing identifier, an open service available priority (OSAF) routing scheduling mechanism is realized; based on deep reinforcement learning and multi-layer graph attention network, a service-oriented routing scheduling (SORS) algorithm is designed to jointly capture resource dependency and queue dynamics, complete multi-computing power instance optimization and path collaborative optimization. The application can significantly improve the service level agreement (SLA) achievement rate, resource utilization rate and load balancing effect, and is suitable for wide-area high-performance computing, immersive video, distributed AI training and other low-latency high-reliability business scenarios.

The application discloses a kind of customized training plan generation methods for operation and maintenance personnel based on performance prism, which comprises the following steps: collecting multi-source heterogeneous demand data of operation and maintenance personnel on multiple platforms, and combining Carnot model and importance performance analysis model for fusion processing, extracting key technology theme through cluster analysis, drawing core training intention and decomposing into key measurement indicators, collecting employee daily operation and maintenance behavior data and calculating ability membership degree vector, combining post demand information, calculating target ability information to be enhanced, generating training content using intelligent workflow engine according to target ability information and key measurement indicators, generating hybrid training plan combining online and offline according to employee historical schedule, issuing the plan to terminal and setting task reminders, which can ensure that training is executed on time, improve training coverage and completion rate, and realize the intelligentization, customization and high efficiency of operation and maintenance personnel training from demand analysis, capability assessment, content generation to plan execution.

The present invention relates to a method, device and storage medium for evaluating the thermal power of a geographical area based on supply and demand relationships, and is applied to the field of online car-hailing technology, including: obtaining online car-hailing order data and driver operation service trajectory data for a period of history, and dividing the time dimension at the same time, extracting the driver order income amount, driver order service time and driver order listening time in each time dimension, obtaining the historical IPH under different time dimensions, and finally fusing the historical IPH and real-time IPH under the same time dimension according to the set weight; the present application combines the history of fine-grained time intervals, and according to the appropriate weight distribution, it can take into account the historical periodic laws of different spatial regions and different time periods, and can also consider the current real-time sudden changes, thereby solving the deficiency of the single dimension of traditional thermal calculation schemes; at the same time, the calculation of thermal level based on IPH index is guided by driver income, which is more in line with driver wishes and service goals.

The embodiments of the present application disclose a virtual item management method, device, storage medium and electronic device. The above method includes: obtaining an account, the above account is used to receive a first virtual item, the above first virtual item is determined based on the execution of the target task by the above account in each second time interval within the first time interval, and the above target task is released by the task release platform; for each second time interval in at least two second time intervals, determine the progress feedback information of the above target task in the above second time interval, and send the first virtual item to the above account according to the above progress feedback information; based on the sent first virtual item, request the second virtual item in the first time interval from the above task release platform, the virtual value of the above second virtual item is greater than the sum of the above first virtual items sent in each of the above second time intervals, and the above first time interval includes the above at least two second time intervals. The present application shortens the virtual item collection cycle.

The utility model relates to the technical field of foam processing, and discloses a cleaning device for silica gel foam processing, which comprises a cabinet body, a motor is arranged on one side in the cabinet body, rotary columns which are symmetrical up and down are rotatably connected in the cabinet body, and the output end of the motor is fixedly connected to one end of the rotary column on the lower side. A driving wheel is fixedly connected to the other end of the rotating column on the lower side, cleaning rollers are fixedly connected to the outer walls of the rotating columns, cleaning brushes distributed in an annular array are fixedly connected to the outer walls of the cleaning rollers, a driven wheel is fixedly connected to one end of the rotating column on the upper side, and a belt is arranged between the driven wheel and the driving wheel; the top of the cabinet body is fixedly connected with a water tank. According to the cleaning device for the silica gel foam, the motor, the cleaning brush, the conveying assembly and other components are matched, so that the effect of cleaning the silica gel foam more thoroughly is achieved, the problem that part of dirt remains when the silica gel foam is washed only through cleaning water is solved, and then the product output quality is improved.

The invention discloses an automatic movie short video explanation editing method and device, a server and a storage medium, and relates to the field of multimedia video editing, and the method comprises the steps: analyzing each plot content of an original movie video; performing shot fragmentation on the original movie video, extracting shot information, and marking starting and ending timestamps of each shot; matching the preset commentary with the shot information in the film; based on a text-to-speech model, converting the matched commentary into a corresponding commentary audio, synchronizing the corresponding commentary audio with a corresponding movie shot, and providing background commentary for the short video; and dynamically planning the rhythm of lens switching based on the identified plot content, the lens matching information and the synchronized explanation audio, and intelligently inserting a transition special effect to assemble the mixed video. According to the invention, by fusing multi-modal feature analysis and intelligent editing logic generation, efficient and high-quality video automatic production is realized, and the manufacturing cost is reduced while the content attraction is improved.

The anti-deviation shell pressing device comprises a bottom plate, a supporting column and a top plate, cushion blocks are symmetrically arranged on the upper end face of the bottom plate, first air cylinders are fixedly connected to the upper ends of the two cushion blocks, and mounting bases are arranged on the upper end faces of the two cushion blocks; piston rods of the two first air cylinders penetrate through the inner walls of the two mounting bases and are fixedly connected with sliding rods, sliding blocks are arranged at the two ends of each sliding rod, first rotating shafts are symmetrically arranged on the outer walls of the sliding blocks, and one first rotating shaft is rotationally connected with a first connecting rod in a sleeving mode. The four clamping plates are arranged on the floor of the pressing device, the shell is fixed at the same time, the stability of the shell on the workbench is guaranteed, position deviation caused by materials or other reasons is eliminated, and the product completion rate is increased.

This invention relates to the field of pulse capacitor charging and discharging technology, specifically to an automated testing method for pulse capacitors that ensures the effective discharge count meets the target. The method includes: S1, charging the pulse capacitor and inputting a trigger pulse to the trigger electrode of the cold cathode discharge tube; S2, determining whether the trigger pulse causes the cold cathode discharge tube to break down to ground. If so, the effective discharge count is incremented by one, and a valid trigger event is recorded. After the pulse capacitor completes its discharge, the process returns to step S1 until the effective discharge count reaches the preset target number; otherwise, an invalid trigger event is recorded, and the pulse capacitor undergoes phased safe discharge; S3, statistical characteristics of the invalid trigger events are extracted based on a sliding time window to determine if the cold cathode discharge tube is faulty. If not, the process returns to step S1 for supplementary testing; otherwise, the test ends. Through a closed-loop control process of invalid trigger statistics, safe discharge, and adaptive supplementary testing, the problem of reduced test intensity caused by occasional failures of switching devices is solved.

The application provides a seat self-learning method and a vehicle, comprising: decomposing a seat self-learning task into multiple sub-tasks, and sequentially executing each sub-task, each sub-task corresponding to controlling a seat motor to move to a limit position direction to complete limit learning of the limit position direction; when detecting an interruption event during execution of each sub-task, controlling the seat motor to stop moving, and saving current learning progress information as a learning snapshot to a preset storage space; in response to a recovery event for the interruption event, reading the learning snapshot from the preset storage space, and continuing to execute an unfinished sub-task according to learning progress information recorded in the learning snapshot to complete the seat self-learning task. Thus, the problem of "progress loss after interruption and starting from the beginning" in the prior art is avoided, and the completion rate and efficiency of self-learning are improved.

This invention discloses a method and system for the dynamic allocation of annual air pollutant targets at each level, relating to the field of air pollution control technology. The method includes: acquiring actual data on the number of days that have passed and calculating the completion rate of three assessment indicators; calculating the daily average target concentration for the remaining period based on the total annual target and the total amount already incurred, and introducing a dynamic safety margin coefficient for correction; allocating the remaining cumulative PM2.5 concentration daily using a time decay weight, and decomposing the remaining days of good air quality monthly using a historical weight; calculating the probability of completion and outputting tiered early warnings based on historical best and historical average data; updating all parameters at the end of each day and re-executing the above steps. This invention can reflect the progress of pollutant control in real time, dynamically adjust subsequent targets, and improve the controllability and adaptability of achieving annual targets.

The invention belongs to the technical field of autism screening, and discloses a rapid interactive screening test method for infant autism, which is suitable for Chinese environment infants of 18-36 months old. The method comprises the following steps: determining an adaptive test object, preparing a test, executing an interactive test and scoring according to a sequence that the first three items are fixed and the other items are flexible, adding scores of the nine items, and judging the risk according to the total score. The method can be completed within 5-10 minutes and is adaptive to Chinese culture, equipment is simple, operation is standardized, and the problems that an existing tool is poor in adaptation and low in efficiency are solved; the method is high in accuracy, is suitable for multiple scenes such as communities, can quickly recognize autism risks, and provides support for early intervention of child autism.

The invention discloses a computer resource allocation management system and method based on data analysis, and belongs to the technical field of resource allocation. The invention analyzes task data, extracts task features, and classifies computer execution tasks; calculates the amount of resources required for each task to be executed at the fastest speed; calculates the total amount of resources required for all tasks being executed, and determines whether the computer can meet the total amount of resources required for all tasks being executed; collects the time data of users executing each task in the computer in history, calculates the time list of users using the computer, calculates the priority of all tasks being executed, obtains the priority of each task being executed, and performs dynamic resource allocation; extracts task error data, calculates the risk threshold, and when it is determined that there is an error risk in the real-time execution of the task, ends the abnormal task process to release computer resources and reallocates resources.

This invention discloses a cross-platform heterogeneous unmanned equipment integrated command system, including a platform side and an equipment side, which communicate with each other through an interconnected patrol operation network. The platform side is equipped with a platform-side command module and a cloud-based intelligent planning module. The equipment side is equipped with a unified robot access node, which is implemented in the form of a software package. The platform side also has a controller and listener mechanism, which achieves closed-loop cooperation through a unified state space field and a unified event semantics. This invention solves the problems mentioned in the background art, such as the difficulty in unified scheduling of collaborative tasks, control source switching conflicts, and discontinuous control and feedback under link changes, by cooperating the platform-side command module, the unified robot access node on the equipment side, and the cloud-based intelligent planning module. It achieves unified scheduling and collaboration of tasks without relying on vendor-specific closed-loop systems.

The invention, which relates to the technical field of unmanned aerial vehicle inspection, discloses a photovoltaic field area unmanned aerial vehicle inspection positioning system comprising an unmanned aerial vehicle body, an environment sensing module, a self-adaptive flight control module, a positioning module and a data acquisition and transmission module. Through the advanced environment perception sensor and the self-adaptive flight control module, the environment condition can be accurately detected, the self-adaptive flight control module adjusts flight parameters in real time according to environment changes, damage to components of the unmanned aerial vehicle due to overheating is avoided at the extremely high temperature, normal operation of key equipment is guaranteed at the low temperature, and the safety of the unmanned aerial vehicle is guaranteed. Flight attitude and parameters can be adjusted in strong wind, stable flight is ensured not to deviate from a route, damage of sand and dust to equipment is prevented in a sand and dust environment, and a flight path is optimized, so that the unmanned aerial vehicle keeps flight stability in severe environments such as high temperature, low temperature, strong wind, sand and dust and the like; flight accidents and inspection interruption caused by severe environments are greatly reduced, and the completion rate of inspection tasks is improved.

The invention discloses a mixed request-oriented large language model adaptive scheduling algorithm. The algorithm comprises the following steps: S1, initializing queues Q1, Q2 and Q3 and a set CB to be empty; s2, when a new request arrives, calculating the priority of the new request, adding Q1 or Q2 according to the type of the new request, and then entering S3; when no new request arrives, entering S3; s3, judging whether the request is true or not, if so, entering S4, otherwise, terminating the algorithm, and starting when a new request exists; s4, updating the priorities of the requests in the queues Q1, Q2 and Q3; s5, the updated Q1, Q2 and Q3 are traversed in sequence, and the request is selected to be added into the CB; s6, inputting the CB into the large language model for reasoning for one time, removing the completed request from the CB, and recording the completion time; s7, when the CB is empty, entering the step S2; when the CB is not empty, k is made to be equal to k + 1, and when k is smaller than Y-1, the step S6 is executed, and otherwise, the step S8 is executed; and S8, for uncompleted requests in the CB, the request from the Q1 is put back to the Q1, the requests from the Q2 and the Q3 are both put back to the Q3, and then the step S2 is returned.

The application provides an AI conversation-driven asset due diligence process arrangement method and system, relates to the technical field of artificial intelligence, and comprises the following steps: analyzing user conversation, constructing an asset knowledge graph and demand constraints, screening and arranging atomized due diligence components to form an execution graph according to the knowledge graph and the demand constraints, and performing information collection and analysis on the components in sequence; when data is missing, the method automatically backtracks the conversation nodes and generates follow-up questions to supplement the information; and finally, the aggregated results are used to generate a due diligence report. The method realizes the automation and self-adaptation of the due diligence process, and improves the efficiency and integrity.

The invention discloses a large-cluster unmanned aerial vehicle management system, relates to the technical field of unmanned aerial vehicle management, and solves the problem that although a large-scale unmanned aerial vehicle management system in the prior art can realize cooperative control of multiple unmanned aerial vehicles to a certain extent, the large-scale unmanned aerial vehicle management system still has obvious defects in the aspect of energy consumption management. According to the method, the charging station most suitable for the current electric quantity condition is dynamically selected by evaluating the relationship between the remaining electric quantity of the unmanned aerial vehicle and the return voyage consumption, so that the unmanned aerial vehicle can find a safe and reasonable charging point even if the electric quantity is insufficient, the electric quantity is not used up due to forced return voyage, and the charging efficiency is improved. And in the task process, the battery use condition of the unmanned aerial vehicle is dynamically allocated, the charging sequence and strategy are reasonably arranged, and the system automatically selects the optimal energy supplement scheme when the electric quantity of the unmanned aerial vehicle is insufficient, so that task failure or interruption caused by the electric quantity problem is avoided.

This invention belongs to the field of wireless communication technology, specifically relating to an integrated air-space-ground intelligent communication monitoring system for areas without public network signals. The system includes: ground-based sensing nodes deployed in specific areas for data acquisition and preprocessing; an airborne mobile relay platform mounted on an unmanned aerial vehicle (UAV) for monitoring, data aggregation, and wireless power replenishment; a space-based constellation backbone network composed of low-Earth orbit (LEO) satellites for constructing backhaul links; and a remote integrated control terminal. The airborne platform dynamically plans its path and performs wireless beam replenishment based on parameters such as the power levels of the ground nodes. When the airborne platform's power is insufficient, it can switch to a fixed base station to enhance coverage. This invention improves the system's lifespan and monitoring flexibility through three-layer air-space-ground collaboration and an on-demand wake-up power replenishment mechanism, achieving reliable backhaul of high-value information and flexible resource reuse in complex environments.

The invention relates to the technical field of short videos, and discloses a method and a system for quickly slicing short videos in batches with a high complete playing rate, and the method comprises the steps: obtaining hotspot keywords of an Internet platform according to a fixed time interval in a first preset time period, and obtaining hotspot videos corresponding to the hotspot keywords; constructing a video material library, wherein the video material library comprises a text unit, a dubbing unit and a dubbing unit; the text unit is used for generating texts for the hot videos, the dubbing unit is used for generating dubbles for the hot videos, and the dubbing unit is used for generating dubbles according to the generated texts; based on the hotspot video, constructing a random population and an optimization population in sequence, and obtaining a target individual according to the optimization population; a plurality of target individuals and corresponding target function values are obtained, the target individuals and the corresponding hotspot keywords are combined to serve as training samples, the corresponding target function values serve as sample labels, and an individual prediction model is obtained through training; and in a second preset time period, obtaining the hotspot keyword.

This invention discloses an intelligent CTD sampling control method and system combining tidal current prediction, relating to the field of marine observation technology. The method includes: generating tidal current predictions by fusing historical tidal elements and numerical model outputs, and planning candidate sampling windows; real-time acquisition of multi-source observation data to form standardized observation vectors, calculating the deviation from the predicted values ​​and fusing the vertical shear force intensity factor to perform rolling corrections on the predictions and update the sampling windows; comparing real-time observation data with safety thresholds to generate operational access instructions, issuing control instructions under permissible conditions, and dynamically adjusting sampling parameters and triggering resampling based on profile data quality coefficients; synchronously monitoring abnormal events, triggering emergency recovery, and deferring the task to subsequent candidate windows. By constructing intelligent closed-loop control, introducing vertical shear force early warning, and a parallel processing architecture, adaptive, highly safe, and highly reliable CTD sampling operations under complex sea conditions are achieved.

The utility model provides an automobile seat framework assembling device which is characterized in that the automobile seat framework assembling device comprises a supporting frame, a workbench is fixed to the supporting frame, an assembling tool for achieving framework assembling work is arranged on the workbench, and the assembling tool is provided with a detection assembly for receiving and judging neglected assembly detection of parts; the supporting frame is connected with a torque gun for installing the parts, and the torque gun is provided with a sensor for obtaining torque data. Efficient and accurate assembly, intelligent detection and early warning, data tracing, energy conservation and environmental protection of the automobile seat framework can be achieved, and the product quality and production benefits are improved.