1074 results about "Demand forecasting" patented technology

The invention relates to the technical field of cloud side end computing, and discloses a cloud side end cooperative computing resource dynamic allocation method based on task-aware migration. The method comprises the following steps: acquiring real-time load characteristics and resource demand characteristics of calculation tasks in a cloud side end system, and dividing task priority queues in combination with task type identifiers; extracting historical execution records of tasks at cloud, edges and terminal nodes, constructing a task execution feature library, and generating a resource demand prediction model in combination with real-time load features; analyzing network transmission time delay characteristics of a cloud end and edge nodes, measuring real-time calculation capability fluctuation data of terminal equipment, and establishing an inter-node resource collaboration degree evaluation matrix; generating an initial migration strategy according to the prediction model and the evaluation matrix, monitoring actual resource occupancy deviation of the task, forming a final decision in combination with a node resource state correction strategy, triggering cross-node migration, and synchronously updating the priority queue and the evaluation matrix.

The invention discloses a commercial building energy monitoring and intelligent control method and device and a storage medium, and belongs to the technical field of building intelligent control, and the method comprises the steps: collecting data, building a nonlinear mapping relation, and generating an energy consumption demand prediction tensor; injecting an adversarial disturbance sample, and evaluating the robustness of the prediction model; in combination with the energy consumption baseline, performing cross confirmation and correction on the prediction data exceeding the threshold value; performing attribution analysis on the corrected energy consumption sequence to generate an energy consumption attribution map; adjusting the solution of a multi-objective optimization function according to the atlas, and generating an optimal cooperative control strategy; and the comprehensive efficiency is used as a reinforcement learning reward, strategy parameters are iteratively updated, and a control knowledge base is formed. According to the method, a closed-loop control framework integrating robust demand prediction, dynamic attribution analysis, collaborative optimization decision and a self-evolution strategy is adopted, intelligent regulation and control of building energy consumption can be realized, and the long-term adaptive optimization capability is improved.

The invention discloses a predictive order dispatching system based on digital twinning, and relates to the technical field of intelligent order dispatching, the predictive order dispatching system comprises four core modules: an urban transport capacity digital twinning construction module integrates order dispatching platform historical orders, real-time traffic flow and urban GIS data, constructs and updates a twinning model, and divides standard grid units; a short-term order demand prediction module obtains grid order, traffic and environment feature sequences based on the model, and predicts order quantity and thermodynamic diagrams in 15-30 minutes in the future; a transport capacity supply and demand gap early warning module analyzes the gap and generates early warning; the active intervention and optimization module generates an instruction according to a preset strategy, dispatches transport capacity through order pre-dispatching and the like, and then evaluates and optimizes the strategy according to feedback. The system realizes order pre-judgment, accurate transport capacity allocation, reduction of supply and demand imbalance, shortening of user waiting time, reduction of operation cost, and improvement of delivery efficiency and service quality.

The invention relates to an AI-based supply chain demand prediction and inventory optimization method and system, and the method comprises the following steps: calling a sales record from a supply chain, recognizing a periodic demand rule of a corresponding commodity, and predicting an estimated demand quantity in a future time period according to the periodic demand rule; carrying out supply and demand difference analysis by combining the current warehouse commodity stock margin, and judging the stock state; if the commodity stock surplus exists, making a corresponding stock digestion strategy; if a shortage quantification result appears, carrying out key degree evaluation and urgency sorting based on a shortage amount in combination with market factors, and generating a key commodity list and an out-of-stock urgency index; and finally, in combination with a preset commodity constraint condition, according to the key commodity list and the urgency index, an optimal replenishment scheme meeting resource and time limit is formulated, and the technical problem of how to realize conversion from passive response to active pre-judgment in a supply chain environment with large demand fluctuation and complex influence factors is solved.

The invention relates to an enterprise multi-dimensional business index analysis data resource optimization method, a medium and equipment, and the method comprises the steps: collecting a multi-dimensional business index analysis request of an enterprise business system, carrying out the mode recognition and feature extraction of the multi-dimensional business index analysis request, and generating business entity features; and inputting the business entity features into the resource allocation model, outputting a priority score and a resource demand prediction result, calculating an optimal resource allocation scheme through an optimization decision model based on the priority score and the resource demand prediction result, and performing automatic resource scheduling according to the optimal resource allocation scheme to generate a business decision instruction. And finally outputting a business decision instruction and a data resource management report. The resource reuse potential among different requests is automatically identified and quantified by adopting a set similarity algorithm, so that the enterprise data resource utilization efficiency is remarkably improved, the redundancy calculation consumption is reduced, and the resource allocation is more reasonable and efficient; manual intervention is reduced, and enterprise data resource management efficiency and response speed are improved.

The embodiment of the invention provides an AI large model training resource dynamic scheduling method and device based on a VGPU, and effective planning of resources is realized through innovatively designing a resource prediction system and through index analysis and demand prediction. A multi-target scheduling mechanism is constructed, and a reliable scheduling strategy is established in combination with competition analysis and scheme screening. Virtualized isolation is introduced, and the training stability is ensured through VGPU and resource control. According to the method, the defects of the traditional technology in the aspects of resource prediction, scheduling optimization, virtualization isolation and the like are effectively overcome, and technical support is provided for large model training.

The invention relates to the technical field of network communication, in particular to a 5G-A network slice dynamic adjustment and resource allocation system oriented to low-altitude economy. The system comprises a mobility management feature integration module, a network slice service level protocol evaluation module, a mobility management demand prediction benchmark construction module, a multi-dimensional trend prediction module, a joint optimization decision generation module and a strategy verification and execution module, and is used for acquiring a mobility management feature data set of a wireless network; performing network slice service level protocol evaluation on the wireless network to obtain a network slice state evaluation result; constructing a mobility management demand prediction benchmark; performing multi-dimensional trend analysis on the network slice state evaluation result to obtain a prediction result; performing joint optimization decision on the standard network slice configuration strategy to obtain a network slice adjustment strategy; and executing the verified slice adjustment strategy to realize dynamic adjustment and resource allocation of the network slice. According to the invention, the efficiency of network slice dynamic adjustment and resource allocation can be improved.

The invention discloses an emergency rescue resource scheduling method based on the Internet of Things, and the method comprises the steps: collecting disaster area environment parameters, images and personnel distribution data through Internet of Things equipment, and associating historical disaster cases to construct a multi-source heterogeneous data set; extracting disaster characteristics through preprocessing, dynamically distributing data source weights, and generating a disaster assessment matrix; outputting a resource demand peak value based on the time sequence prediction model and performing dynamic correction; constructing a deep reinforcement learning model optimization decision strategy, and generating a scheduling scheme in combination with priority matching and an improved A * algorithm; according to the method, the disaster sensing precision and the demand prediction accuracy are improved, the scheduling decision adaptability is enhanced, the response time of a high-priority region is shortened, and the efficient demand of emergency rescue in a complex disaster is met.

The invention discloses a fused salt heat storage and Carnot cell combined coal-fired unit peak-load regulation operation method, device and equipment, a medium and a product, and relates to the field of unit peak-load regulation operation. Acquiring information data; based on a power regulation demand prediction model, performing peak regulation demand prediction according to the information data to obtain a prediction result; the power regulation demand prediction model is obtained by adopting a multi-objective optimization model and performing time sequence analysis on a long-short-term memory network model; determining an energy scheduling strategy based on the prediction result; the energy scheduling strategy comprises a heat storage power coordination instruction and a heat release power coordination instruction; and according to the energy scheduling strategy, operation adjustment is conducted on the coal-fired unit, the fused salt heat storage subsystem and the Carnot battery subsystem based on the control system. The invention aims to improve the peak regulation performance of the coal-fired unit.

The invention relates to the technical field of intelligent prediction, and discloses an emergency resource reserve intelligent prediction system and method based on a graph neural network, and the system comprises a space-time diagram construction module which collects port, berth, channel segment, emergency warehouse and ship information, passing time window information and emergency information, and constructs an event flow heterogeneous space-time diagram; the backbone screening module is used for screening backbone sub-graphs and setting a relation feature conversion rule and a time feature representation rule; the graph memory updating module is used for updating node and edge features; the multi-task prediction module is used for outputting quantile demand prediction, reachability scores, transportation timeliness and safety stock threshold values; the consistency constraint training module is used for executing consistency constraint training; the inventory allocation decision-making module is used for generating allocation distribution quantity and outputting predicted arrival time; and the berth time window recharging module is used for generating a berth operation time window and recharging the actual arrival time. According to the invention, emergency resource demand prediction, inventory threshold calculation and allocation and berth operation linkage decision are realized.

The invention relates to the crossing field of supply chain management and block chain technologies, discloses a block chain-driven circulating packaging ownership tracing and reuse optimization system, and solves the problems that existing circulating packaging ownership tracing is not credible, the reuse rate is low, data sharing is barrier and settlement is low in efficiency. The system comprises a sensing layer, a network layer, a block chain layer and an application layer, wherein the sensing layer collects and packages full life cycle data; the network layer carries out preprocessing and 5G data transmission; the block chain layer adopts an alliance chain to deploy ownership change, full life cycle records and privacy protection contracts; the application layer is internally provided with a core unit for XGBoost demand prediction and genetic algorithm scheduling, and functions of ownership tracing, multiplexing scheduling and the like are realized. Implementation shows that the ownership tracing credibility is 100%, the reuse rate is improved to 90% or above, the settlement time is shortened from 3 days to 1.5 hours, and the method is suitable for circulating packaging scenes such as cold chains.

The invention relates to the technical field of intelligent wharfs, in particular to a wharf container truck dynamic optimization scheduling method and system combining machine learning and path planning. Comprising a behavior data acquisition and feature coupling unit; a learnable incentive and behavior guide unit; a scheduling demand prediction unit; and a path planning and scheduling unit. According to the method, on the basis of the coupling characteristics, the excitation coefficient is optimized through reinforcement learning, the excitation instruction is dynamically pushed, and targeted guidance of the non-operation staying behavior of the container truck is achieved; according to the method, based on standardized time series data, an association rule of a historical staying period and a working condition is learned through an LSTM model, a prediction result is optimized in combination with real-time data, a prospective constraint basis is provided for scheduling, and meanwhile, a time, space, resource and priority multi-dimensional path constraint system is constructed through a structural causal model; container truck-berth matching and dynamic path planning are completed by matching with an improved A * algorithm fused with dynamic weights, and scheduling conflicts are effectively avoided.

The invention relates to the field of direct drinking water of a fusion calculation model, and discloses a double-membrane mixing direct drinking water system and an operation method.The method comprises the steps that online data of a water supply end and a water quality monitoring end are collected for cleaning and time alignment, and a basic working condition data set is formed; the method comprises the following steps: establishing an operation characteristic model comprising water demand prediction, raw water quality prediction and membrane flux characteristic sub-models, calculating double-membrane target water yield and water quality indexes based on the model to generate a control parameter set, evaluating membrane pollution to generate an online cleaning task queue, performing joint simulation to generate a pressure stability control instruction set, and dynamically correcting the model. The conversion from passive response to active pre-judgment of the double-membrane mixing direct drinking water system is realized, and the operation stability, the water quality guarantee capability and the water-saving and energy-saving effects are improved.

The invention discloses a mixed traffic flow-oriented automatic driving taxi dynamic scheduling system and a collaborative decision-making method, belongs to the field of intelligent traffic systems, and solves the problem that in the coexistence environment of manual driving vehicles and automatic driving taxies, the automatic driving taxies cannot be automatically scheduled. The technical problem of how to efficiently and cooperatively dispatch vehicles, accurately predict demands, optimize energy management and improve the overall operation efficiency of the system is solved. The system comprises a scheduling server which is connected with a road side sensing unit, a vehicle-mounted control unit and a charging station management platform. The scheduling server comprises a traffic flow analysis module; a demand prediction module; a dynamic scheduling module; and an energy collaboration module. The system is mainly used for realizing real-time, dynamic and intelligent scheduling and energy management of the automatic driving taxis in the mixed traffic flow, improving the operation efficiency, relieving the traffic jam and optimizing the charging resource utilization.

The invention discloses a purchase demand prediction method combined with PMC material management and control, and belongs to the technical field of material management and control data processing, and the method comprises the steps: obtaining a material consumption sequence and corresponding timestamp data, and employing a time sequence decomposition method to separate trend components and seasonal components, and obtaining a consumption rule feature sequence; if the fluctuation amplitude of the consumption rule feature sequence exceeds a preset threshold value, determining the consumption rule feature sequence as a basic demand predicted value sequence; obtaining a market price index sequence, and calculating a correlation coefficient between the basic demand predicted value sequence and the market price index sequence; if the absolute value exceeds a set threshold value, updating a basic demand predicted value sequence through a time sequence model; generating a purchase plan sequence according to the basic demand predicted value sequence and the inventory average value; and generating a final output report. The invention discloses a problem that a current material management and control mode cannot capture key changes in time and is difficult to balance contradictions between inventory and production.

The invention discloses an optical storage and charging dynamic power distribution control method and system based on AI prediction, and the method comprises the steps: collecting photovoltaic, energy storage, charging, power grid and environment multi-dimensional operation data through a sensor network, and obtaining a historical and real-time data set through preprocessing; training an AI joint prediction model by using the historical data set, and outputting a photovoltaic output and charging load demand prediction result; inputting system operation constraint condition parameters to delimit an optimization boundary; solving a power instruction reference value of each power module through a multi-objective optimization algorithm in combination with a prediction result and a constraint condition to realize global optimal distribution; and correcting the reference value based on the hardware operation parameters and the deviation between the predicted value and the real-time value, issuing an instruction, monitoring operation, and triggering fault protection. The method effectively solves the problems that a traditional system is rigid in power dispatching, low in photovoltaic consumption rate, large in energy storage life loss, remarkable in power grid impact, insufficient in prediction precision, high in energy transmission loss and weak in multi-target coordination.

The invention relates to the technical field of storage intelligence, and discloses an early warning and allocation system and method for intelligent inventory of inspection and quarantine reagent consumables, and the system comprises a consumption trend extraction module, a demand prediction module, a risk early warning module, an early warning trigger module, a supply feasibility evaluation module and an intelligent allocation module. Performing seasonal decomposition on the historical consumption record to obtain a consumption trend component; periodically superposing the consumption trend component to obtain a demand prediction result; based on the demand prediction result, performing inventory risk research and judgment on the current inventory to obtain an inventory early warning index; when the inventory early warning index exceeds a safety threshold value, inventory early warning information is generated; according to the consumable demand information, carrying out feasibility evaluation on inventory data and logistics constraint conditions of the supply point to obtain a feasible allocation point; generating a deployment scheme according to the safety inventory standard of the feasible deployment point; according to the invention, the perspectiveness and accuracy of early warning and allocation of the intelligent inventory of the inspection and quarantine reagent consumables can be improved.

The invention relates to the technical field of intelligent logistics, and discloses a dynamic optimization method and system for a logistics tail end distribution path. The method comprises the steps of collecting order distribution and transport capacity data in a delivery area, and dividing an order dense area and an order sparse area through a clustering algorithm; extracting order dense area fluctuation data according to a classification result, dynamically adjusting path planning granularity, establishing a prediction model in combination with historical orders and user consumption habits, and outputting future demand prediction; if the predicted value exceeds the current transport capacity, triggering resource configuration to generate an extension scheme, matching the transport capacity with the task in combination with the real-time scheduling request, and generating an optimized scheduling sequence; a response time index is extracted, and the prediction model is calibrated through comparison of an actual response and a predicted value; and on the basis of the calibrated model and performance indexes, integrating regional feature data to iteratively optimize path parameters, and forming a dynamic path optimization mechanism. According to the invention, the path planning accuracy and the resource utilization rate of logistics end distribution are improved.

The invention discloses a sports market demand prediction method and device based on deep learning and a medium, and relates to the technical field of market demand prediction, and the method comprises the steps: collecting sports demand data, and carrying out the preprocessing; performing relation mining and graph structure learning on the preprocessed sports demand data through a graph attention space-time network to generate a macroscopic demand potential energy graph; performing potential area identification on the macroscopic demand potential energy diagram by adopting a pre-trained sports market simulation model, and outputting local demand prediction data; performing weighted fusion and error correction on the macroscopic demand potential energy map and the local demand prediction data, and outputting a sports demand prediction score; and making a sports market demand strategy according to the sports demand prediction score and the multi-granularity demand prediction report, and transmitting the sports market demand strategy to an enterprise manager through an enterprise decision support interface. According to the method, multi-level accurate prediction and decision support of sports market demands are realized through dual-mechanism cooperation of the graph attention space-time network and the space-time convolution.

The invention discloses an intelligent charging management system of a new energy automobile, and relates to the technical field of charging control of the new energy automobile, and the system comprises a user identification and authentication module which supports three authentication modes, account association related information and authentication encryption transmission; the charging demand prediction module collects multiple types of data, and predicts charging related parameters through a fusion algorithm; the charging pile and battery state monitoring module collects various operation data in real time; the power grid load sensing module collects power grid information and synchronizes peak and valley periods; the charging strategy optimization module dynamically adjusts a charging scheme in combination with multi-source data; the safety protection module has six protection functions and a fault diagnosis capability; the data storage module adopts a distributed database and an encryption technology; the man-machine interaction module supports multiple operation modes and displays key information; and the communication coordination module is responsible for multi-terminal data interaction and protocol conversion. According to the invention, intelligence and accuracy of charging management are improved, safety protection and data safety are enhanced, and multi-scene charging requirements are met.

The invention relates to the technical field of manufacturing execution systems, in particular to an intelligent scheduling method integrating demand prediction and supply matching, which comprises the following steps of: acquiring to-be-scheduled order information and a workshop real-time resource state in a manufacturing execution system, constructing an order demand feature vector, the method comprises the following steps: acquiring an equipment precision retentivity value, a current load rate value, an energy consumption level value and an operator skill proficiency value, and constructing a resource supply capability vector; according to the method, when an urgent order insertion instruction is received, a clear decision basis is provided, a rigid scheduling locking block which must be kept unchanged and an adjustable liquid adjustable scheduling block which can be adjusted can be distinguished, only the latter is subjected to accurate splitting and backward moving operation, and the target-clear local adjustment mode is adopted, so that the operation efficiency is greatly improved. The huge calculation overhead and the violent fluctuation of the production plan caused by the traditional global rescheduling are avoided, and the agility of the manufacturing system to deal with the market change is enhanced.

The invention relates to the technical field of environment monitoring and automatic control, and specifically provides a dust monitoring analysis control system based on a port bulk cargo storage yard electronic greenhouse. The system comprises an environment multi-dimensional sensing module, a dust field dynamic reconstruction module, a dust suppression demand prediction module and a precise execution control module, and through multi-dimensional data acquisition, dynamic three-dimensional dust field modeling, future dust suppression demand prediction and space and time sequence optimization regulation and control of spraying resources, the dust suppression demand prediction and the precise execution control module are optimized. Accurate monitoring and prediction of dust concentration and on-demand throwing of dust suppression resources are achieved, and control timeliness, foresight and operation energy efficiency are effectively improved.

The invention discloses a modeling and scheduling method and device for a dynamic load and a shared distributed energy storage system under shared power conversion, and relates to the technical field of shared power conversion and distributed energy storage systems, and the modeling and scheduling method comprises the steps: building a multi-target optimization scheduling model based on user power conversion information and power conversion demand prediction data under a shared power conversion scene; and constructing a collaborative solving mechanism, solving the comprehensive scheduling model, and obtaining and selecting an optimal shared energy storage scheduling scheme. The optimal scheduling model comprehensively considers the economic cost of optimal scheduling and the operation and scheduling balance of the shared distributed energy storage system, so that the cost of configuring energy storage by a single user can be reduced, the operation coordination and load balance capability of the whole system can be optimized and improved, good resource sharing and system coordination effects are embodied, and the user experience is improved. The method is suitable for building a flexible energy system with a user side as a core in the future.

The invention discloses a smart home energy efficiency optimization method and system based on multi-source perception learning. The method comprises the steps of multi-source data acquisition and preprocessing, multi-source perception feature extraction, user behavior and environment dynamic learning, demand prediction, scene recognition, multi-target collaborative optimization and control strategy generation and strategy execution and feedback learning. The invention relates to the technical field of smart home, in particular to a smart home energy efficiency optimization method and system based on multi-source perception learning, and according to the scheme, multi-source perception data are collected, a multi-target optimization model is constructed, smart home scheduling is driven by using real-time electricity price, and user satisfaction is improved; an improved self-adaptive optimization algorithm is constructed, parameters of a multi-objective optimization model are optimized, dynamic electricity price and user behavior changes are responded in a mode of searching an optimal solution, energy efficiency is optimized, and interference on living habits of users is reduced to the maximum extent.

The invention discloses a new energy automobile charging scheduling method and system based on deep learning, and the method comprises the steps: carrying out the preprocessing of data, and obtaining multi-source data; a self-attention mechanism of Transform is combined with a GNN graph neural network to establish a deep learning model, a time sequence is processed, a topological relation between geographic distribution of charging piles and a power grid load is modeled through GNN, and spatio-temporal joint features are output; meta-learning is introduced to dynamically adjust the weight of the spatio-temporal joint feature according to a real-time environment, and a charging demand prediction index is obtained; and solving an optimal charging pile distribution scheme through MIP mixed integer programming according to the charging demand prediction index, and generating a charging scheduling scheme by adopting a greedy algorithm. The total charging cost is reduced, the average waiting time of users is shortened, and the power grid load variance is reduced.

The invention relates to the technical field of artificial intelligence, in particular to a distributed edge computing resource scheduling method, system and device, a medium and a product. According to the method, the resource demand prediction value of each edge node is obtained by obtaining the real-time resource use data of each edge node and inputting the resource use data into a trained resource demand prediction model, so that the resource scheduling scheme is generated according to the resource demand prediction value and the resource use data in combination with a preset multi-target optimization strategy; and resource allocation processing and task migration processing are performed on each edge node according to a preset distributed coordination mechanism, so that the resource scheduling efficiency and accuracy of the distributed edge nodes are improved.

The invention relates to the technical field of data management, and discloses a power grid material supply chain-based data intelligent management synchronization system, which comprises a weight distribution module, a material demand prediction module, a material allocation scheme preliminary generation module, a data feedback module and a target material synchronization scheme generation module, wherein an attention coefficient matrix of a power grid material supply chain is constructed based on an attenuation weight and a dynamic weight, and weighted fusion is performed on supply data to obtain a material demand prediction result; generating a preliminary material allocation suggestion scheme of the power grid material supply chain; collecting feedback information of the power grid material supply chain after the preliminary material allocation suggestion scheme is executed, and transmitting the feedback information to the preliminary material allocation suggestion scheme; dynamically adjusting the preliminary material allocation suggestion scheme by taking resource constraints and operation priorities of nodes in a power grid material supply chain as constraint conditions to obtain a target material synchronization scheme; the efficiency based on material supply management can be improved.

The invention discloses a cloud resource dynamic scheduling method, system and device and a storage medium, relates to the technical field of cloud computing resource management, and provides basic data support for resource demand prediction by collecting multi-dimensional core resource real-time data. The input feature weight is dynamically adjusted by means of an LSTM neural network model fused with a dynamic feature selection mechanism, and the short-term fluctuation and long-term trend of the resource demand are respectively captured through an LSTM structure, so that the time sequence analysis better fits the multi-scale characteristics of the load, and the accuracy of resource demand prediction is improved; a prediction result and a current resource configuration state are integrated through a benefit function, overall consideration of economic constraints and service quality requirements is realized, iterative solution is carried out in combination with an optimization algorithm, and a resource scheduling scheme can balance multiple objectives; meanwhile, the model design naturally has adaptive capacity to complex load changes, the deviation between resource configuration and actual requirements is reduced, and accurate prediction and efficient dynamic configuration of cloud resource requirements are achieved.

The invention belongs to the technical field of hotel management systems, and particularly relates to a hotel storage scheduling and prediction optimization method based on artificial intelligence, and the method comprises the following steps: obtaining storage ledger data, business behavior data and external environment data, and fusing the data to generate a feature vector; constructing a semantic mapping relationship between the business behavior and inventory consumption, and outputting a mapping result; inputting the feature vector and the mapping result into a multi-model fusion prediction architecture, and outputting a demand prediction value and a prediction confidence interval of each material in a future preset period; comparing the predicted value with the actual inventory consumption, and triggering the self-learning process of the model; and regenerating a prediction result from the prediction model subjected to self-learning updating, solving an optimal replenishment and allocation scheme based on the prediction result and an inventory constraint condition, and constructing a reinforcement learning model for strategy parameter self-updating by collecting an actual execution result. According to the method, inventory ledger data and consumption behavior data are fused through AI, and a dynamic self-learning warehousing prediction system is established.

The invention discloses a clothing demand dynamic prediction method and system based on multi-source data fusion and machine learning, and the method comprises the steps: collecting and preprocessing multi-source heterogeneous data, and obtaining basic time sequence features, static attribute features and external situation features containing fashion trend quantification features; inputting the basic time sequence and the static attribute characteristics into a time sequence processing network to obtain a reference trend prediction value; inputting the external situation features into a situation feature processing network to obtain a situation influence vector; generating a dynamic adjustment coefficient by the vector through a gating unit; and finally, performing fusion calculation according to the reference trend prediction value and the dynamic adjustment coefficient to obtain a final demand prediction quantity. According to the invention, through a double-flow network structure and a gating fusion mechanism, effective modeling is carried out on an internal sales law and external situation impact, the prediction accuracy and the response speed to market changes are significantly improved, and accurate and dynamic decision support is provided for a clothing supply chain.