9results about How to "Avoid service interruption" patented technology

The invention provides a conference room water cup water replenishing strategy optimization method and system based on reinforcement learning, and relates to the technical field of intelligent services, and the method comprises the steps: obtaining the liquid remaining amount feature and the spatial position feature of a container in a target space, and constructing a state representation vector to represent the incidence relation between the two features; predicting the liquid consumption trend of each container in the target time window through the time sequence analysis model to obtain a demand prediction result; inputting the state representation vector and the demand prediction result into a reinforcement learning strategy network, and based on the multi-container liquid residual amount distribution state and the prediction consumption trend, generating a decision sequence containing a supplement priority and a supplement opportunity by optimizing a long-term accumulation reward function; and dynamically adjusting the supplementary execution sequence of each container according to the supplementary priority in combination with the available supplementary resource constraint, and generating and executing a supplementary execution scheme. According to the invention, intelligent optimization of the water replenishing strategy is realized through reinforcement learning, and the resource utilization efficiency and the service quality are improved.

The invention provides a main node election method of a multi-node network, the multi-node network and electronic equipment, and relates to the technical field of wireless communication. The method comprises the following steps: if a current main node judges that a preset election triggering condition is met, broadcasting an election instruction containing a score calculation rule to the whole network; scoring response data including node identifiers and energy consumption scoring values returned by the nodes are received, the energy consumption scoring values are scoring values calculated by the nodes according to energy consumption parameter sets of the nodes on the basis of scoring calculation rules, and the energy consumption parameter sets include the remaining battery capacity, the historical energy consumption rate and the data load; sorting all the nodes according to all the energy consumption score values so as to determine a new main node; and generating and broadcasting election result data containing the node identifier of the new master node so as to trigger the network to perform master-slave role switching and network synchronization according to the data. The scheme can solve the problems of unbalanced energy consumption and insufficient adaptability, so as to effectively prolong the service life of the network and improve the reliability of the network.

The invention discloses a quantum key distribution method and a quantum key distribution system. The quantum key distribution method comprises the following steps: in response to a received service demand sent by a first service terminal, determining a quantum key distribution target scheduling path according to a first network state and the service demand; and issuing a quantum key scheduling instruction to the target ground quantum key distribution network and / or the target non-ground quantum key distribution network based on the target scheduling path. Thus, through overall planning and scheduling of global resources, unified management and optimal configuration of global quantum key distribution resources are realized, the transmission distance bottleneck that a single ground quantum key distribution network is limited by optical fiber loss is effectively broken through, and the inherent shortages that the coverage range is fixed are effectively broken through; the service requirement that a cross-regional ground quantum key distribution network is difficult to touch is met, and the overall utilization rate of key resources and the satisfaction rate of user requests are improved.

This application provides an environment deployment method, apparatus, device, storage medium, and program product, which can be applied to the field of big data technology. The method includes: in response to an environment deployment instruction, performing a rolling deployment operation on N preset containers for the same service, where N is a positive integer greater than 1; wherein the rolling deployment operation includes: for any of the preset containers, determining a target container among the N preset containers, wherein the target container is used for service shutdown, and the other preset containers among the N preset containers (excluding the target container) are used to provide the service; performing version deployment on the target container to obtain a deployed target container; performing verification on the deployed target container; and, if the deployed target container passes verification, bringing the deployed target container online.

The application discloses a high-availability and anti-split-brain industrial data acquisition system, comprising: a basic server cluster, which comprises a business server cluster and a history server cluster, the business server cluster comprises a plurality of business instances, the business instances are used for undertaking business logic processing tasks, the business instances are separately deployed from a database master node, and at least one special master node is included in the plurality of business instances; and an acquisition server cluster, which is used for acquiring real-time data from field devices or IoT terminals; wherein the business server cluster is based on an election mechanism of a Redis cluster, and when network partitioning occurs, an external coordinator is used for judging and forcibly closing the write permission of a non-master node. Compared with the prior art, the business service and the database master node are separately deployed, and a three-node anti-split-brain mechanism is introduced, so that the stability and data consistency of the system in a complex network environment are remarkably improved.

The invention discloses a prediction content cache scheduling system based on machine learning. The prediction content cache scheduling system comprises a content feature perception module, a user demand prediction module, a dynamic cache decision module, a cache synchronization optimization module, a tracing audit module and an emergency guarantee module. According to the method, the content feature sensing module collects three types of core features of cloud game attributes, operation dependency and user interaction through multi-dimensional feature support, Min-Max normalization and one-hot coding standardization processing are carried out, a high-quality data set with a unified structure and a clear hot degree and delay sensitivity scoring rule are output, game priority quantification is controllable, and the game quality is improved. Decision subjectivity is avoided, a fusion model is constructed through a time sequence diagram neural network and a gradient boosting decision tree, association of regions and time accessed by players is captured, a nonlinear demand rule is mined, the prediction precision is remarkably improved after weighted fusion, a high-frequency iteration mechanism can quickly respond to sudden demands, and the problem of resource supply and demand mismatching is solved.

The invention provides a terminal scheduling method, device and equipment and a readable storage medium, and the method comprises the steps: collecting the operation information of a terminal under the statistical dimensions of state, business, queue occupation and the like, and determining a comprehensive abnormal score of the terminal based on the matching condition of the operation information and a preset abnormal condition; according to the method, abnormal occupation of a terminal on shared communication resources can be accurately identified without depending on independent queue hardware of each STA with high cost, scheduling intervention is implemented on an abnormal terminal, and the scheduling efficiency is improved on the premise of not increasing hardware queue resources. The situation that a single abnormal terminal occupies too many packet sending queue resources, and consequently forwarding of the whole machine is blocked or stuck is effectively avoided, and the system robustness, the resource utilization efficiency and the multi-user fairness are improved. The method can be applied to various IP network environments and various network topology structures, including optical fiber access scenes, such as fiber to the room (FTTR) networking scenes.

The invention relates to the technical field of intelligent cabinet management, and discloses an automatic replenishment method for an intelligent cabinet, and the method comprises the steps: firstly carrying out the structural processing of an original signal flow and SKU main data, so as to obtain a refined inventory list containing the batch, expiry date and other information; then, on the basis of historical consumption data, key index calculation is carried out on each batch in the list, so that information such as a dynamic consumption rate is obtained; secondly, quantifying dimensions such as inventory quantity, shelf life and the like to obtain health degree sub-scores of each dimension; fusing the sub-scores into a unified inventory health degree score through a preset synthesis algorithm; and finally, differential replenishment decisions are made according to the unified health degree score, and a replenishment instruction matched with the actual risk level is generated. Thus, by constructing the multi-dimensional inventory health assessment model, accurate, dynamic and prospective management of the inventory state is realized, and the accuracy and operation efficiency of automatic replenishment are significantly improved.

The embodiment of the invention provides a communication link switching method and device, equipment and a medium, and the method comprises the steps: building a second communication link with terminal equipment based on a second communication protocol after building a first communication link with the terminal equipment based on a first communication protocol; determining data characteristics of data communicating based on the first communication link, and determining a migration strategy of the data according to the data characteristics of the data; executing data migration from the first communication link to the second communication link according to the migration strategy; and after data migration is completed, when the second communication link meets a link stability condition, disconnecting the first communication link. According to the embodiment of the invention, when data migration is carried out, the dual communication links are kept to act together, and the migration strategy is formulated based on the data characteristics, so that service interruption in a network switching process is effectively avoided, the first link is disconnected after the second communication link is ensured to be stable, the service continuity and the control reliability are ensured, and the user experience is improved.