130 results about "Parallel search" patented technology

The invention relates to the technical field of computers, and provides a request management method and device, electronic equipment and a storage medium, and the method comprises the steps: when it is detected that a data access request is subjected to a cache miss, based on a first address carried by the data access request, executing the cache miss on the data access request; checking whether an uncompleted request entry corresponding to the first address exists in a preset request management structure or not, wherein the preset request management structure is a pre-constructed multi-path group associative cache structure used for performing grouping management on a cache miss request; and adding the data access request into a waiting queue of the uncompleted request entry under the condition that the uncompleted request entry exists in the preset request management structure. According to the method and the device, the cache miss requests are subjected to grouping management by adopting a multi-path group associative cache structure, so that the parallel search of the cache miss requests can be realized, and the processing efficiency of the cache miss requests is greatly improved.

The invention discloses an intelligent-based multi-task self-adaptive scheduling system for a pension accompanying robot, and belongs to the technical field of multi-task self-adaptive scheduling. Comprising a multi-modal task perception fusion module, a dynamic task evaluation module, a dynamic task priority generation module, a quantum causal reasoning module, a quantum optimization task scheduling module, a multi-task collaborative optimization module and a man-machine collaborative interaction module. Task features are mapped to a multi-dimensional weight space, a time attenuation coefficient is introduced to dynamically adjust the influence of historical causal association, a dynamic adjustment mechanism ensures timely execution of high-priority tasks, a quantum superposition state is utilized to parallelly search a global optimal solution through a quantum optimization task scheduling module, and the task scheduling efficiency is improved. Through a dynamic priority queue and a conflict solution feedback mechanism, the scheduling scheme is iteratively optimized, and the feasibility and the resource utilization rate of the scheduling scheme are remarkably improved.

The invention discloses a multi-robot path planning method based on multi-heterogeneous population coevolution and parallel search. The method comprises the following steps: firstly, modeling an environment by adopting a two-dimensional grid method, and generating a multi-robot initial path of a total population by utilizing a risk perception A * algorithm; then, dividing the total population into three sub-populations and corresponding external cooperative populations through a hierarchical roulette selection strategy; the three types of heterogeneous algorithms are executed in parallel, namely the collaborative genetic algorithm, the collaborative particle swarm algorithm and the collaborative ant colony algorithm are executed in parallel and synchronously, and a global solution set is generated in combination with a double-solution pool fusion mechanism. And then, a priority-based conflict resolution strategy is adopted, and time-space conflicts are solved by adopting time migration, path detour and local path planning strategies. And finally, path smoothing is realized through a cubic B-spline curve based on path reconstruction. According to the method, heterogeneous population coevolution is combined with parallel calculation, so that the efficiency, safety and cooperation capability of multi-robot path planning are remarkably improved.

The invention provides a big data standardization method and system based on data integration, and the method comprises the steps: converting original data of different data sources into intermediate format data through a dynamic format adapter based on a data conversion mapping rule in combination with a reinforcement learning Schema algorithm; generating a quantum superposition state of each field in the intermediate format data and a node in the knowledge graph by using quantum state coding; solving an optimal mapping node of the field in the knowledge graph based on a quantum superposition state by utilizing a quantum annealing algorithm; the multi-source data is integrated through the dynamic format adapter and analyzed into the intermediate format data, the problem of data heterogeneity is solved, multiple mapping possibilities are reserved at the same time on the basis of the quantum superposition state, and the multi-source data are integrated through the dynamic format adapter and analyzed into the intermediate format data. And a global optimal solution is obtained through rapid parallel search of a quantum annealing algorithm, and the accuracy and efficiency of node standardized mapping are improved.

Systems and methods for machine learning-informed response and augmentation of incomplete queries and query artifacts that include executing parallel search engines in a machine learning pipeline based on an input of a query to retrieve one or more corpora of candidate digital artifacts, constructing a merged corpus of candidate digital artifacts based on the retrieved corpora of digital artifacts, ranking each candidate digital artifact of the merged corpus of candidate digital artifacts based on a computed relevance to the input query, and returning, via a user interface, a response to the input query based on a subset of prioritized candidate digital artifacts most relevant to the input query.

The invention discloses a competitive evolution multi-task optimization method, system and equipment, and aims to solve the problem that an existing method is difficult to carry out collaborative optimization on fuel cost and gas emission under the condition that power balance and unit capacity constraint conditions are met. The method comprises the following steps: constructing main and auxiliary double-population parallel search; state indexes representing population convergence, diversity and feasibility are calculated in real time; based on the current state, a cooperation mode and evolution operator combined action is adaptively selected through a deep reinforcement learning agent; rewards are calculated according to improvement of the population on cost, emission and constraint satisfaction after action execution; and training the intelligent agent by using the state, the action and the reward, and iteratively optimizing the decision strategy until the Pareto optimal scheduling scheme meeting the constraint is output. According to the method, adaptive intelligent guidance of the evolutionary process is realized, and the optimization quality, the convergence speed and the algorithm robustness of the economic emission scheduling scheme of the power system are remarkably improved.

The invention discloses a method, a device and equipment for realizing a large-scale hybrid strategy. The method comprises the following steps of: receiving a message and analyzing the message to extract a field representing a data stream feature; generating intermediate state information according to an entry number, a port attribute and the field obtained through analysis, and obtaining a label through table look-up operation; fields, intermediate state information and labels of the messages are combined according to strategy rules defined by different storage blocks and compared with strategy entries in the storage blocks; and implementing a corresponding processing behavior on the message according to a comparison result. According to the method, parallel searching and processing behaviors of large-scale mixed policies can be realized, the method is suitable for more than hundreds of thousands of policy routing deployment and access control lists of various matching field combinations, and high-power-consumption and high-cost TCAM (Ternary Content Addressable Memory) storage can be avoided.

The invention relates to the technical field of unmanned aerial vehicle inspection, in particular to a transformer substation unmanned aerial vehicle dynamic route planning method and system. A patrol demand is converted into a detailed point location data set, a unified safety patrol point location set is generated by using coordinate conversion, safety radius calculation and shooting parameter setting, safety constraints are introduced through a path selection probability formula based on an improved ant colony algorithm, and a multi-ant colony parallel search and an adaptive pheromone updating mechanism are combined to obtain a safety patrol point location set. The method comprises the steps of generating an optimized shortest route, then performing to-be-examined identification generation and verification on an optimized waypoint set, performing path smoothing processing on the waypoint set, fitting a starting point, an end point and a middle patrol point through a third-order Bezier curve, smoothly joining the route, meeting a safe steering angle, path smoothness and a safe radius constraint, and finally obtaining the optimal route. According to the method, the requirements of the substation patrol task on safety, high efficiency and accuracy are met.

The invention discloses a broadband spread spectrum signal capturing device based on a rapid lock judgment feedback mechanism, and belongs to the technical field of spaceflight measurement and control communication. According to the method, a pseudo code parallel multi-channel capturing method and a carrier serial search capturing method are adopted, compromise processing is carried out between signal capturing time and consumed resources, and rapid carrier capturing can be completed through a rapid judgment locking feedback mechanism. According to the invention, rapid carrier capture of high-bandwidth spread spectrum signals can be completed under the condition of less resource consumption. According to the invention, the engineering realization occupies less chip resources, a low-performance processing chip can be adopted, the cost is saved, and the method is suitable for low-cost application scenarios. Through a rapid carrier wave lock judgment feedback mechanism, carrier wave serial search is combined with pseudo code parallel search, so that a corresponding carrier wave frequency is rapidly found during carrier wave frequency search, and a corresponding chip phase is found through pseudo code parallel search. The method is simple in design, easy in engineering realization and low in difficulty.

The invention discloses a commodity type selection recommendation method and system based on big data, and relates to the technical field of big data intelligent recommendation, and the method comprises the steps: carrying out the feature screening through CIIF and Shannon entropy, carrying out the optimization through an edge contribution degree and an adaptive pruning threshold value, carrying out the adjustment through an RL-QN strategy, carrying out the aggregation through HGCN, and generating a behavior enhancement rule set through NVR logic regularization. The dynamic weight of the rule is calculated through incremental reasoning, PoNSGA-III is used for optimization, a Combb-K objective function and a particle optimization algorithm are combined for global search, and a virtual center is adjusted to optimize a particle update path. According to the method, the personalized precision of commodity recommendation is improved through a collaborative linkage mechanism among feature screening, heterogeneous graph optimization and HGCN aggregation, and the diversity and stability of commodity recommendation are improved through fusion of a rule logical reasoning engine and a neural network regularization technology in combination with a parallel search mechanism of a multi-objective optimization algorithm.

The present invention provides a big data standardization method and system based on data integration, including: using a dynamic format adapter based on data conversion mapping rules combined with a reinforcement learning Schema algorithm to convert original data from different data sources into intermediate format data; using quantum state encoding to generate a quantum superposition state of each field in the intermediate format data and a node in a knowledge graph; using a quantum annealing algorithm based on the quantum superposition state to solve the optimal mapping node of the field in the knowledge graph; based on the optimal mapping node combined with the semantic constraints and standard format of the node, the field is standardized and mapped to obtain standardized data. In the present invention, multi-source data is integrated and parsed into intermediate format data through a dynamic format adapter, thereby solving the problem of data heterogeneity, retaining multiple mapping possibilities simultaneously based on the quantum superposition state, and obtaining the global optimal solution through rapid parallel search through the quantum annealing algorithm, thereby improving the accuracy and efficiency of node standardized mapping.

The invention discloses a bee colony navigation method and system based on multi-source grouping-evidence fusion, and belongs to the technical field of multi-robot collaborative search and environmental perception. According to the method, parallel search can be carried out in a multi-source environment through a grouping-sharing mechanism, and the multi-source recall rate is remarkably improved. And through IP evidence fusion and multi-model Bayesian fitness, the robustness of the algorithm to odor intermittency, turbulence disturbance and model mismatch is enhanced, and the false drop rate is reduced. Finally, through virtual potential field navigation, reachability and safety of a search path are guaranteed, and search failure caused by obstacles or cooperation conflicts is avoided. The method is suitable for multiple fields of industrial leakage detection, environmental pollution monitoring, dangerous goods search and rescue, public safety monitoring and the like, and has a wide application prospect.

The invention discloses a high-dimensional space-oriented distributed graph index nearest neighbor search method, which comprises the following steps of: firstly, designing a boundary sensing balanced partition strategy, and measuring boundary characteristics of quantized nodes through a connection ratio; secondly, constructing a hierarchical mixed index architecture, vertically integrating a sparse global navigation layer and a fine-grained local precision layer, managing high-importance boundary nodes by adopting an M-Tree structure, and remarkably improving the cross-partition retrieval efficiency by utilizing the relative insensitivity of a tree structure to a curse in a boundary region; finally, a distributed multi-starting-point parallel search framework is provided, complementary entry points are dynamically generated based on query features and graph topology, vector space multi-region parallel exploration is achieved, irrelevant partitions are filtered through a navigation layer, and then accurate similarity calculation is executed in selected partitions. According to the method, on the premise that the high recall rate is guaranteed, the throughput of high-dimensional vector distributed search is remarkably improved, and an efficient and reliable solution is provided for large-scale high-dimensional vector retrieval.

The invention discloses a fine-grained flow scheduling priority marking system for a hardware unloading network, which comprises a packet analysis module for extracting quintuple flow identification information from a data packet and supporting multi-layer protocol field extraction, and a flow table module for maintaining the real-time state of the current active data flow by adopting a double-hash table structure and supporting parallel search and update, the priority judgment module is used for allocating scheduling priorities to current messages through configurable boundary matching rules on the basis of state information returned by a flow table, the message recombination and marking module is used for encoding priority labels into DSCP fields of an IPv4 message header, and the whole system adopts a full-assembly-line hardware architecture and can be ensured to continuously run at a linear speed under a 100Gbps network load. Dynamic packet-by-packet scheduling control is supported, a scheduling priority based on a real-time flow state can be given to each data packet, and fine scheduling of network resources is realized.

The invention discloses a hot word-fragment mapping management method, which comprises the following steps that: a hot word-fragment mapping list is constructed, a plurality of records are recorded in the hot word-fragment mapping list, and each record comprises a hot word and a fragment list corresponding to the hot word; the hot word is a certain word / statement of which the search frequency reaches a search threshold value in a set time window; according to a search request, preferentially matching the hot word-fragment mapping list, and if the hot word-fragment mapping list is hit, distributing the search request to a corresponding fragment for processing; and if not, broadcasting all the fragments to execute the search request. The invention also discloses an intelligent fragment scheduling method for parallel search, which comprises the following steps of: when a search request arrives, preferentially matching the hot word-fragment mapping list according to the search request, if the hot word-fragment mapping list is hit, acquiring the comprehensive score of each fragment in the corresponding fragment list, processing the search request by the fragment with the highest comprehensive score, and constructing a fragment request; and merging the fragmentation results to complete the search request.

The present invention discloses a signal detection system and method for a MIMO system based on parallel Cholesky matrix decomposition, which mainly solves the problems of high complexity in signal detection of existing MIMO systems and doubling of detection delay when performing signal detection in the real number domain. The implementation steps are as follows: the channel matrix conversion module converts the channel complex matrix into a channel real matrix and generates an iterative matrix; the channel matrix decomposition module parallelly decomposes the iterative matrix by using the Cholesky matrix decomposition method; the K-Best parallel search module parallelly searches for paths between adjacent two layers by using the K-Best parallel search method; the detection result decision module completes the detection of signals of all antennas and determines the detection value of the transmitted signal. The parallel design of the signal detection process in the present invention effectively reduces the detection complexity, reduces the detection delay, and can realize signal detection with fewer resource quantities.

This invention discloses a GNSS integer ambiguity fixing method integrating intelligent optimization algorithms and VIB strategies, belonging to the field of satellite navigation and positioning technology. This method combines the advantages of VIB strategies and swarm intelligence fusion optimization algorithms. By directionally grouping the ambiguity floating-point solution covariance structure, the ambiguities are guided into groups. The PSOAF and AWDE intelligent optimization algorithms are used to perform parallel searches on the grouped ambiguities, improving search efficiency and positioning accuracy. Simultaneously, the probability assessment and fixing judgment of the integer solutions of each sub-block ambiguity and the overall ambiguity ensure the reliability of the ambiguity solutions, making it suitable for complex and dynamic GNSS positioning scenarios.

The present disclosure relates to a method for performing a parallel search for a first positioning fix in a Global Navigation Satellite System (GNSS) receiver. The method includes, in some instances, determining prepositioning information, wherein the prepositioning information includes a receiver information and a satellite information for each satellite in a plurality of satellites. The method further includes determining a code phase search range and a frequency search range, based on the prepositioning information, for each satellite in the plurality of satellites. The method further includes determining a starting point information for each satellite in the plurality of satellites, wherein each respective starting point information is representative of an offset from a center of a search range of the respective satellite. The method further includes performing the parallel search for all satellites in the plurality of satellites based on the respective code phase search range, the respective frequency search range and the respective starting point information.

The invention discloses a discrete space global optimization method based on neighborhood mapping and multi-path parallel search. The method comprises the following steps: acquiring a discrete state variable to be optimized as an initial search starting point; generating an initial candidate set by applying a neighborhood mapping rule, and executing structure pre-screening; constructing a multi-path parallel computing architecture, and distributing the effective candidate states to a plurality of independent computing threads by taking the effective candidate states as local initial starting points of the independent search paths; executing deep iterative search on each search path, wherein each round of iterative search is screened according to an evaluation index system comprising a first evaluation index for macroscopic evaluation and a second evaluation index for microscopic difference distinguishing; and periodically synchronizing local optimal solutions output by all parallel search paths, and screening to obtain a global optimal solution. According to the method, the problem of search stagnation during construction of the high-performance password component is solved, and the high-security password component with both high nonlinearity and low difference uniformity can be efficiently and stably generated.

A system and method for performing parallel local and remote searches, in particular to searching a local cache and remote database in parallel while accepting a search input, to obtain additional data for the local cache. The method includes, responsive to detecting a first portion of a search input, searching a local cache using the first portion to populate a search results list and initiating a remote search using the first portion; and responsive to receiving remote search results from the remote search, updating the search results list based on the remote search results and updating the local cache with data associated with items in the remote search results.

The invention provides an intelligent query system and method for cam configuration of a circular knitting machine, and is applied to the technical field of data processing. According to the method, multi-source data such as triangle models and process requirements and parameters such as configuration priorities are collected and are preprocessed and standardized into a unified query data set; core attributes are analyzed and extracted and are sorted according to parameter importance, and an optimized data set is generated; determining a retrieval strategy, a retrieval range and the number of iterations in combination with the demand complexity and the computing power to form a query batch; efficiency is improved through parallel retrieval, a query algorithm is optimized based on a multi-dimensional architecture, and a strategy is verified and dynamically adjusted through double indexes; and finally, according to a model rule optimization result, a redundant scheme is eliminated, and accurate, efficient and unique triangular configuration intelligent query is realized.

The invention discloses an e-commerce advertisement putting strategy generation method and system based on quantum computing optimization, and particularly relates to the technical field of quantum computing, comprising parameter coding, quantum optimization, line evolution, strategy resolving, dynamic optimization and strategy verification and output. According to the method, a high-dimensional parallel search mechanism based on a parameterized quantum circuit is constructed by introducing a quantum calculation normal form, the problem of calculation complexity index increase of a classical computer in e-commerce advertisement multi-parameter combination optimization is effectively solved, and the multi-parameter combination optimization of e-commerce advertisements is realized by designing a classical and quantum mixed architecture and a dynamic feedback optimization mechanism. The real-time adaptive adjustment of the advertisement putting strategy is realized, the practicability and robustness of the system are remarkably improved, the precise alignment of the technical means and the commercial target is realized by deeply coupling the quantum optimization process with the specific e-commerce advertisement business index, and a quantifiable business improvement effect is provided.

The invention discloses a concurrent implementation method of motion vector search, which relates to the technical field of H264 video compression implementation, and comprises the following steps of: judging whether a current macro block meets a condition of becoming a PSKIP macro block according to PSKIP macro block type information of a block adjacent to the current macro block, if the current macro block is judged to be in a PSKIP macro block type, ending search, and otherwise, entering the next step; under the assumption that the partitioning mode is a P16 * 16 partitioning mode, taking a pskipmv variable value obtained by predicting an adjacent block of the current macro block as a prediction vector value of the current macro block, and calculating and judging whether the type of the current macro block is a PSKIP macro block type or not; and when the judgment result does not meet the condition of the PSKIP macro block, carrying out parallel search processing, and carrying out concurrent motion vector search in different blocking modes. Under the condition that a small number of code rates are increased, the compression speed is greatly increased, and the method is more suitable for the use scene of real-time video compression.

The invention discloses a fluid pulsation accelerator architecture for reinforcement learning of a large language model, and belongs to the technical field of artificial intelligence hardware acceleration. The system aims at solving the problems that when a current universal processor executes an RLHF working load, an instruction driving normal form is not matched with data flow calculation, fixed parallel granularity cannot adapt to a dynamic load, and the efficiency is low due to the fact that a framework does not sense data statistical characteristics. The core of the system is a liquid systolic array calculation fabric capable of being dynamically reconstructed, execution is triggered through data flow, and kernel boundaries and instruction overhead are eliminated. According to the system, a global lookup table subsystem is fused, and nonlinear calculation is optimized through merging query and parallel lookup by using data distribution prior; and a dynamic scheduling unit is configured, and elastic parallelism is realized by adopting a resource allocation algorithm supporting work stealing. According to the method, the throughput and the energy efficiency of RLHF training and a large language model reasoning stage can be remarkably improved, and a new design normal form is provided for a next-generation AI special computing architecture.

The technical scheme of the invention discloses a compact type 5T0C eDRAM (Dynamic Random Access Memory) gain unit. According to the other technical scheme, the invention discloses a high-density content addressing circuit adopting the compact 5T0C eDRAM gain unit. Aiming at the problems of large storage unit area and limited density in the traditional CAM design based on SRAM (Static Random Access Memory), the invention provides a novel architecture based on eDRAM (Enhanced Dynamic Random Access Memory), and by optimizing the storage unit structure and a reading circuit, the storage density is obviously improved, and CAM search and in-memory logic operation are supported at the same time. Compared with an existing 6T SRAM, the area of the CAM array based on the eDRAM is reduced by 15%. According to the invention, the highest search and logic operation frequency of the binary CAM (BCAM) of 637 / 658 MHz is realized, and the search power consumption and the logic operation power consumption under the voltage of 1.1 V are respectively 0.91 fJ / search / bit and 27.47 fJ / bit. The method is suitable for data intensive applications such as artificial intelligence accelerators, network routers, genome analysis and the like which need efficient parallel search and near-memory calculation.

A photonic content addressable memory (CAM) (100) is provided. The photonic CAM (100) comprises: N search lines (101) and a plurality of match lines (102), each match line comprising N photonic CAM cells (110). Each search line (101) is configured to receive a search optical signal and each search optical signal comprises a different wavelength encoding one of N parallel search symbols. Each match line (102) is configured to receive: a wavelength division multiplexed content optical signal comprising a plurality of wavelengths, each wavelength encoding one of N parallel content symbols; and the N search optical signals from the N search lines (101). Each photonic CAM cell is configured to compare a respective search symbol with a respective content symbol and provide an output optical signal that is responsive to the comparison. Each match line (102) also comprises a content optical demultiplexer (120), configured to demultiplex the wavelength division multiplexed content optical signal and provide each CAM cell (110) with the respective content symbol.

The invention discloses a GNSS integer ambiguity fixing method fusing an intelligent optimization algorithm and a VIB strategy, and relates to the technical field of satellite navigation positioning, the method fuses the advantages of the VIB strategy and a swarm intelligent fusion optimization algorithm, directivity grouping is carried out on an ambiguity floating point solution covariance structure, ambiguity is subjected to guiding grouping, and the ambiguity is subjected to guiding grouping. The PSOAF and AWDE intelligent optimization algorithms are adopted to perform parallel search on the grouped ambiguity, so that the search efficiency and the positioning precision are improved. Meanwhile, probability evaluation and fixed judgment of real solutions are carried out on integer solutions of the ambiguity of each sub-block and the overall ambiguity, the reliability of the ambiguity solutions is ensured, and the method is suitable for GNSS positioning scenes in complex and dynamic environments.

The application discloses a kind of wideband spread spectrum signal acquisition device based on fast lock judging feedback mechanism, belong to spaceflight measurement and control communication technical field.The method uses the method of parallel acquisition of pseudo code multiple ways, carrier serial search acquisition, compromise between signal acquisition time and resource consumption, and can complete the fast acquisition of carrier by fast lock judging feedback mechanism.Can complete the fast acquisition of carrier of high bandwidth spread spectrum signal under the condition of less resource consumption.The engineering implementation of the application occupies less chip resources, can use low-performance processing chip, save cost, and can be applied to low-cost application scenarios.Can find corresponding carrier frequency quickly when searching carrier frequency by fast carrier lock judging feedback mechanism, and find corresponding chip phase by pseudo code parallel search.The application is simple in design, easy to implement in engineering, and has low difficulty.

The application discloses a kind of energy-saving optimization method and system of manufacturing factory water-cooled air conditioner, it is related to intelligent energy-saving optimization technical field, including collection multimodal data, construct initial gene bank, form cluster gene bank by spectrum decomposition and K-means, using genetic algorithm to update cluster gene bank, output configuration parameter table, set target function and heat balance constraint condition, convert into quadratic binary optimization problem, using quantum annealing algorithm parallel search, by discretization mapping and screening, output initial operation optimization parameter, carry out use and calculate deviation rate, update GP model parameter, output operation optimization parameter.The application forms configuration parameter table by exergy analysis combined with advanced algorithm, improves the accuracy and adaptability of system initial configuration, introduces quantum annealing and GP model dynamic optimization operating parameter, improves the dynamic adaptability and optimization efficiency of operating parameter.