36results about How to "Efficient decision-making" patented technology

A repair system for flat panel display (FPD) substrates performs a number of operations, such as automatic image capture and processing, automatic defect classification, automatic repair classification, and repair macro (instruction) generation software. Defect classification, repair classification, and repair macro generation are based on an open architecture and can address any number of use-cases through the use of multi-tiered classifiers, and thus a wide variety of panel designs may be repaired within a single repair tool. The multi-tiered set of classifiers, e.g., defect classifier, repair classifier, enables an efficient decision-making repair process with capability for customization. The multi-tiered classifiers are optionally extended to support statistical learning (both online & batch) and active learning, in the context of a supporting database of defects and associated tools. The classifiers and recipe rules may be automatically or semi-automatically improved over time, incorporating lessons learned by operators while judging defect or repair needs.

The invention discloses a multi-unmanned aerial vehicle task decision method based on MADDPG, which introduces an MADDPG algorithm into multi-unmanned aerial vehicle task allocation, and comprises thesteps of firstly establishing a two-dimensional combat environment model required by deep reinforcement learning according to the actual combat environment of multiple unmanned aerial vehicles, secondly establishing mathematical description of various threats such as air defense missiles in the combat environment of the multiple unmanned aerial vehicles, and finally, taking the tracks and distances of the multiple unmanned aerial vehicles and the defense threats of the battlefield as constraint conditions, and performing learning training to obtain a multi-unmanned aerial vehicle task decision model. According to the method, an experience pool and a double-network structure are adopted, so that the operation and convergence speed of the whole network is greatly improved, a result can be obtained more quickly in the high-speed flight process of the unmanned aerial vehicles, the purpose of autonomous decision making of the multiple unmanned aerial vehicles can be achieved, and the highefficiency of task decision making of the unmanned aerial vehicles can also be guaranteed in an unknown combat environment.

The invention discloses a preventative device maintenance method based on dynamic reliability and belongs to the technical field of petrochemical equipment management. The preventative device maintenance method based on the dynamic reliability comprises the following steps of 1 establishing a device basic database including equipment accounts, historical maintenance and repair data, a fault strength analysis data, operation monitoring data and the like; 2 utilizing basic data to conduct operation time analysis, fault strength analysis and alarm state and operation trend analysis on devices; 3 utilizing results to establish a device reliability level and fault strength incidence matrix, formulate DRBPM data analysis logic, automatically screen preventative maintenance devices and generate preventative maintenance planning reports of the devices; 4 examining, verifying and executing the preventative maintenance planning reports of the devices. The preventative device maintenance method based on the dynamic reliability can be used for dynamic analysis and management of states of petrochemical equipment, facilitate formulation of scientific and effective device maintenance measures and provide guarantee for timely device fault elimination and production continuity reliability and can be widely applied to the technical field of petrochemical equipment management.

The invention discloses an answer aggregation method and system facing crowdsourcing scene question-answering system. The method comprises the steps of obtaining multiple reply texts with respect to a specified question, wherein each reply text comprises a pro vote and a con vote of other users; conducting text analysis of a theme and key words on each reply text, and further creating a vector set; clustering all vector sets, combining all similar replies into a sort, and calculating the pro votes and the con votes of each sort of similar replies; supplying a questioner with the replies with a preset number of replies in the each sort of similar replies and a calculation result corresponding to the pro votes and the con votes. The answer aggregation method and system facing crowdsourcing scene question-answering system has the advantages that by supplying the questioner with sorts of the answers and the calculation result of the pro votes and the con votes corresponding to the sorts of the answers, redundant time for the questioner to read a lot of answers with similar meanings is saved, a more general level of support of public opinion is presented for the questioner, and the users can be helped to make own judgment and determination more efficiently.

The invention relates to a smart park management system and method based on a three-dimensional scene. The system comprises a data storage module and a data management module which are mutually connected, and a three-dimensional scene data acquisition module and an Internet of things service data acquisition module which are connected with the data storage module, wherein the three-dimensional scene data acquisition module is used for acquiring the three-dimensional scene data of the park, the Internet of things service data acquisition module is used for acquiring the Internet of things service data of the park, the data storage module is used for storing the three-dimensional scene data and the Internet of things service data, and the data management module is used for processing and displaying the three-dimensional scene data and the Internet of things service data, representing park scene pictures through taking the three-dimensional scene as a carrier and monitoring and managing the park. The system is advantaged in that the three-dimensional scene is taken as the carrier, various service data are fused in the three-dimensional space in multiple levels, each service and the park status of each park are represented in a more stereo mode, and management personnel is facilitated to carry out multi-level efficient monitoring, analysis and operation of the smart park.

The invention provides a bottleneck upstream lane level traffic state prediction system and an implementation method. The prediction system comprises an information acquisition module, a processing module, a release module and a transmission module; the speed and passing time of each vehicle are obtained via the information acquisition module, then the initial time-space distribution state and traffic of bottleneck upstream traffic flow are obtained, and the time-space distribution state of the traffic flow in different lanes at any moment is predicted in the information processing module based on a driver risk threshold in combination with a confluence model at the bottleneck and a lane selection model at the upstream of the bottleneck. The confluence model and the lane selection model involve in the influence of different advantages of different lanes at the upstream of the bottleneck at a confluence point, and the precision of bottleneck upstream traffic state prediction can be improved to the lane level, so that the accuracy of traffic state prediction can be greatly improved, theoretical and technical supports are provided for dynamic control of traffic managers, intelligent induction, efficient management and quick decision, and more accurate travel information is provided for road users.

The invention relates to a medical image sample screening method and device, computer equipment and a storage medium, for carrying out intelligent screening on unlabeled medical image samples. The medical image sample screening method includes the steps: carrying out model training on a labeled sample set by utilizing a Mask-RCNN model so as to obtain a focus target detection depth model; predicting the unlabeled medical image sample set according to the focus target detection depth model to obtain a prediction result of each medical image sample and judge a labeling value; and selecting a medical image sample with high annotation value to perform annotation confirmation, performing iterative updating on the focus target detection depth model, and ending the iterative updating until the performance of the focus target detection depth model cannot continue to annotate a new sample. According to the medical image sample screening method, under the condition of limited computing resourcesor annotation cost, the high-value small data set is actively mined and extracted, and efficient diagnosis and decision making are achieved by simulating a medical expert intelligent learning mode, and the intelligent degree is high, and the processing speed is high, and the problem that the annotation efficiency is low is effectively solved.

The invention discloses an intersection traffic signal lamp regulation and control method with a position sensing function. The method comprises the steps of: 1, carrying out mathematical modeling ona traffic signal control problem through employing a reinforcement learning network model, specifically, the modeling a traffic network into a graph mode , and defining a state space, an action spaceand an award according to traffic signal control problems; 2, preprocessing the original observation value oi of an intelligent agent; step 3, acquiring edge features with position perception among intelligent agents; step 4, realizing a Pos-Light message transmission model between the intelligent agents; step 5, realizing an intersection traffic signal lamp regulation and control decision of a Qnetwork; and step 6, carrying out regulation and control target training based on the Q network. Compared with the prior art, the method is more efficient in intelligent agent decision, has higher convergence speed, and can finally obtain a strategy which effectively relieves the traffic jam; and (2) the method is superior to a method for solving the traffic signal lamp control problem by using agraph neural network at present, has higher convergence speed and can obtain more stable results.

The invention discloses an improved PMADDPG multi-unmanned aerial vehicle task decision-making method based on transfer learning. The method comprises the steps of firstly, performing reasonable mathematical modeling for threats such as air defense missiles and radars in an environment under the background of a multi-unmanned aerial vehicle actual combat environment, then creating a plurality of different two-dimensional combat environment models, designing constraint conditions, and learning and training the plurality of combat environments in sequence, thereby obtaining a final multi-unmanned aerial vehicle task allocation model. According to the method, the defect that in the prior art, task decision making can only be carried out in a known or static combat environment is overcome, decision making can also be carried out efficiently in an unknown combat environment, tasks of unmanned aerial vehicles are achieved, and the viability of an unmanned aerial vehicle group in the unknowncombat environment is greatly guaranteed.

The invention relates to a method for managing government vehicles based on a mobile phone APP. The method includes the following steps: 1. installing a government vehicle management system APP on a mobile phone; 2. an applicant filling in information of an application, namely reasons for applying, time of use, and starting and terminating mileages, and then submitting the application for examination; 3. sending an alert if an examiner's mobile phone APP receives the information; 4. the examiner examining the applicant's application information on the APP; and 5. if the applicant's mobile phone APP receives approval information, allowing the applicant to use government vehicles. According to the invention, the method herein transforms conventional paper applications and examination to electronic applications and examination, so that users can get vehicles rapidly, vehicles are managed in a scientific manner and leaders are able to make efficient decision-making. The method herein alsoadopts a PGIS system, so that the vehicles are scheduled in a scientific manner and the information of each government vehicle can be known in time.

The invention discloses an improved order relation method based comprehensive evaluation method for voltage states of medium-voltage distribution lines. The method comprises the steps of firstly, proposing a voltage state characteristic index system of the 10kV distribution lines in three aspects of line attributes, charge characteristics and management factors, and calculating a centesimal system score of each index of each 10kV distribution line; secondly, determining a weight of each index by applying an improved order relation method; and finally, performing multiplication and summation calculation on a value of each index and the weight corresponding to the index to obtain a voltage state comprehensive evaluation value of each 10kV distribution line, and quantitatively evaluating the voltage states of the numerous varied 10kV distribution lines. The method has the advantages of comprehensive evaluation indexes and easiness for acquiring evaluation data, has relatively high operability, practicality and applicability, and can effectively improve the voltage control efficiency and management level of the 10kV distribution lines.

The invention discloses a user exploration type information exhibiting method based on a visualization technology. The user exploration type information exhibiting method comprises the following steps: 1, preprocessing collected data backstage and then visually displaying the processed data on a map interface, wherein the collected data comprises GIS (Geographic Information System) data and statistical data, and the statistical data comprises different dimensions of information related to the GIS data; 2, establishing a color lump view by a plurality of display areas according to the statistical data, and representing the different dimensions of information of the statistical data by using color and transparency of a color lump; 3, providing selection of an area-of-interest area for a user. The user exploration type information exhibiting method can be used for presenting various statistical data more intuitively and conveniently, and provides convenience for the user to observe the relationship among various areas and the comparison condition; in addition, prior domain knowledge of the user can be added to a whole model in the inquiry process, so that the knowledge of the user and machine intelligence are combined, the advantages are exerted, and the information exhibiting method is helpful for the user to make a decision more efficiently.

The invention discloses a method of hazardous waste emergency early warning and response through 3S technologies and a system. The method includes (1) a step of monitoring and early warning hazardous waste, (2) a step of emergency command of the hazardous waste, (3) a step of making decisions of the hazardous waste, and (4) a step of emergency site treatment of the hazardous waste, namely, detecting sites, collecting video information and carrying out final treatment on the hazardous waste pollution accidents. A monitoring and early warning system, an emergency command system, a decision support system and an on-site treatment system are connected with a communication network information platform to form a complete emergency response system. The 3S technologies are used in a hazardous waste pollution early warning and emergency response system, response and coping capability for hazardous waste emergencies are improved, requirements of environment management modernized development of China are conformed to, guarantees are provided for city public security, and the level of managing grave accident potential risk sources is improved.

The invention discloses a self-learning event extraction method and application thereof. The event extraction method comprises the steps that candidate words are subjected to self-learning screening through sentence vectors of event information and candidate word role mapping contained in each sentence vector, and a trigger word set is obtained; according to the invention, industrial production abnormal logs and other information can be quickly and accurately interpreted, and an appropriate decision is further made.

The invention relates to the technical field of urban geographic information processing, and discloses a road network density estimation method, a device, equipment and a storage medium. The method comprises the steps of collecting first road raster data according to the electronic map raster data; performing binarization processing on the first road raster data to obtain second road raster data;creating a road center line vector diagram layer, and extracting first road center line data according to the second road raster data and the road center line vector diagram layer; performing elementverification on the first road center line data to obtain second road center line data; and estimating the road network density according to the second road center line data. According to the road network density estimation method, apparatus and equipment, and the storage medium, the road center line vector data is extracted based on the electronic map raster data, so that quick estimation of theroad network density can be realized and the labor cost can be reduced.

The invention relates to the technical field of data management, and discloses a big data management system based on a smart shopping mall, and the system comprises a monitoring system which is used for obtaining target feature data; a member system which is used for generating member dynamic data; a transaction system which stores shopping mall transaction data; a data resource center which communicates with the monitoring system, the member system and the transaction system, collects target feature data, member dynamic data and transaction data, and integrates the data according to classification; a data analysis system which communicates with the data resource center and performs interactive statistical analysis on the integrated data; and a data display system communicates with the data analysis system to realize visual display of the data. By associating the data of different dimensions, the behaviors of shoppers are better measured, customer group portraits and consumption activity preferences are depicted, and the health condition of shop operation is monitored, so that rent pricing and overall business planning management are conveniently guided according to the operation condition of shops and customer requirements.

The invention discloses a multi-dimensional comprehensive display visual data display platform, which comprises an electric power data multi-dimensional expression module, and is characterized in that the electric power data multi-dimensional expression module is used for providing a plurality of electric power data expression methods; the power data multi-dimensional expression optimization module is used for optimizing a power data expression method by utilizing an algorithm; the power data multi-dimensional expression feature extraction module is used for extracting visual features of the optimized power data; and the power data multi-dimensional expression feature fusion module is used for fusing the visual features and the traditional features. According to the method, multi-dimensional comprehensive display of the electric power data is realized, high-dimensional data visual representation can be seen from a multi-dimensional comprehensive display result, electric power workers can quickly understand data information and capture hidden features, categories, relations, trends and the like in the data, more efficient decisions are generated, and the method is applied to actual production and life.

The invention provides a visual conversion system and method for real-time data image, belongs to the technical field of data conversion, and the system comprises a data collection module, a model calculation module, a model data correction module, a graph conversion module, and a display module. Two-dimensional and three-dimensional views are combined with virtual reality technology, and urban panoramas including urban blocks, landmark points, buildings, maneuvering targets, pipeline facilities and the like are completely and vividly presented; The data according to two dimensions of time andspace is synchronously presented and a data change trend is comprehensively mastered. real-time monitoring, historical playback and simulation deduction of spatial data are supported, so that rules are clear and visible, decisions can be made based on the rules, and the method is more efficient; a plurality of chart presentation modes is provided, and besides planar charts, chart superposition based on a three-dimensional space scene is also supported.

Disclosed is an ontology-driven mass data event decision-making method. According to the method, a supply chain event ontology base is established, a supply chain event ontology is taken as the upper ontology, the two-layer semantic mapping technology is adopted, uniform semantic translation from a local ontology to a target ontology is achieved, and finally a factbase is formed. SWRL language is adopted to define the processing rule of supply chain event decision-making so that an event decision-making rule base can be formed. Based on the factbase and the rule base, mass data-oriented fact distribution and rule matching are achieved by means of the distributed Rete algorithm, and then an optimum matching result is reasoned out, high-efficiency processing of mass data is achieved and event decision-making is assisted. The fact and rule combined supply chain distributed information processing mechanism is adopted, the problems including information heterogeneity and massiveness existing during manufacturing enterprise perceptual information processing are effectively solved, and the agility and accuracy of enterprise decision-making are improved.

The invention discloses a preventative device maintenance method based on dynamic reliability and belongs to the technical field of petrochemical equipment management. The preventative device maintenance method based on the dynamic reliability comprises the following steps of 1 establishing a device basic database including equipment accounts, historical maintenance and repair data, a fault strength analysis data, operation monitoring data and the like; 2 utilizing basic data to conduct operation time analysis, fault strength analysis and alarm state and operation trend analysis on devices; 3 utilizing results to establish a device reliability level and fault strength incidence matrix, formulate DRBPM data analysis logic, automatically screen preventative maintenance devices and generate preventative maintenance planning reports of the devices; 4 examining, verifying and executing the preventative maintenance planning reports of the devices. The preventative device maintenance method based on the dynamic reliability can be used for dynamic analysis and management of states of petrochemical equipment, facilitate formulation of scientific and effective device maintenance measures and provide guarantee for timely device fault elimination and production continuity reliability and can be widely applied to the technical field of petrochemical equipment management.

The embodiment of the invention discloses a material putting method and device and a related product, and at least relates to machine learning or natural language processing in artificial intelligence. The method comprises: obtaining user response data after material delivery; enabling the user response data and the delivery effect index to have an association relationship; enabling the material and the target material to have an association relationship; obtaining a predicted value of a delivery effect index of the target material in a future delivery time window according to the user response data; according to the predicted value of the delivery effect index, respectively simulating and executing various undetermined delivery actions on the target material, and obtaining a delivery target of the target material in a future delivery time window under the various undetermined delivery actions; and determining a target delivery action enabling the delivery target to meet the tuning requirement from the multiple undetermined delivery actions, and executing the target delivery action on the target material.

The invention provides a bottleneck upstream lane level traffic state prediction system and an implementation method. The prediction system comprises an information acquisition module, a processing module, a release module and a transmission module; the speed and passing time of each vehicle are obtained via the information acquisition module, then the initial time-space distribution state and traffic of bottleneck upstream traffic flow are obtained, and the time-space distribution state of the traffic flow in different lanes at any moment is predicted in the information processing module based on a driver risk threshold in combination with a confluence model at the bottleneck and a lane selection model at the upstream of the bottleneck. The confluence model and the lane selection model involve in the influence of different advantages of different lanes at the upstream of the bottleneck at a confluence point, and the precision of bottleneck upstream traffic state prediction can be improved to the lane level, so that the accuracy of traffic state prediction can be greatly improved, theoretical and technical supports are provided for dynamic control of traffic managers, intelligent induction, efficient management and quick decision, and more accurate travel information is provided for road users.

The invention discloses a panoramic electric energy quality monitoring system with a real-time monitoring function. The system comprises a monitoring system, a database server, an application server and a human-computer interface interaction workstation, wherein the monitoring system comprises an electric energy acquisition module, an electric energy quality storage module, a voltage monitoring module, an electric energy abnormity alarm module, a power supply service command module, a low voltage application module, a three-phase imbalance management module, a power distribution network time household budget type management module, a power distribution network on-load capacity regulation and voltage regulation suitability calculation module, a comprehensive maintenance evaluation application module, a work order application module and a geographic information positioning module. The invention belongs to the technical field of electric energy quality monitoring, and particularly provides a system for integrating marketing allocation system data, utilizing the concept of the electric electric energy internet of things and the big data technology and researching and developing a panoramic and visual interaction system, so that the service quality is integrally improved, the satisfaction degree of users is enhanced, the effect of precise treatment can be achieved, and the operation and maintenance efficiency is also remarkably improved.

The invention provides an order delivery method and system for multi-robots. The system includes terminal devices, order delivery servers and robots. The method includes: according to the order request message sent by the user and the location information of the robot, the order delivery server first uses the order sorting algorithm to sort the importance of the user's order, and then matches the order delivery robot for the user order according to the importance of the user order, and then The improved global path planning algorithm is used to plan the order delivery global path corresponding to the matching order for the order delivery robot, and then local collision avoidance optimization is performed on the planned order delivery global path, and then the order delivery optimization path is obtained, and then The order delivery optimization path is sent to the corresponding robot; the robot executes the delivery task for the corresponding order according to the order delivery optimization path. This technical solution can also make efficient decisions in complex environments, and then realize the use of multiple robots to quickly complete order delivery tasks.

The invention discloses a multi-UAV task decision-making method based on MADDPG. The MADDPG algorithm is introduced into the multi-UAV task assignment. First, according to the actual combat environment of the multi-UAV, a two-dimensional combat required for deep reinforcement learning is established. The environment model, secondly, establishes the mathematical description of various threats such as air defense missiles in the multi-UAV combat environment, and finally takes the trajectory, distance of the multi-UAV and the defense threat of the battlefield as constraints, conducts learning and training, and then obtains Multi-UAV mission decision model. This method adopts the experience pool and dual network structure, which greatly improves the operation and convergence speed of the entire network. In the process of high-speed flight of the UAV, the results can be obtained faster, and the purpose of autonomous decision-making of multiple UAVs can be realized. , it can also ensure the efficiency of its mission decision-making in the unknown combat environment.

The invention discloses a method and system for dynamically adjusting the frequency of a chain accounting block, and the method comprises the steps that a committee for dynamically adjusting the frequency of the block is selected to continuously collect and mine block information, and a continuously updated block information historical data set is formed; based on the historical data set, the committee predicts the increase and decrease trend of the transaction number; according to the transaction number prediction result, the committee determines the demand and amplitude of block frequency adjustment in advance; the committee broadcasts and publishes a frequency standard of a subsequent block, and a block frequency adjustment consensus is achieved in a whole chain range; the whole chain generates and accepts the block with the new frequency according to the new block frequency standard; and the committee evaluates the block frequency dynamic adjustment strategy and carries out necessary correction. According to the method, the frequency of the chain accounting block can be dynamically adjusted, the method has the advantages of being dynamic, flexible, efficient, good in expansion and the like, and the method has great significance in building a block flexible expansion accounting technology and solving the problem of cloud transaction payment accounting.

The invention belongs to the field of information safety, and particularly relates to a measurement data safety protection method, which comprises the following steps of: 1, user authentication: when a user accesses a data resource for the first time, firstly applying a safety certificate from a safety certificate authority so as to prove the identity of the user when the user accesses the data resource, safety assurance is carried out on malicious access which does not obtain the safety certificate by adopting a mode of denying to access data resources; and step 2: domain positioning: when a user passing identity authentication accesses data resources, an access request Req (SAVt, RAVm, EAVn, AAVk) is forwarded to a domain positioning server, the domain positioning server analyzes a subject attribute and a resource attribute, the subject attribute and the resource attribute respectively correspond to the SAVt and the RAVt, and whether local access or cross-domain access is determined according to an analysis result. According to the method, an attribute-based access control model is adopted, the problems of user authentication, domain positioning, access decision and module association are solved, the data safety in a big data environment can be effectively protected, and quick decision and efficient access can be realized.

The invention discloses a finite state machine-based intelligent vehicle driving behavior decision-making method and system. The method includes the steps of: 1) defining six types of driving behavior based on a priori rules: cruising, following a car, changing lanes and going out, accelerating and overtaking , changing lanes, returning, and braking; 2) Determine the judgment conditions for the current driving scene of the vehicle based on the current position, heading angle, driving speed, and driving environment of the vehicle; 3) Use event sets to describe the mapping relationship between the driving scene and driving behavior of the vehicle , use the finite state machine to judge the optimal driving behavior of the current vehicle. The invention realizes the effective decision-making of the driving behavior of the intelligent vehicle in the complex environment.