351 results about "Load scheduling" patented technology

The invention relates to the technical field of server resource scheduling, in particular to a server resource scheduling system integrating AI and edge computing. The server resource scheduling system integrating the AI and the edge computing comprises a computing node module, a resource calling module, a task traversal module, a demand prediction module, a resource allocation module and a load scheduling module, the computing node module is used for monitoring the real-time running state of a server computing node and performing distributed node architecture fitting, and the load scheduling module is used for scheduling the load of the server computing node; constructing a distributed server node network; the resource calling module is used for acquiring a server running log; performing server resource calling behavior analysis on the server running log, performing resource calling time sequence characteristic evolution, and generating a server resource calling time sequence characteristic; and the task traversal module is used for carrying out unprocessed task traversal identification on the server running log and constructing an unprocessed task state mapping sequence. According to the invention, efficient and load-balanced server resource scheduling is realized.

The present invention relates to the technical field of new energy vehicles, and specifically to a load balancing and scheduling method for charging piles of new energy vehicles based on intelligent algorithms, comprising the following steps: S1, data collection: real-time collection of operating data of charging piles; S2, charging demand prediction: prediction of the charging demand of each charging pile within a predetermined time in the future through a graph convolutional network model; S3, charging pile spatial scheduling optimization: spatial optimization of the scheduling of charging piles according to the geographical distribution of charging piles, the flow of surrounding vehicles and the results of charging demand prediction; S4, load scheduling and resource allocation: based on the results of charging demand prediction and charging pile spatial scheduling optimization, combined with the load of the power grid, resource allocation is performed. The present invention improves the utilization efficiency of charging pile resources, optimizes the layout of the overall charging network, and ensures the fairness of charging services and the balance of power grid load.

The invention provides a task load scheduling optimization method and system for distributed network nodes, and the method comprises the steps: determining a task transmission path between the distributed network nodes according to historical task congestion data and a real-time link load prediction result; adjusting a modulation coding strategy in a task transmission process according to the channel quality parameter on the task transmission path to obtain an adjusted modulation coding strategy; according to an energy consumption demand corresponding to the adjusted modulation coding strategy, combining load distribution of each communication link and each node on the task transmission path to generate a voltage frequency adjustment instruction corresponding to the distributed network node; performing collaborative optimization on the task transmission path, the channel quality parameter and the voltage frequency regulation instruction to generate a task load scheduling strategy; according to the invention, through collaborative optimization of dynamic path selection, adaptive modulation coding adjustment and precise energy consumption regulation, the transmission reliability, energy efficiency balance and system stability of the distributed network in a high dynamic environment are improved.

The invention discloses a power system load dynamic optimization method based on reinforcement learning. The method comprises the following steps: S1, collecting power system data to construct a state space; s2, constructing a hierarchical reinforcement learning model based on the state space, and dividing a high-level decision and a low-level execution task; s3, training a high-level decision model, and outputting a scheduling task target category instruction in a high-level state; s4, training a low-layer execution model, and outputting a control action in combination with a current node state and a high-layer instruction; s5, introducing an evolutionary mechanism to generate a strategy population and optimizing a low-layer execution model; s6, fusing an evolutionary mechanism and a strategy gradient to synchronously optimize individuals with excellent performance; s7, deploying the trained model to a power dispatching system; s8, performing model parameter fine tuning based on scheduling feedback; and S9, continuously applying the fine-tuned model to load scheduling control. According to the invention, power load accurate scheduling and strategy efficient adaptive optimization are realized, and system responsiveness and operation stability are improved.

The invention discloses a green data center scheduling optimization method under computing power and electric power coupling, data load types of a data center comprise a transferable load, a reducible load and a rigid load, and the method comprises the following steps: constructing a double-target scheduling optimization model by taking the minimum power consumption cost of the data center and the maximum green power fitting degree as targets; the power consumption cost of the data center is used for representing the sum of expenses generated by the data center for purchasing power from a power grid in order to meet the power required for processing the data load in each time period, and the green power fitting degree is used for evaluating the matching degree between the power consumption demand of the data center and the renewable energy source supply; solving the dual-target scheduling optimization model by using a multi-target dogvessel swarm algorithm to obtain a Pareto frontier; and determining an optimal scheduling scheme based on the Pareto frontier, wherein the optimal scheduling scheme comprises a data center power consumption curve, an energy storage power station charging and discharging curve and a transferable load scheduling strategy. According to the invention, the power consumption cost of the data center can be reduced, and the renewable energy consumption capability can be improved.

The invention discloses an emergency demand response resource optimal allocation method under a hierarchical aggregation architecture, which comprises the following steps of: by constructing the hierarchical aggregation architecture, dividing flexible resources into a plurality of regions according to geographic positions and communication characteristics, and enabling a region coordination center to be responsible for local power allocation and load scheduling in each region; a power grid dispatching center is responsible for coordinated management of global demand targets, dispatching complexity is remarkably reduced, real-time performance is improved, an optimization model adopts a multi-target optimization method, power response cost, response reliability and time delay penalty are integrated, and through power balance, capability constraint and time constraint, accuracy and economical efficiency of an allocation scheme are ensured, and the allocation efficiency is improved. According to the method, rapid adjustment of resource state abnormity, response deviation or node faults is realized through dynamic adjustment and a distributed consensus algorithm, the surplus capacity in the region is preferentially utilized to compensate the insufficient demand, and redundant power support is scheduled through an inter-region coordination mechanism when necessary.

The invention belongs to the technical field of power system load prediction, and discloses a cross-dimension multi-scale fusion load prediction method based on multi-user load time-space correlation, which comprises the following steps of: firstly, preprocessing user load statistical data, extracting time sequence dependence and periodic characteristics in a time sequence, and calculating the time sequence dependence and periodic characteristics of the user load statistical data; introducing a channel attention mechanism to adaptively mine key variable information; then, a multi-scale space-time fusion module is combined with frequency domain analysis and a graph convolutional network to realize depth feature interaction under different time scales and space levels; and finally, outputting a load prediction result under a plurality of time granularities in the future through a linear projection structure. Compared with an existing method, the method has the remarkable advantages in the aspects of capturing a complex load mode, improving model prediction precision and enhancing generalization ability, and is suitable for various application scenes such as power consumer energy consumption management and power grid load dispatching.

The invention discloses a power grid data acquisition and analysis system based on big data, particularly relates to the technical field of power network monitoring, and is used for solving the problem of key feature loss caused by cross-hierarchy data semantic association missing of an existing hierarchical aggregation mechanism. Performing standardized filling and time label layered alignment processing on the original data through the equipment acquisition module; the feature extraction module screens an abnormal waveform fragment and a steady-state parameter offset based on the equipment type and the real-time fluctuation amplitude; the semantic association module generates a space-time association weight in combination with the regional physical topological relation and the reactive circulation path sensitivity; the feature clustering module reconstructs a region-level feature aggregation packet through kernel density estimation and spatial weighted fusion; the cross-layer analysis module dynamically corrects the aggregation weight based on the current flow direction and the transient energy distribution; and the decision generation module generates an equipment positioning instruction and load scheduling strategy combination through alarm template matching, and finally realizes accurate positioning of fault equipment and generation of a scheduling strategy.

The invention relates to the technical field of power systems, and discloses an industrial park distributed power utilization network simulation analysis power utilization behavior evaluation method comprising the following steps: S1, data acquisition: power load data, equipment operation state data and environment data of each power utilization unit in an industrial park are acquired in real time, and the environment data comprise temperature and humidity; s2, data preprocessing: preprocessing the collected data to obtain complete data; and S3, simulation modeling: establishing a power load flow model based on the park information and the real-time data, and performing simulation modeling to generate a dynamic model under various power consumption situations. According to the invention, through real-time data acquisition, dynamic load scheduling and a deep reinforcement learning algorithm, flexible optimal distribution of electric power resources is realized, the efficiency and stability of park electric power management are improved, stable operation of equipment is ensured, and energy waste is reduced.

Provided in embodiments of the present disclosure are a method for scheduling workloads of a data center in an electricity market environment, and a device. In the embodiments of the present disclosure, a multi-stage random optimization solution that allows a data center to participate in a multi-level electricity market is provided. The types and data dependency relationship of workloads are taken into consideration, and a load scheduling model, in a time dimension, for the workloads is constructed; on the basis of electricity source information of the data center, a multi-stage electricity cost optimization model is constructed, so as to comprehensively consider the exogenic uncertainty of an electricity price and carbon emissions and the endogenic uncertainty of the workloads; on the basis of the multi-stage electricity cost optimization model, multi-stage optimization solving is performed on the load scheduling model, so as to obtain a load scheduling scheme and an electricity purchase strategy. Therefore, a more accurate and more reliable estimation is made concerning the electricity flexibility of the data center, such that the electricity cost of the data center can be reduced, support is provided for the reduction of carbon emissions, and the consumption of renewable energy sources is encouraged, thereby achieving the energy use flexibility of the data center in the multi-level electricity market environment.

The invention belongs to the technical field of intelligent power grids, and discloses an intelligent power grid load balance control method and system. The method comprises the steps of dividing a power grid into a plurality of regions based on a predetermined rule; based on the regional power supply quantity and the regional power consumption quantity, identifying an overload region with insufficient power supply quantity; formulating and executing a load balancing strategy for the overload area by adopting a natural heuristic optimization algorithm, wherein the load balancing strategy is a load scheduling strategy taking scheduling time minimization as a target; calculating the priority coefficient of the electric equipment in the overload area, and supplying power to each electric equipment according to the priority coefficient; effective balance and optimal allocation of intelligent power grid loads are achieved, emergency power supply is guaranteed to the maximum extent in natural disaster time, and therefore the power grid operation level and livelihood guarantee capacity are improved, and the power supply reliability and the power utilization efficiency of the power grid are improved.

The invention discloses a virtual power plant building key load priority distribution method and system, and relates to the technical field of load dispatching management. The virtual power plant building key load priority distribution method comprises the following steps: S1, acquiring load switching scheduling data, and preprocessing the load switching scheduling data; s2, identifying an executable switching window between the main power supply and the standby power supply, quantifying a switching impact index of a load, and screening a candidate load set; s3, evaluating the switching priority of the load and generating a sorting list, and sequentially selecting according to the upper limit of the residual impact capacity of the bus to form a current period execution plan; and S4, executing a scheduling instruction, recording operation parameters before and after switching in real time, constructing a disturbance evaluation input set, evaluating a load switching disturbance degree, judging a switching state based on an evaluation result, and triggering a feedback mechanism to adjust a subsequent sorting weight and an impact tolerance. The problem of bus impact caused by concurrent access of multiple loads in the main and standby power supply switching process is solved.

The invention provides a flexible load-oriented day-ahead intra-day inter-group secondary cooperative regulation and control method, and belongs to the technical field of power flexible load scheduling control. The problem that active participation of user side flexible resources in power grid regulation and control is uncertain is solved; comprising the following steps: S10, obtaining related characteristic parameters and historical electric quantity data of user side flexible resources; s20, establishing a flexible resource evaluation index, dividing flexible resources into different flexible load clusters through an improved K-means clustering method, and forming a layered regulation and control framework of flexible loads after aggregation; s30, in combination with a day-ahead load prediction method, calculating the maximum adjustable capacity of user side flexible resources, and in combination with a parallel Markov chain model, preliminarily making a flexible load day-ahead scheduling plan; s40, tracking the load power change of the user side in the day in real time, and correcting the adjustment capacity in time based on a flood filling algorithm to realize secondary cooperative regulation and control of the flexible load cluster; the method is applied to a power system.

The invention provides a network load balancing algorithm based on a credential platform and a credential terminal, and relates to the technical field of credential, and the algorithm comprises the following steps: collecting real-time traffic data from a plurality of ports; carrying out preprocessing of normalization, abnormal value elimination and protocol field standardization on the collected data; performing multi-dimensional protocol fingerprint feature extraction on the flow of each port based on a protocol analysis algorithm; modeling historical and real-time traffic of each port by using a time sequence neural network, and outputting a traffic prediction value and abnormal traffic probability assessment in a future short time period; dynamically generating a multi-port distribution weight factor based on the protocol characteristics and the flow change trend of each port; inputting the prediction result and the real-time load state into a reinforcement learning scheduling agent; upon detection of an abnormal traffic trend, feed-forward load distribution adjustments are performed immediately. According to the method, the data processing efficiency can be improved, the load scheduling intelligence is realized, the resource allocation priority of the port of the credential terminal is ensured, and the service continuity is guaranteed.

The invention discloses an energy consumption optimization method and system for a computing power server, and relates to the technical field of server optimization, and the method comprises the steps: monitoring the operation data of the computing power server in real time, and recognizing the task and load conditions; constructing a load prediction model to predict the future load of the computing power server, and carrying out the energy consumption optimization of the computing power server according to a predicted load scheduling task; continuously optimizing the load prediction model based on a reinforcement learning algorithm; and displaying the load condition of the computing power server in real time, and storing the operation data of the computing power server to a database. The load prediction model is constructed by collecting the processing tasks and the load data of the computing power server to predict the future load, the accuracy of load prediction is improved, task processing scheduling is performed by predicting the future load to achieve load optimization of the computing power server, and load energy consumption adjustment of the processing tasks of the computing power server is effectively achieved.

The invention discloses a multi-model dynamic image analysis system, and relates to the technical field of artificial intelligence. The system adopts a heterogeneous computing architecture and comprises a preprocessing model, a multi-modal analysis model, an anisotropic correction model, a large language model, a task knowledge base embedding model and a KV cache management system. An incongruous task reconstruction mechanism based on hardware isomerism is constructed, a large language reasoning model is nested for analysis, and efficient reuse of computing resources is achieved by using a cross-model KV cache optimization technology. The system is provided with multiple graphics cards and a double-channel NUMA, and a multi-mode large language model carries out unsupervised or weak supervised training on a target detection model in the operation of the system. According to the system, through a dynamic load scheduling strategy, a mechanism for reasoning by using a processor and a memory enhanced display card in a full-load mode is innovatively designed, and three-section real-time streaming enhanced output is achieved. According to the system, through a dynamic load scheduling strategy, a mechanism for reasoning by using a processor and a memory enhanced graphics card in a full-load mode is innovatively designed.

The invention relates to the technical field of distributed computing, and discloses a heavy truck charging load scheduling-oriented distributed energy intelligent matching method, which comprises the following steps of: establishing a charging service computing power mapping table at a scheduling node; maintaining a shadow counter in a local memory, extracting a pre-estimated computing power consumption value according to an event type and accumulating the pre-estimated computing power consumption value to the shadow counter, and executing linear numerical deduction on the shadow counter according to a reference logic subtraction rate so as to simulate a scheduling data throughput evolution process; adjusting a linear deduction rate parameter according to the deviation between the state feedback data and the numerical value of the shadow counter; and distributing the charging matching task to a charging station edge computing node of which the shadow counter value does not exceed a preset logic saturation threshold value. According to the invention, through an open-loop estimation and closed-loop calibration mechanism of a local logic state, an instantaneous congestion risk caused by physical feedback lag is eliminated; and logic state consistency and self-adaptive distribution of the whole network computing power resources are realized.

The invention relates to the technical field of energy storage load scheduling, in particular to an energy storage load scheduling method and system based on multi-objective optimization, and the method comprises the steps: obtaining historical electricity price, load power and renewable energy power generation power data of each time period before the current time as input variables, and carrying out the normalization processing to obtain input data; inputting the input data into an LSTM-Transformer prediction model, and predicting to obtain the power grid electricity price, the load demand and the renewable energy power generation amount in a future time period; establishing a multi-objective optimization model including electric charge cost, operating profit and energy storage life; substituting the prediction data into the multi-objective optimization model, and solving a Pareto frontier based on a multi-objective optimization algorithm to obtain a non-inferior solution set satisfying the multi-objective optimization model; and taking each solution in the non-inferior solution set as input, and screening the solution with the maximum Sharpley value as an optimal energy storage charging and discharging adjustment strategy. By accurately controlling the working state of the energy storage system, the interactive operation mode of the energy storage system and the power grid is optimized.

An urban area optimization control system and method based on phase balance of distribution network, the method includes: obtaining distribution network operation data in real time through intelligent sensors and measurement and control devices deployed at key nodes of distribution network; performing load forecasting based on the collected distribution network operation data and clustering loads through clustering algorithms, and evaluating load status based on load clustering results and load results; taking into account power grid topology, distribution network operation data, load forecasting results and load evaluation results, adopting multi-agent collaborative optimization to determine the optimal load dispatching instructions; converting the obtained optimal dispatching instructions into actual control actions based on phase balance to optimize the power grid. The present invention can reduce the three-phase imbalance of the distribution network from 23.6% to 8.4%, effectively improving the power supply quality and operation efficiency.

Systems and techniques that facilitate dynamic load scheduling are provided. For example, one or more embodiments described herein can comprise a system, which can comprise a memory that can store computer executable components. The system can also comprise a processor, operably coupled to the memory that can execute the computer executable components stored in memory. The computer executable components can comprise a carbon footprint component that generates estimated carbon footprints of a plurality of tasks for a plurality of datacenters based on carbon intensity forecasts and estimated workload requirements of the task, wherein the carbon intensity forecasts comprise ancillary carbon output, marginal carbon output, and average carbon output; and a scheduling component that schedules the plurality of tasks for one or more datacenters selected from the plurality of datacenters based on the estimated carbon footprints.

The invention relates to the technical field of hydropower station intelligent monitoring and dispatching, and discloses a hydropower station computer monitoring method based on domestic hardware, which comprises the following steps: dynamic system modeling: establishing nonlinear dynamic models, including water pump and generator power models, of equipment in a hydropower station, and describing the operation state of the equipment; the multi-objective optimization scheduling is used for calculating load scheduling of equipment based on a multi-objective optimization model of power consumption, equipment loss and system energy efficiency; and the fault prediction and early warning module is used for carrying out real-time prediction on the equipment state through Kalman filtering, carrying out fault mode prediction based on historical data and sending out fault early warning in time. According to the invention, through dynamic system modeling, multi-objective optimization scheduling, Kalman filtering fault prediction and H-infinity robust control, efficient scheduling, accurate fault early warning and robust control strategy optimization of hydropower station equipment are realized, and the operation efficiency and stability of the system are significantly improved.

The present invention discloses a method for flexible load control of distribution network substations under new energy access, comprising the following steps: S1, aggregation modeling of typical flexible loads of users in the substation, wherein the typical flexible loads include electric vehicle charging and discharging models, temperature control load power consumption models and energy storage equipment charging and discharging models; S2, flexible load dispatching of users in the substation for regional load aggregators; S3, joint dispatching of multiple flexible load groups of users in the substation; S4, day-ahead dispatching of distribution network substation loads with multi-agent game; S5, real-time control of substation loads with uncertainty of source, load and storage based on day-ahead dispatching; the present invention realizes active distribution network optimization dispatching based on flexible loads on multiple time scales, implements the use of flexible loads to smooth fluctuations of renewable energy, and solves the problem of difficulty in real-time dispatching caused by uncertainty of source, load and storage.

The invention discloses a space-time price-based computing power and electric power space-time collaborative flexible scheduling method and device, and relates to the technical field of electric power systems, and the method comprises the steps: obtaining the architecture and operation characteristics of a data center, pre-establishing an energy consumption model and a load scheduling model of the data center, and pre-establishing a computing power-electric power collaborative regulation double-layer optimization model, the calculation power-electric power coordinated regulation double-layer optimization model is based on an energy consumption model and a load scheduling model of the data center, takes the initiative of the data center as an independent operator and the influence of the load of the data center on the electricity price of the system into consideration, and is constructed by taking the data center as a decision-making main body; a lower-layer power grid direct current optimal power flow problem is converted into a KKT condition to serve as a constraint to be substituted into an upper-layer problem, a strong dual theorem and a Big-M method are utilized to linearize and solve the double-layer optimization model of computing power-power coordinated regulation, and a computing power-power space-time coordinated flexible scheduling result is obtained.

The invention relates to the technical field of power load scheduling, and discloses a power equipment load scheduling method and system based on big data analysis. The method comprises the following steps: predicting a power load at a future moment in a selected dispatching area through a multi-feature time sequence machine learning model, judging whether a power load insufficient supply prompt is generated or not, and judging whether an available standby power supply exists or not by combining power supply parameters of the standby power supply according to the power load insufficient supply condition; performing power load scheduling on the electric quantity in the standby power supply according to the scheduling duration; compared with the prior art, the potential safety hazard of power load insufficiency possibly occurring in the selected dispatching area in the future can be found in time in advance, and power load dispatching remedial measures are taken for the power load insufficiency, so that the situation that the power load in the selected dispatching area increases instantly at a certain moment is avoided. Therefore, the negative influence caused by the lagging of the power load dispatching is avoided.

The invention discloses a glass factory load low-carbon regulation and control method considering capacity compensation and carbon quota, and aims to realize refined carbon emission control and flexible scheduling in the production process of a glass factory. By fusing multi-dimensional input of historical load power, load change rate, environment temperature, humidity, wind speed, electricity price, time and the like and utilizing an STFormer architecture to model a complex time dependency relationship, a kilowatt-hour electricity carbon emission factor in a future time period is predicted, and prior information is provided for low-carbon scheduling; with minimization of carbon emission and electric charge cost and maximization of capacity income as targets, a future load scheduling strategy is optimized under constraints of carbon quota, an electric power regulation range and equipment operation; and globally searching an optimal load control scheme through a quantum particle swarm algorithm, and generating a multi-scene-oriented dynamic regulation and control strategy under the constraints of a multi-target fitness function and a penalty mechanism. The method not only improves the precision and robustness of carbon factor prediction, but also realizes flexible scheduling of the glass factory load driven by carbon economy.

The present invention discloses a Hadoop cluster task scheduling method based on a positive and negative feedback load scheduling algorithm, which is applied to a Hadoop resource allocation center. The cluster task scheduling method comprises: an idle computing node applies to a management node for a task execution request; the management node receives the task request and queries whether there is a historical record of running the task request; the historical records are compared and analyzed, and if the task configuration meets the resource requirements, the task is calculated in this computing node; otherwise, the computing node is re-matched with a new task request; and computing tasks can be dynamically allocated in cluster task scheduling according to the resource usage of different servers and the differences in server performance, thereby achieving efficient completion of computing tasks and improving CPU and resource utilization efficiency.

The invention provides a large language model application workload scheduling method, system and device, and relates to the technical field of model load scheduling, and the method comprises the steps: modeling a composite large language model application into a directed acyclic graph comprising a conventional stage, an LLM stage and a dynamic stage; modeling execution correlation among stages in the directed acyclic graph through a Bayesian network, dynamically predicting duration distribution of uncompleted stages, and calibrating a duration estimated value of an LLM stage in combination with a real-time batch processing size of an LLM executor; the uncertainty reduction amount of each ready stage is quantitatively scheduled based on information entropy; an epsilon-greedy strategy is adopted, and a JCT priority queue and an uncertainty reduction priority queue are combined to allocate scheduling resources; and assigning the task to a corresponding executor for execution, and repeating the above process until all operations are completed. The technical problem that the scheduling technology in the prior art is difficult to effectively deal with the execution time uncertainty and the structure uncertainty of the composite LLM application is solved.

The invention discloses a cloud-side collaborative task allocation method and device based on task difficulty analysis, terminal equipment and a storage medium, and belongs to the technical field of load scheduling, and the method comprises the steps: obtaining a calculation amount index, a data amount index and a real-time requirement index of a to-be-allocated task, and carrying out the calculation to obtain task complexity; and according to the task complexity and the terminal computing power index of the terminal where the task to be allocated is located, whether the task should be independently processed by the terminal equipment or cooperatively processed by other terminals in the cloud server or the local area network is judged. According to the method, the task complexity is considered in the task allocation process, particularly the calculated amount index, the data volume index and the real-time requirement index of the task are concerned, and the cloud edge collaborative task allocation result obtained by implementing the method better meets the calculated amount condition, the data volume condition and the real-time requirement of the to-be-allocated task; the problem that the cloud edge collaborative task allocation result is low in practicability due to the fact that the task difficulty cannot be effectively perceived can be solved.

The invention discloses a microgrid storage load coordination control optimization method and system, and relates to the technical field of power grid control, and the method comprises the steps: analyzing a target historical generating capacity record of a target microgrid, and obtaining a predicted generating capacity; analyzing the target historical electricity utilization record to obtain a target electricity utilization type set; performing correlation analysis on the power consumption time and the power consumption types to obtain a correlation result, and establishing a schedulable power consumption type set; calling a load prediction plan to obtain a predicted electrical load; and comparing the predicted generating capacity with the predicted electrical load to obtain a predicted comparison result, and generating a coordination control plan. According to the method, the technical problems of low energy utilization efficiency and operation stability of the micro-grid caused by inaccurate power generation capacity and load prediction and insufficient energy storage and load scheduling optimization in the prior art are solved, and coordinated optimization control of energy storage and load of the micro-grid is realized; and the energy utilization efficiency and the operation stability of the micro-grid are improved.

The invention provides a database query prediction and load scheduling method and device. The method comprises the following steps: obtaining a plurality of query tasks corresponding to a database system; respectively inputting the plurality of query tasks into a preset memory prediction model, and obtaining memory prediction results which are output by the memory prediction model and respectively correspond to the plurality of query tasks; the memory prediction result is a memory demand prediction value; performing load scheduling on the plurality of query tasks based on the memory prediction result and a dynamic scheduling strategy to obtain load scheduling information, and performing memory allocation processing on the plurality of query tasks based on the load scheduling information; wherein the dynamic scheduling strategy is a load scheduling strategy based on memory perception of the database system; the load scheduling information comprises a query execution sequence of a plurality of query tasks. According to the method provided by the invention, the memory prediction precision and the load scheduling efficiency of the query task can be effectively improved.