78results about How to "Efficient aggregation" patented technology

The invention discloses a multi-modal emotion understanding method based on cross-modal semantic alignment and interactive learning. The method comprises the following steps: acquiring multi-modal initial data; performing semantic quantification processing on the initial data of each mode; according to the semantic feature set, text modal dominant cross-modal feature aggregation learning is carried out; according to the multi-modal semantic aggregation feature expression, performing intra-modal semantic relationship learning of the multi-modal features, determining a context relationship between intra-modal semantics Tokens, and performing semantic alignment with a text modal to enhance semantic consistency of different modal feature expressions; carrying out fusion processing on the multi-modal features by adopting a multi-modal fusion strategy to obtain fusion features; and establishing an emotion understanding model according to the fused features, and realizing multi-modal emotion understanding processing according to the emotion understanding model. The embodiment of the invention can improve the sentiment analysis performance, is beneficial to promoting effective fusion of different modes, and can be widely applied to the technical field of computers.

The invention discloses an ODPS-based customer service data aggregation method and server, and relates to the field of electric digital data processing, and the method comprises the steps: carrying out the field classification of collected source data, and obtaining a plurality of source data sets; loading the source data sets to an ODPS platform, and performing first aggregation processing on the source data in the plurality of source data sets by using the ODPS platform to obtain a plurality of field wide tables; pulling the field wide table to a DRDS platform, and performing distributed data processing on the field wide table by using the DRDS platform to obtain a plurality of pieces of distributed data; determining a target data feature corresponding to a target report in a report request instruction according to the report request instruction sent by the user terminal; and extracting multiple pieces of target data corresponding to the target data features in the multiple pieces of distributed data, and performing second aggregation processing on the target data to generate a target report. By implementing the method, the data processing efficiency can be improved when the multi-azimuth data display of the customer service staff is carried out.

The invention discloses a tomato maturity detection system and method based on deep learning, and relates to the technical field of fruit and vegetable detection models. Comprising the following steps: acquiring data from a robot perspective based on a natural scene of a tomato plantation, and making a tomato maturity detection data set; on the basis of a YOLO v5s model, an improved model YOLOv5s BCD model fusing a CA attention mechanism, a BiFPN neck feature fusion network and a Decup led Head head decoupling network is provided and serves as a tomato maturity detection model of the invention; experimental research of model training and effect comparison is introduced, the conditions of model evaluation and optimization are performed through a model comparison test, the tomato maturity grade is judged on the basis of the fruit red surface coloring proportion and in combination with fruit appearance characteristics, tomatoes with the required maturity grade are collected, the light weight and accuracy of edge side deployment are adapted, and the tomato maturity grade evaluation method is suitable for the tomato maturity grade evaluation and the tomato maturity grade evaluation and the tomato maturity grade evaluation and the tomato maturity grade evaluation and the tomato maturity grade evaluation. The system can be used as a visual system of a tomato harvesting robot, and technical support is provided for automatic picking.

The invention belongs to the field of traffic control systems, and relates to a vehicle driving track prediction system and method based on an improved graph neural network model, and the system comprises an FPGCN model and a Social LSTM model. The FPGCN model comprises a graph convolution layer, a Dropout layer, a differential pooling layer, a feature fusion module and a full connection layer; obtaining node feature vectors from graph structure data through a graph convolution layer, a Dropout layer and a differentiable pooling layer, then dividing and conquering node features, generating global features through a dense attention mechanism and principal component analysis, and generating local features through virtual filling and matrix calculation; after global features and local features are obtained, a feature fusion module is used for fusion, then the features are input into a full connection layer, and the influence score of a historical trajectory diagram on a current state diagram is calculated; integrating the preprocessed driving data and the data of which the score is calculated by the FPGCN model, and inputting the integrated data into a Social LSTM model; according to the system, while the model expression ability is maintained, the parameter quantity is remarkably reduced, the calculation logic is optimized, and the prediction accuracy is improved.

The invention discloses a continuous production process of a CO2-based multipolymer, and belongs to the technical field of polycarbonate synthesis. The technological process comprises the following steps: firstly, preparing an auxiliary agent, a solid-state monomer, a liquid-state reaction monomer and any solvent into a mixed solution in a premixer (1); continuously feeding carbon dioxide and the mixed solution into a multi-stage continuous reactor for mixing and heating to initiate a reaction, wherein the total reaction residence time of the materials in the multi-stage continuous reactor is controlled to be 10 minutes to 5 hours; and after the reaction is completed, carrying out post-treatment to complete the continuous production of the CO2-based multipolymer. The process particularly optimizes the effective continuous polymerization of high-molecular and high-viscosity materials of the CO2-based multipolymer. Through accurate control of the multi-stage tubular reactor, appropriate reaction viscosity is realized, further reaction uniformity of materials in the screw extrusion reactor is ensured, and reaction efficiency is improved.

The invention discloses efficient oil removal and purification equipment in a deep sea gas field ethylene glycol enriching system, and belongs to the technical field of petroleum treatment, the efficient oil removal and purification equipment comprises purification equipment composed of multiple stages of oil removal devices, the oil removal devices are connected stage by stage through communicating pipes, and each oil removal device comprises a separation pipe, a coalescence module and an oil removal film module, and the coalescence module and the oil removal film module are arranged in the separation pipe; the coalescence module and the oil removal film module are arranged, a coalescence technology and an oil removal film technology are combined, a set of dual separation system is constructed, the combined technology shows obvious advantages when an oil-water mixture with high emulsification degree is treated, a super-hydrophilic fiber bed layer is arranged in the oil removal film module, the fiber bed layer utilizes a water film barrier effect, and the oil removal effect of the oil-water mixture is improved. According to the oil-water separator, the oil-water mixture can be further captured and polymerized without being completely polymerized by the coalescer, and a multi-layer design is adopted, so that the oil-water mixture is treated for multiple times when passing through the equipment, the whole-process treatment from crude separation to fine separation is realized, and the overall separation efficiency of the equipment is greatly improved.

The invention discloses a double-bridge normal type U-shaped medical image segmentation network, which comprises an encoder, a double-bridge framework and a decoder, the encoder adopts a backbone network, the double-bridge architecture comprises a multi-scale bridge and a single-scale bridge, the multi-scale bridge adopts an SCAM, the single-scale bridge adopts a CAM, and the decoder adopts an RMSCAM; the encoder extracts multi-stage image features from the medical image of the single-channel gray scale, and the features extracted by the encoder in four stages are fed into the double-bridge architecture; the multi-scale bridge receives features extracted from the four stages of the encoder and captures global features; the single-scale bridge receives features extracted by the encoder in a single stage and captures local features; the global features and the local features are added and fused and then sent to a multi-stage decoder; the decoder restores the feature map to obtain a final prediction map; according to the method, the problems of semantic difference between different stage features extracted by the encoder and semantic difference between the encoder and the decoder can be effectively solved.

The invention discloses a generalized rapid Mesh extraction method based on Gaussian, and the method comprises the steps: configuring a training data collection system, and carrying out the marking and preprocessing of data; the method comprises the following steps: constructing a generalized rapid Mesh extraction network based on Gaussian, and training the rapid Mesh extraction network; a training model of a generalized rapid Mesh extraction network based on Gaussian is used, testing is carried out under multiple data sets and different scenes, index evaluation is carried out on a result, the performance of the Mesh extraction method based on the Gaussian model can be further improved through a self-adaptive adjustment strategy and an optimized parameter estimation method, and the performance of the Mesh extraction method based on the Gaussian model is improved. The method has important research significance in theory, particularly has wide practical application prospects in scenes needing to process large-scale and high-dimensional complex data, aims to overcome the defects in the prior art and improve the adaptability and efficiency in different data sets and application scenes, and provides an innovative solution for the field of three-dimensional data processing. And further development and application expansion of the Mesh extraction technology are promoted.

The invention provides a method for preparing a composite solid electrolyte through a Lewis fluorine supply nano additive, and the composite solid electrolyte comprises the following components in parts by weight: 10-20 parts of a polymeric monomer, 2-6 parts of a lithium salt, 1-4 parts of a nano functional additive and 5-12 parts of an organic solvent. When the prepared composite solid electrolyte is used for a lithium ion battery, the composite solid electrolyte has the characteristics of no need of an additional initiator, uniform polymerization reaction, high ionic conductivity and the like, and the stability, the rate and the cycle performance of the semi-solid lithium ion battery can be improved; according to the preparation method, no extra initiator is added into the preparation raw materials of the in-situ cured solid electrolyte, the lithium salt is instably decomposed into the lewis acid in a heating state, the lewis acid is utilized to efficiently initiate ring-opening of cyclic ether to perform polymerization reaction, no initiator is left, and no extra side reaction is generated to cause reduction of battery performance.

The invention provides a method and equipment for aggregation optimization of distributed resources of a virtual power plant under homomorphic encryption. The method comprises the steps of obtaining distributed resource data of a virtual power plant; performing classification; homomorphic encryption processing is carried out to obtain encrypted data; performing homomorphic encryption optimization to obtain a ciphertext of the distributed resource data of the virtual power plant; based on the encrypted virtual power plant distributed resource data, determining a target and constraint conditions of virtual power plant distributed resource aggregation optimization, and constructing a virtual power plant distributed resource aggregation optimization model; and solving the virtual power plant distributed resource aggregation optimization model through a heuristic algorithm to obtain a virtual power plant distributed resource aggregation optimization scheme. Under the condition that original data are not exposed, efficient aggregation and optimization of the distributed resources of the virtual power plant are achieved. The security and privacy of the distributed resource data of the virtual power plant are ensured, the energy utilization efficiency and the system stability are improved, and technical support is provided for the construction of an intelligent power grid and the virtual power plant.

The invention provides an antistatic heat-insulating coating and a preparation method thereof, and relates to the technical field of coatings. The antistatic heat-insulating coating is prepared from polyacrylate, titanium dioxide, antistatic heat-insulating filler, hydrophilic aerogel, diatomite, water and other raw materials, the polyacrylate comprises Vitrimer xylose polyacrylate with a specific structure, the Vitrimer xylose polyacrylate is used as a substrate material of the coating, and the mechanical property, heat resistance, durability and the like of the coating can be remarkably improved; furthermore, the antistatic heat-insulating filler comprises an antistatic heat-insulating composite material which is formed by compounding a light material, polypyrrole and a carbon material, the composite material is prepared by an in-situ polymerization method, and a unique structure that a composite microlayer is formed by compounding the polypyrrole and the carbon material on the surface of the light material is formed; the anti-static heat-insulating composite material and other raw materials in the coating cooperate with each other to act together, so that the prepared coating has good anti-static performance and heat-insulating performance.

The invention discloses putty for a special vehicle body material and a preparation method of the putty, and relates to the technical field of putty. The putty for the special vehicle body material is prepared from modified unsaturated polyester resin, a dispersing agent, modified talcum powder, bentonite, fumed silica, titanium dioxide, cobalt iso-octoate, cyclohexanone peroxide and dibutyl phthalate. The modified unsaturated polyester resin is obtained by reacting a modified unsaturated polyester resin precursor with hydroxyalkyl-terminated modified polysiloxane and then reacting with isocyanate terpyridyl; the modified talcum powder is obtained by reacting talcum powder modified by p-aminophenyltrimethoxysilane with chlorobenzene, cyclizing by using powdered sulfur, reacting with 1-iodoheptane, formylating by using phosphorus oxychloride, and grafting p-trifluoromethylaniline and arginine. The putty for the special vehicle body material prepared by the invention has the advantages of quick drying, antibacterial property and heat resistance.

The present invention relates to a novel compound which modulates the function and differentiation of a cytoskeleton involved in pulmonary fibrosis and thus can be used for preparing a medicament capable of treating related diseases. It has been determined that: the novel compound according to the present invention or a pharmaceutically acceptable salt thereof has an effect of modulating actin polymerization required for cytoskeleton formation; and the expression of alpha-SMA and F-actin, which are main factors causing fibrosis, is effectively inhibited in human lung tissue-derived fibroblasts (MRC5). Furthermore, it has been determined that when activated macrophages are treated with the compounds, the level of inflammation of inflammatory cytokines caused by the macrophages is reduced. Therefore, the novel compound according to the present invention is expected to be effectively used in the prevention and treatment of diseases associated with pulmonary fibrosis as a preparation having an effect of inhibiting or reducing pulmonary fibrosis.

The invention provides a cross-domain transfer learning laser ultrasonic layering defect detection method suitable for a composite material and application thereof.The method comprises the steps that firstly, multi-channel input containing a frequency domain, a time-frequency domain and Koopman operator characteristics is constructed for the problem of data distribution deviation caused by the batch difference of the composite material; secondly, extracting and fusing high-dimensional spatial-temporal features through a multi-channel attention mechanism and a graph convolutional network; and finally, migrating the defect features of the source domain to the target domain in combination with dynamic distribution self-adaption and domain adversarial training strategies. The method can effectively overcome model performance reduction caused by material attribute change, realizes high-precision layering defect identification and imaging under the condition of no target domain label, and is suitable for composite material detection in a complex industrial scene.

The application provides a kind of azabicyclometalated macrocyclic complex of formula (I) and its preparation method and application. As a catalyst, it can effectively catalyze the polymerization of ethylene and other olefins, and is suitable for continuous or intermittent solution method to prepare olefin materials, and has better catalytic effect than mononuclear metal complexes.

The invention discloses transparent photosensitive resin for DLP 3D printing and a preparation method of the transparent photosensitive resin. The transparent photosensitive resin for DLP 3D printing comprises the following components in percentage by weight: 40%-60% of tetrahydrofuran acrylate, 20%-30% of bifunctional polyurethane acrylate, 19.5%-29.5% of a reactive diluent, 0.02%-0.03% of a light absorber, 0.2%-0.3% of a photoinitiator Irgacure-2959, 0.1%-0.3% of a defoaming agent and 0.1%-0.3% of a flatting agent, and the transparent photosensitive resin for DLP 3D printing is prepared by fully stirring and mixing the components at room temperature. The photosensitive resin for DLP 3D printing is transparent, low in viscosity and high in product speed, and the rapid formability of a 3D printing product is guaranteed; after curing, the volume shrinkage change is small, and the dimensional stability of the product is high; the mechanical property is stable, and the anti-aging property is excellent.

This invention relates to the field of digital archives management technology, specifically to an artificial intelligence-based system for analyzing and warning of abnormal archive behavior. The invention uses a graph construction module to transform archive access logs into a relational topology graph; an attribute propagation module performs weighted transfer and accumulation of weights based on a graph traversal algorithm to calculate the cumulative relational index value of nodes; a connectivity retrieval module retrieves cross-regional links based on partition isolation rules; and an early warning output module generates early warning signals based on index values ​​and link status. This invention utilizes the flow mechanism of feature values ​​in a graph structure to aggregate minute risky behaviors within discrete long-term windows, accurately identifying complex abnormal patterns where a single operation may not violate regulations but the long-term cumulative risk exceeds the limit. This solves the problem that existing static statistical rules are unable to detect hidden logical violations, effectively reducing false alarms and false negatives.

This invention relates to the field of energy forecasting technology, and in particular to a method and system for short-term energy load forecasting based on an improved Transformer using SE-Block. The method includes reversible normalization preprocessing of acquired multivariate load sequence data; feature extraction and fusion of the preprocessed data using improved cross-scale interactive patching, including multi-scale feature extraction, cross-scale interactive alignment, residual correction, and dynamic fusion; and feature filtering of the fused features based on a channel attention mechanism, including feature response based on improved SE-Block and nonlinear interaction of context vectors. This invention addresses the non-stationarity of actual load caused by meteorological conditions and user behavior. By automatically eliminating noise interference among multiple variables, the model accurately depicts the fluctuation details of the load curve, demonstrating its robustness in multivariate load forecasting for integrated energy systems.

This invention provides a method for predicting the TCM functions of a single herbal drug and for drug screening. The method includes: treating each atom in the molecule of the herbal drug to be predicted as a node in a graph; constructing feature vectors for each node based on the physical and chemical properties of each atom; constructing a feature vector matrix based on the feature vectors of all atoms; treating the chemical bonds between atoms as edges in the graph; constructing an adjacency matrix based on the bonding types between atoms; constructing a graph structure matrix based on the feature vector matrix and the adjacency matrix; inputting the graph structure matrix into a graph convolutional neural network model to obtain a function score vector, the length of which is equal to the total number of preset TCM functions, and each score in the function score vector representing the probability that the herbal drug to be predicted belongs to a corresponding preset TCM function; comparing each score in the function score vector with a preset threshold, and determining the predicted TCM function value of the herbal drug to be predicted based on the comparison result. This invention improves the accuracy of function prediction.

The invention relates to poly (p-hydroxystyrene) photoresist film-forming resin and a preparation method thereof, relates to the technical field of photoresist materials, and is used for solving at least one of the problems of relatively wide molecular weight distribution, uncontrollable molecular weight and the like of the film-forming resin prepared by the existing method. According to the method disclosed by the invention, free radical polymerization reaction is carried out under photo-initiation through a reversible addition-fragmentation chain transfer polymerization technology, so that the reversible inactivation reaction of free radicals in a polymerization reaction system is controlled, the molecular weight distribution of the synthesized deep ultraviolet photoresist film-forming resin (poly-p-hydroxystyrene photoresist film-forming resin) is narrow, the PDI is less than 1.2, the reaction is instantly controlled, and the film-forming effect is good. The reaction conditions are mild; the energy consumption is low; and the product is easy to separate and purify.

The utility model relates to a water supply sewage treatment filtering device which comprises a supporting assembly and a filtering barrel rotationally arranged on the supporting assembly, a barrel is coaxially arranged in the filter barrel, an annular filter cavity is formed between the barrel and the filter barrel, an annular sealing plate is arranged above the cavity, a plurality of layers of concentric circle filter materials are arranged in the cavity in the radial direction, and through holes communicated with the annular filter cavity are formed in a barrel body of the filter barrel and the barrel. 4-12 blades are arranged on the upper portion of the inner wall of the barrel in a circumferential array mode, an end cover is coaxially arranged above the barrel through a bearing, a water inlet is formed in one side of the end cover and connected with a water pipe, the water outlet direction of the water pipe is tangent to the inner wall of the side face of the end cover, and tangential water outlet is matched with the blades to drive the filtering barrel to rotate. The bottom of the filter barrel is conical or arc-shaped, and a drain outlet is formed in the center. The device solves the problems that a traditional filtering device is low in efficiency and needs extra power, hydraulic self-rotation driving and multi-stage annular filtering without external power are achieved, the filtering effect is remarkably improved, and the operation and maintenance cost is reduced.

This invention discloses a centrifugal clutch-type drainage manhole cover, comprising: a supporting base; a manhole cover body, the manhole cover body including a cover body, the cover body having annular and spaced primary drainage outlets and secondary drainage outlets, and an opening and closing grid plate connected to the bottom of the manhole cover body below the primary drainage outlets; and a rotating valve plate assembly, the rotating valve plate assembly being placed below the secondary drainage outlets and connected to the bottom of the manhole cover body. This invention uses gravity as the primary trigger signal and water turbine rotation speed as the secondary trigger signal, achieving normal micro-drainage and increased drainage during heavy rainfall through a filter chassis linked to the opening and closing grid plate. It also utilizes water flow to drive gears for active unblocking. When rainfall further increases, the water turbine drives the centrifugal clutch mechanism to open the secondary drainage outlets, enabling the manhole cover system to automatically complete a three-stage stepped response from micro-drainage to primary strong drainage and then to secondary strong drainage according to the rainfall intensity, ensuring drainage effectiveness.

The application provides a forest fire smoke detection method based on semantic segmentation, and belongs to the technical field of fire warning, and comprises the following steps: feeding a real forest fire data set into a feature extraction network, inputting the extracted features into an improved DeepLabV3+ network for training, and outputting detection and recognition results. According to the complex and changeable forest fire scene, the non-rigid structure, the opacity and the variable mode of smoke and other characteristics, the application deeply studies a feature fusion method, adopts an improved ASPP structure to extract multi-scale spatial features, simultaneously introduces a bottom feature extraction module and a double-path convolution aggregation module to effectively extract smoke boundary information, and uses a lightweight backbone network to replace the original feature extraction network of DeepLabV3+, so that the detection speed is accelerated by using fewer parameters. The application can solve the problem that the false alarm and missed detection rates of the smoke detection method are high, thereby providing technical support for forest ecological protection.

The invention belongs to the field of intelligent algorithm fraud identification, and provides an imported commodity price fraud identification method based on FTransform-LightGBM, which improves the feature expression of price fraud data through the proposed double-branch architecture and dynamic feature fusion, and further improves the category identification accuracy of imported commodity price fraud. The method comprises the following steps: firstly, designing a double-branch architecture constructed based on a machine learning method and a deep learning method, and by fusing output features of LightGBM and FTModel models, better coping with data diversity and enhancing feature expression of fraudulent price data; secondly, a feature extraction branch based on FTModel is designed based on a full connection layer and Transform, feature expression of fraud price data is enhanced by focusing on extraction of key features of commodity price fraud data in a complex scene, and extraction of global features of the fraud data is enhanced; and finally, the proposed dynamic feature fusion can more autonomously select effective features through dynamic weight adjustment, deep mining of associated features is realized, and the model recognition performance is improved.

This invention discloses a real-time semantic segmentation algorithm for remote sensing images based on a dual-path structure, comprising the following steps: S1, acquiring Potsdam and GID-15 remote sensing image datasets and preprocessing them, then dividing them into training, validation, and test sets according to a set ratio; S2, constructing a real-time semantic segmentation network for remote sensing images based on a dual-path structure, including spatial path, context path, and dual-path feature aggregation; S3, inputting the preprocessed training and validation datasets into the real-time semantic segmentation network for training, calculating the cross-entropy loss function and performing backpropagation, updating the network parameters, obtaining a model that meets the requirements, and saving the model; S4, inputting the preprocessed test set into the model obtained in step S3, and outputting accurate segmentation of the remote sensing image. This invention improves the richness and effectiveness of semantic feature extraction while enhancing segmentation accuracy.

The invention relates to the field of data visualization, in particular to a scatter diagram visualization task-oriented data sampling quality evaluation method and system. The method comprises the following steps: giving a high-dimensional data set, and generating a high-dimensional data subset by using a data sampling method; obtaining a low-dimensional data subset according to the high-dimensional data subset; randomly extracting clusters of the high-dimensional data set to obtain high-dimensional space sub-clusters; obtaining a low-dimensional space sub-cluster according to the high-dimensional space sub-cluster; measuring the difference between the high-dimensional space sub-cluster and the low-dimensional space sub-cluster to obtain the difference between different dimensional space sub-cluster pairs; obtaining a global error according to the difference of the sub-cluster pairs in the different dimension spaces; and evaluating the sampling quality of the high-dimensional data through the global error. According to the method, by comparing the sub-clustering differences of different dimension spaces and measuring the retention conditions of different sampling methods after data projection, the authenticity of the high-dimensional data in the projection scatter diagram is effectively evaluated.

The invention discloses an urban disaster emergency management overall circle identification method, a terminal and a storage medium, and the method comprises the steps: obtaining mobile phone signaling data, and recognizing the main residence of residents before and during a disaster according to the mobile phone signaling data; according to the change characteristics of the main residences of the residents before the disaster and during the disaster, drawing residence change OD, and performing clustering analysis on the residence change OD to obtain resident residence change OD clusters; overlapping the boundary of the residential place change OD cluster with a community administrative boundary to determine a disaster prevention overall circle, and visually displaying the disaster prevention overall circle; the invention provides a new disaster emergency management overall planning district division method, the configuration of disaster prevention resources is optimized, and the evacuation efficiency is improved.

The invention discloses an encryption aggregation method and system for a household micro-grid, and is applied to the technical field of micro-grid group operation control, and the method comprises the steps: obtaining adjustable data of a plurality of household micro-grids; classifying the adjustable data of the plurality of household microgrids to obtain a plurality of to-be-adjusted data groups; and inputting the plurality of to-be-adjustable data groups into a micro-grid group system at least comprising a plurality of household micro-grids for processing to obtain to-be-aggregated ciphertext data, and performing threshold decryption processing on the homomorphic aggregated ciphertext by using a distributed key generation technology to obtain an aggregation result of the micro-grid group system. According to the method provided by the invention, the technical problems of privacy leakage and data security caused by plaintext uploading and user side sensitive operation data transmission in an existing household micro-grid aggregation mode can be solved, so that privacy protection of the original sensitive data of the household micro-grid is realized.

The application relates to an active load response system based on low communication demand, wherein a central control module and a plurality of frequency modulation modules are connected through distributed control communication; the frequency modulation module comprises a data receiving and transmitting unit, a data processing and storage unit and an active load response adjusting unit; the data receiving and transmitting unit is used for obtaining operation data of the central control module, other frequency modulation modules and power loads; the data processing and storage unit is used for determining consumed power of the active load response adjusting unit according to the operation data obtained by the data receiving and transmitting unit and the operation state of the active load response adjusting unit; and the active load response adjusting unit is used for adjusting consumed electric energy of the load through a communication mode or a direct adjustment mode of load power voltage and frequency. The application solves the technical problems of a large number of devices, small capacity and wide distribution in the prior art, improves the technical performance of device clusters, and provides frequency support and adjustment for a power system.

The present application relates to the technical field of address translation request scheduling, and particularly relates to an address translation request aggregation scheduling method and system. The present application maintains a pending request filter table, each table entry corresponding to a 4KB address range, recording address translation requests that have been sent but have not received responses. When a new request is received, the 4KB base address to which the virtual address belongs is calculated and the filter table is queried: if there is a corresponding table entry, the request is added to the associated waiting queue, and the new request is not sent; if there is no corresponding table entry, the new request is sent and a new table entry is created. Thus, during the waiting response period, subsequent requests in the same address range are aggregated, the number of repeated requests is reduced from N times to 1 time, the root complex load is significantly reduced, redundant transactions are reduced, and the hardware overhead is extremely low.