17 results about "Online algorithm" patented technology

The invention belongs to the technical field of petroleum and natural gas engineering, and particularly relates to a priori knowledge embedded unsupervised learning geology-engineering dessert comprehensive evaluation method, which comprises the following steps: S1, obtaining an evaluation data set of a research area, and preprocessing and standardizing the evaluation data set; s2, dividing the evaluation data set into a positive index and a negative index according to the prior knowledge, constructing positive and negative ideal points through the prior knowledge, and fusing to generate an unsupervised evaluation data set; s3, clustering is carried out on the unsupervised evaluation data set by adopting principal component analysis and an incremental online k-means algorithm; and S4, according to category labels obtained by clustering, determining geological-engineering dessert grades by adopting a standardized variable distance grading evaluation method, and carrying out segmented clustering optimization and perforation position optimization on the geological-engineering dessert grades. According to the method, a supervised evaluation effect is achieved through an unsupervised algorithm, comprehensive evaluation of the dessert of the fractured horizontal well section is achieved, and the problem of unlabeled data evaluation is effectively solved.

The application discloses a cloud measurement and control arrangement architecture, method and system based on micro services, belongs to the field of combination of spaceflight measurement and control and cloud computing, and comprises: a service arrangement architecture after cloudization of a measurement and control baseband, which adopts micro services and a layered thought, performs hierarchical and service division of the arrangement architecture, and realizes high availability, high cohesion, low coupling and scalability of the cloud baseband arrangement architecture; a customizable arrangement rule model, which realizes self-arrangement and scalability of rules, and improves richness of the micro service arrangement architecture; and open access capability of a scheduling algorithm service, which can randomly access diversified scheduling services online, forms a dynamically scalable online algorithm library, provides flexible, balanced and optimized deployment for measurement and control application arrangement, and enhances scheduling capability of the micro service arrangement architecture. The application can comprehensively realize service, diversification and open arrangement of measurement and control resources.

The present application relates to a kind of airplane simulation model confidence optimization method, for airplane simulation model confidence is not high and airplane online data transmission delay problem, proposed the method for improving simulation model confidence and solving data delay problem in combination with simulation system, airplane transmission data.The characteristic deficiency phenomenon of simulation output data in real data exists in offline network correction, online network can be combined with delayed transmission real data to the secondary correction of offline correction result, so the method involved in the present application combines two parts of offline correction network and online correction network.Random forest algorithm and MDA method are used to extract the features of offline network and online network.The simulation data error is preliminarily corrected by bidirectional long short-term memory neural network (BiLSTM) network and time attention mechanism in offline network;And the error of offline correction result and real data is secondarily corrected by online algorithm FTRL.

This invention discloses an online algorithm development method, system, and device for dynamic management of the runtime environment. Addressing the problem that current algorithm development platforms often have pre-installed algorithm components, making it difficult to expand and flexibly support rapidly changing business scenarios, this invention establishes a basic algorithm development environment, constructs algorithm components, uploads the required dependency files for the algorithm components, and obtains an environment library dependency list. It then creates an algorithm runtime environment container, selects programming language versions and corresponding dependency files from the environment library dependency list, and establishes an extended algorithm runtime environment. The extended algorithm is coded, debugged, and deployed online using an algorithm editor. During task modeling, multiple algorithm components are selected, and the training and runtime environments of the algorithm components are dynamically managed. This provides an environment expansion scheme, generating extended algorithm environments based on selected dependency packages to achieve algorithm environment expansion.

The invention discloses a drilling engineering design on-line cooperation method and system. The method comprises the steps of collecting and storing a formulated template, geological data analysis, algorithm packaging, well bore structure on-line design, tubular column checking, drilling tool combination design, drilling parameter design and report automatic generation. The system is integrated with an application layer, a service layer, an engine layer and a data management layer through hierarchical design and modular construction. The method has the technical effects and advantages that (1) the template is edited online, and different format requirements are supported; (2) the algorithm is online and can be directly used without depending on external software; (3) data structuring: all data are formatted automatically in the design process, and extra work is not needed; and (4) multiple persons are assisted, and each person can complete one of the modules.

The application discloses an internet harmful event detection method, device and storage medium, constructs or updates a keyword knowledge graph, divides the knowledge graph into a plurality of subgraphs, clusters documents into harmful events by using a clustering algorithm, inserts each harmful event into a harmful story tree by updating each harmful event, or creates a new harmful story tree according to the harmful event, realizes harmful content detection and classification in mobile internet and internet, and the purpose is to find harmful events from massive webpage and document data, organizes harmful events in a reasonable harmful story tree in an online mode, proposes a two-layer webpage document clustering algorithm based on a knowledge graph, extracts harmful events from a large amount of webpage text or webpage related news, and organizes events into a story tree by using an online algorithm after new webpage and document data arrive, and is more effective than a traditional keyword-based harmful algorithm in the aspect of harmful event extraction.

The invention discloses a method and system for correcting the topological relation between a distribution network area transformer and a line, and relates to the technical field of distribution network transformer topology correction, and the method comprises the steps: obtaining a to-be-corrected area transformer in a target area, and injecting a target current signal with a preset frequency and amplitude at a preset side of the area transformer; based on a feeder terminal unit, acquiring first current signals of a plurality of section switches of the target area; extracting a high-frequency component amplitude of the first current signal, and obtaining a first signal amplitude of the section switch based on the high-frequency component amplitude; if the first signal amplitude is greater than a preset threshold value, obtaining a target line of the section switch; and correcting the topological graph of the target area based on the target line, thereby solving the problems of high cost caused by the fact that a monitoring device needs to be additionally arranged on a transformer node in a current correction method for the topological relation between the power distribution network line and the transformer area, and complexity and difficulty in compatibility caused by the fact that different threshold values and parameters need to be set in an online algorithm.

The present application relates to the technical field of administration, in particular to a security personnel intelligent scheduling system and method based on hierarchical architecture. The system adopts five-layer architecture of data management layer, feature calculation layer, intelligent core layer, decision support layer and user interaction layer to realize business logic decoupling. The method core includes: converting requirements into structured portraits through a project analyzer; the intelligent scheduling engine adopts a multi-objective evolutionary algorithm integrating robust optimization, generates a resilient scheduling scheme capable of resisting the worst absence scenario by defining an uncertain set of personnel availability; the dynamic scheduler adopts an online algorithm with a guaranteed competitive ratio to provide real-time decisions with bounded performance for unpredictable temporary tasks. The scheme realizes the change from post-event remediation to pre-event prevention, guarantees business continuity, and provides quantifiable performance guarantees for scheduling decisions in dynamic environments.

The invention relates to the technical field of administrative management, in particular to a security personnel intelligent scheduling system and method based on a layered architecture. The system adopts a five-layer architecture of a data management layer, a feature calculation layer, an intelligent core layer, a decision support layer and a user interaction layer to realize business logic decoupling. The core of the method comprises the following steps: converting a demand into a structured portrait through a project analyzer; the intelligent scheduling engine adopts a multi-objective evolutionary algorithm fused with robust optimization, and generates a flexible scheduling scheme capable of resisting the worst absence scene by defining an uncertain set of personnel availability; and the dynamic scheduler adopts an online algorithm with competition ratio guarantee to provide real-time decisions with bounded performance for unpredictable temporary tasks. According to the scheme, shift arrangement is changed from post-event remedy to pre-event prevention, and quantifiable performance guarantee is provided for scheduling decisions in a dynamic environment while service continuity is guaranteed.

The application provides a multi-cloud storage processing method, device and equipment and storage medium, and belongs to the technical field of multi-cloud storage. The method comprises the following steps: obtaining target data to be stored, and dividing the target data into multiple data segments; performing a multi-cloud storage segmented online algorithm strategy to determine target storage nodes of the multiple data segments corresponding to the target data, including performing an allocation strategy and performing an allocation adjustment strategy; wherein the allocation strategy is used to determine the target storage nodes of the multiple data segments corresponding to the target data; the allocation adjustment strategy is used to adjust the target storage nodes of the multiple data segments of the target data according to a storage balance point when there is any access request, and the storage balance point is used to represent a balance relationship between the number of access requests and the storage cost of the target data. The application aims to reduce the storage cost.

The invention provides a method for realizing flexible deployment of a target detection and tracking algorithm and a computer program product, which construct a full-link automatic process from data annotation, model training to reasoning deployment, realize rapid construction and one-button deployment of target detection, tracking and time sequence prediction algorithms, take a data warehouse as a unified entry of data, and realize flexible deployment of the target detection and tracking algorithm. The data set can be automatically annotated to obtain annotated data capable of being used for model training, and versioning control is implemented; the task management module can create a training task and implement trajectory tracking and time sequence prediction algorithm training through a selected data version and algorithm, the online algorithm is packaged layer by layer when the training task is completed, and finally, the obtained multi-layer detachable algorithm package can provide inference service for any application through online deployment at an inference end. The problems that a traditional model is long in development and deployment period and high in integration complexity are solved, and therefore flexible deployment of the target detection tracking algorithm is achieved for any application and requirements thereof.

The application discloses a kind of distributed online optimization methods based on differential privacy mechanism, it is related to multi-agent collaborative control technical field, builds distributed online optimization problem mathematical model;According to the difference between the cost generated by the current decision of all nodes and the cost generated by the after-optimal fixed decision, the performance evaluation index of online optimization method is constructed;Differential privacy mechanism and bandit feedback mechanism are introduced to design distributed online optimization algorithm, and the distributed optimization model is iteratively solved according to the designed distributed online algorithm. Using bandit feedback mechanism to solve a distributed online problem with time-varying cost function, the real gradient information is estimated, so as to guide the update of decision variable, so that it is applicable to the application of expensive or impossible gradient calculation;Differential privacy mechanism makes the information exchanged between nodes be perturbed by random noise conforming to Laplace distribution, effectively solving the privacy problem in distributed online optimization.

The invention relates to the technical field of server-free computing resource scheduling, and provides a server-free computing resource scheduling method and system based on an online algorithm, and the method comprises the steps: recording whether each job experiences cold start time or not when passing through different functions in a DAG application operation process, and marking the jobs as qualified jobs and unqualified jobs; setting a time window for the DAG application; at the end of each time window, obtaining the use condition and the concurrency degree of the quantitative instances, and calculating an observation value in combination with the marking condition of the completed operation; and comparing the observed value with a performance index target value, and if the observed value is smaller than the performance index target value and a certain function uses a quantitative instance during operation processing, starting a new reserved instance for the function at the end of the time window. The operation is prevented from experiencing cold start time in the processing process, and the scheduling timeliness and reliability are very high.

The invention discloses a bilateral matching market online learning method based on relative victory and defeat feedback. The method comprises the following steps: initializing a preference sequence position, preference probability estimation and a confidence radius of a requester; the requester initiates an application based on the position which is not definitely rejected; the provider randomly selects a receiver and outputs a relative victory or defeat feedback; the requester updates preference probability estimation and confidence radius according to the preference probability estimation and confidence radius, and constructs a preference learning mechanism; and based on the updated data and the preference probability difference lower limit, judging whether the most preferred requester of the provider is clear or not, if so, matching, otherwise, updating the position and repeating the process until all the requester are matched. The method depends on relative victory and defeat feedback, efficient preference learning and stable matching of the requesters are achieved through the online Gal-Shapley algorithm, multiple requesters are supported to initiate applications in parallel, the learning efficiency is defined by introducing a relative preference minimum gap, and the method is suitable for a limited feedback scene.

The present invention specifically relates to a motion recognition method and system based on multi-device collaboration with integrated sensory and communication capabilities, a storage medium, a computer program product, and an electronic device. This method proposes a deep reinforcement learning online algorithm that dynamically allocates sensory and communication resources to adapt to time-varying sensory environments and network conditions. This algorithm optimizes the allocation of sensory and communication resources across multiple devices over multiple time periods, thereby maximizing the accuracy of human motion recognition.

The application discloses an algorithm correction method, device, system and storage medium, which are used for optimizing a product quality online detection algorithm and improving the accuracy of online detection. The method comprises the following steps: in the process of detecting the product quality value according to a preset algorithm, acquiring a first offline value of a key impurity component of the product according to a first preset time interval, and acquiring a second offline value of all impurity components of the product according to a second preset time interval; correcting a first coefficient in the preset algorithm according to the first offline value; and correcting a second coefficient in the preset algorithm according to the second offline value. By using the scheme provided in the application, the coefficients of the online algorithm are corrected by more accurate offline values, the algorithm is optimized, and the accuracy of online detection is improved.

The application provides an online scheduling method considering periodic maintenance of a batch processing machine, taking the batch processing machine as a research object, considering the batch processing machine in the periodic maintenance process, and solving the actual problems of online scheduling of orders arriving over time, batch grouping, processing and the like. For various common scheduling objectives, such as maximum completion time of a workpiece, maximum flow time of a workpiece in a production system and the like, an effective online algorithm is designed through a waiting strategy to generate a real-time scheduling scheme, and the best value that can be reached by the algorithm performance under various objectives is analyzed, so that the online scheduling rules and methods for solving the online scheduling problem of the batch processing machine under the periodic maintenance are provided, guidance is provided for production planning of an enterprise, and a referable technical means is provided for other batch processing scheduling problems of the enterprise.