39results about How to "Relieve overload" patented technology

The invention provides a source driver applied to TFT-LCD. The source driver comprises a data register, a data latch, a digital to analog converter and an output buffer, wherein a first load pulse is supplied to the output buffer, so that the output buffer responds to a second edge, from a second level to a first level and closely following a first edge, of the first load pulse, and begins to output the greyscale voltages of odd numbered output ends to a corresponding TFT source; a second load pulse is supplied to the output buffer, so that the output buffer responds to a second edge, from a second level to a first level and closely following a first edge, of the second load pulse, and begins to output the greyscale voltages of even numbered output ends to a corresponding TFT source; at least the second edge of the first load pulse and the second edge of the second load pulse are asynchronous. The invention further provides a corresponding drive circuit and a drive method. The source driver, the drive circuit and the drive method can alleviate the adverse consequences arisen by excessive differences of two adjacent lines of data displayed.

The invention relates to a multi-root peer analytic method for identifications in the internet of things, comprising the steps of: peer authorization files are established and stored into each analytic root node, and the analytic root node edits a DNAME record in a root area space according to the corresponding peer authorization file; a client acquires an object identification code of an item and converts the object identification code of the item into URI in the form of standard FQDN, and sends an item object identification code query message to a local analytic root node NX; the NX extracts a country code from the message, and forwards, according to the DNAME record, the message to a root node NY having the right to analyze the message; the NY parse a DNS-based layer-by-layer until a certain authoritative server returns a name authority pointer record containing an information server address; and finally, the client communicates with the corresponding information server to acquire the detailed attribute information of the item. With the present invention, the single point failure defect existing in the existing analytic system can be overcome, and the efficiency of querying the object identification of the item can be improved.

The invention discloses an overload protection method, device and server, belonging to the field of Internet communication. The method includes: separately accumulating the number of requests received in consecutive m time intervals; extracting n time intervals from the m time intervals, and accumulating the number of requests received in each of the n time intervals The sum of the n time intervals includes the time interval to which the current time belongs; if the sum of the cumulative number of requests exceeds the predetermined threshold, reject receiving requests after exceeding the predetermined threshold in the time interval to which the current time belongs The request received; wherein, m and n are integers greater than or equal to 2, and m is greater than or equal to n. The present invention solves the problem that the existing overload protection method cannot effectively relieve the server overload during the peak period; realizes overload protection by comprehensively accumulating the accumulated numbers in multiple time intervals, thereby effectively alleviating the server overload during the peak period Effect.

The invention provides a method and a system for obtaining a user demand based on a context environment of a household workplace. The method comprises the steps of obtaining mobile user network usage information based on a dataset of a mobile network operator; performing data cleaning to remove abnormal user Internet-surfing records; extracting and remaining attributes related to user interests from the user Internet-surfing records; discovering user interest space; discovering the position of the household workplace by utilizing a squeezing principle; performing data integration and data dimension reduction; and performing model training by utilizing a random forest and classifying. According to the method and the system for obtaining the user demand based on the context environment of the household workplace, a new technical scheme is provided for mining user demands and interests conveniently and efficiently by utilizing the dataset which is composed of the user network usage detailed information and comes from a mobile network operator and based on the position discovery of the household workplace and the context environment thereof.

The invention provides a method for breaking a high-voltage (HV) superconducting current-limiting direct current (DC) circuit breaker. The method comprises the following steps: when a HVDC side is in short-circuit fault and current / current rising steepness rises rapidly and reaches a critical value of a current limiter, a superconducting material quenches immediately, impedance rises steeply, short-circuit current is limited within set values, and meanwhile, an instruction is sent to an ultra-quick mechanical switch; the ultra-quick mechanical switch acts immediately, enables contacts to separate at an ultrahigh rigid separating speed so as to form a physical gap and generate electric arcs, and immediately sends a breaking instruction to a power electronic switch; the power electronic switch quickly breaks the short-circuit current and enables the electric arcs between the contacts to be extinguished, so as to form an isolated fracture; when the contacts are separated to final open positions and current in a lightning arrester is attenuated to zero, the HV superconducting current-limiting DC circuit breaker completes a whole breaking process, and a HVDC transmission power grid is isolated from a faulty line. The method has the advantages that the cost performances of different components are taken into account comprehensively, so that the cooperation is optimized; the number of components and devices is reduced as much as possible, so that the size is reduced, the weight is reduced, and the reliability is improved.

The invention discloses a site selection and capacity calculation method for a flexible DC access point, which includes determining the scope of the power grid to be analyzed and constructing an overload scene set and an overload line set; calculating the weight factor of the overload line in each overload scene to obtain all flexible The power sensitivity of the DC access point position to the overloaded line; calculate the comprehensive sensitivity of the VSC-HVDC located at each flexible DC access point to all overloaded lines; construct a candidate access point set; construct a different The multi-objective optimization model of the scene is solved to obtain the capacity configuration result of each candidate access point; the candidate access point set and the capacity configuration result of the candidate access point are evaluated to obtain the final site selection result of the flexible DC access point and capacity allocation results. The method of the invention can effectively reduce the overload problem of the AC line in the N-1 scene, and has high reliability and good effect.

The invention provides a method and a system for controlling the instant congestion of a server. In the method, controlling state comprises a first state, a second state and a third state, in each state, a data packet is processed by different methods under different conditions. The method comprises the following steps of: 1) receiving a data packet, then discarding or forwarding the data packet according to the controlling state; and 2) comparing the average request return time delay of the data packet with a detection threshold value, then changing controlling state or keeping an original state according to a comparison result, when the average request return time delay is greater than the detection threshold value, switching from the first state to the second state, or switching from the second state to the third state, and when congestion condition is improved, switching from the third state to the first state. By the method and the system, session integrity can be protected, the retransmission of massive data packets caused by the oscillation of server load can be avoided, and learning or training is not required and computation cost is low.

The present invention discloses a method and apparatus for determining, processing and requesting information, in the above method, determining whether the user surface data carried by MBMS is congested or overloaded, configuring the first indication; The first indication is sent to GCS AS, wherein the first indication is used to provide a reference basis in the process of GCS AS requesting MCE or eNB to stop the cluster traffic hosted by MBMS due to congestion or overload of user plane data. According to the technical solution provided in the present invention, it is possible to alleviate MBMS carrying congestion / overload, thereby reducing the impact on non-congestion / overload area UE.

The invention relates to an automatic summarization generating method and device. The method comprises the steps of modeling a text to generate a sentential network diagram which comprises edge weights; computing first weights of sentences in the text according to the edge weights and initial weights of the sentences in the text; computing term weights according to frequencies of terms in the sentences and the first weights of the sentences corresponding to the terms; computing second weights of the sentences according to the frequencies of the terms in the sentences and the term weights of the terms; computing final weights of the sentences according to the first weights and the second weights; and if the difference between the final weight after nth iterations and the final weight after (n-1)th iterations is less than a preset threshold value and the difference between the term weight after nth iterations and the term weight after (n-1)th iterations is less than a preset threshold value, generating an automatic summarization. By means of the method, the problem that the quality of single long text summarizations is low is solved; the information overload is alleviated; and the summarization quality is improved.