136 results about "Combination strategy" patented technology

The invention discloses a cold station multi-equipment combination energy efficiency optimization method and system, relates to the field of cold station energy efficiency optimization, and aims to solve the problems of low energy efficiency and response lag caused by manual adjustment. The method comprises the following steps: S1, acquiring key operation data (power, state, temperature, flow and frequency) of a water chilling unit, a cooling tower and a water pump and external environment data (temperature and humidity, wind speed and load prediction) in real time; s2, constructing an equipment energy efficiency model based on historical data, and quantifying the change of the equipment efficiency along with the load rate, the temperature difference and the environment temperature; s3, according to the current load and environment prediction, simulating and traversing the energy consumption of the equipment start-stop combination and the water pump frequency combination by applying a Monte Carlo algorithm, and screening a combination strategy with the lowest total energy consumption of the system; s4, converting the optimization strategy into a control instruction, and issuing the control instruction to equipment for execution; and S5, comparing the actual energy consumption with the predicted value, and if the deviation exceeds 5%, correcting model parameters by adopting a least square regression algorithm to realize closed-loop optimization. And the energy efficiency and the response speed of the cold station are obviously improved.

The invention discloses a multi-knowledge-base scheduling routing method based on intention recognition and feedback optimization, and the method comprises the steps: S1, receiving multi-modal input from a user, carrying out the deep semantic analysis of the input content through natural language understanding and multi-modal analysis, and outputting a standardized intention label; s2, constructing a capability portrait for each knowledge base, and then generating a route matching score matrix according to a similarity relationship between the user intention tag and the capability portrait of each knowledge base; s3, screening out one or more most relevant candidate knowledge bases according to the scoring matrix and a set matching threshold value, and selecting a retrieval combination strategy to realize information recall; s4, information retrieval is executed based on the selected knowledge base, recall results and original user input are submitted to a large language model for content generation, and answers or suggestions are formed and output; and S5, user feedback collection and strategy optimization are carried out. According to the method, a more accurate, efficient and self-evolutionary knowledge base scheduling mechanism can be realized.

The invention provides a software development data management method and system based on cloud service. The method comprises the following steps: acquiring code version difference data generated in a multi-version development process from a cloud service environment, and allocating the code version difference data to a corresponding storage position according to a version sequence identifier; a version relation graph is constructed based on the calling relation between the version branches, and the data flow direction and the logic dependency relation between versions are dynamically recorded; carrying out sorting and logic dependence verification on conflict contents in combination with the conflict state and the merging strategy, and generating a traceable merging operation instruction set; and finally, executing feedback data according to an instruction set to update a graph conflict state, synchronously triggering storage data update and redistribution of version identifiers, and realizing full-link closed-loop management of development data. According to the technical scheme provided by the invention, the efficiency and accuracy of software development data management can be improved.

The invention relates to a user operation risk dynamic monitoring method and device based on data consanguinity and a medium, and the method comprises the following steps: obtaining operation behaviors of all users in real time, and extracting a segment of continuous operation behaviors of each user as a group of behavior fragments with semantic relevance; a plurality of continuous behavior segments of each user are matched with a preset behavior template, so that a plurality of intermediate behavior blocks are obtained through dimension reduction; according to a time sequence, traversing and extracting a predetermined number of intermediate behavior blocks from the plurality of intermediate behavior blocks, and respectively combining to generate a behavior fragment sequence; and respectively inputting the behavior fragment sequences into a pre-trained neural network model to obtain a risk classification result or a risk probability distribution result corresponding to each behavior fragment sequence. The method has the advantages that each operation behavior can be subjected to structured expression, the potential behavior chain can be reconstructed through the combination strategy, and the high-risk behavior path hidden in the data noise is recognized by means of the deep model.

The invention relates to the technical field of software engineering and artificial intelligence crossing, in particular to an API test case rapid generation method and system based on an AI large model.The method comprises the following steps that joint analysis and semantic fusion are conducted on an API document and a demand document through a multi-modal data analysis layer, and a complete test context is constructed; generating a test case based on the pre-trained large model by using a hybrid inference engine; the method has the beneficial effects that the association constraint of the API document and the demand document is automatically extracted through semantic understanding; a parameter combination strategy is optimized through reinforcement learning, and the problem of combination explosion is solved; constructing a self-adaptive test case maintenance mechanism, and responding to API version iteration; an executable test code that can be directly integrated into the CI / CD pipeline is generated.

The invention discloses an intelligent item selection recommendation system of a live broadcast room based on AI digital people, relates to the technical field of intelligent item selection recommendation, and solves the technical problems of generating a dynamic and high-confidence matching scheme based on highly matched commodities and adjusting a combination strategy in combination with a real-time scene and user feedback. Through the multi-dimensional weighted calculation of the feature matching score, the crowd matching score and the scene association score, the hidden integrating degree of the commodity and the user demand is quantified, the highly matched commodity is screened out, the invalid recommendation is reduced, the comprehensive confidence degree of the commodity combination is calculated based on the association rule algorithm, the personalized matching scheme conforming to the real-time scene is generated, and the user experience is improved. A high-potential scheme is screened through a self-adaptive threshold value, the matching rationality and transformation potential are improved, a comprehensive feedback processing module is introduced, a user is supported to customize combined commodities, and optimization suggestions are provided through system evaluation; and for user change feedback, generating hidden collocation based on a user portrait, and realizing real-time iteration of a recommendation scheme.

The invention discloses a cooperative control and disturbance suppression method for a multi-axis servo system, and relates to the technical field of cooperative and disturbance control of servo systems, position and speed signals of each servo axis are acquired in real time, a cooperative error vector is calculated based on a graph theory model, the vector reflects the deviation between the state of each axis and an expected cooperative trajectory, and the cooperative control and disturbance suppression of the multi-axis servo system are realized. And synchronous movement is ensured. The collaborative error is used for generating a preliminary control signal, and a proportional, differential and integral term combination strategy is adopted to quickly correct the error and eliminate the steady-state error. Meanwhile, disturbance vectors are estimated through an observer, and external interference or internal changes are compensated. And after the preliminary control signal and the disturbance compensation are added, amplitude limiting processing is carried out to prevent saturation of the driver. The control parameters are adjusted through an off-line optimization algorithm to minimize errors and disturbance influences, so that the tracking precision, the anti-interference capability and the adaptability of the multi-axis servo system are improved, and the method is suitable for the fields of high-precision manufacturing, robot control and the like.

The invention relates to the technical field of natural language processing, and provides a multi-granularity text compression method and system for large-model multi-round dialogues, and the method comprises the steps: carrying out the semantic coding of historical question and answer pairs and a current dialogue question through a pre-training model, obtaining a semantic vector, generating a historical dialogue semantic representation through a self-attention mechanism, and carrying out the semantic representation of the historical dialogue. A semantic vector of a current dialogue problem is combined, a correlation enhancement vector is obtained through a target attention mechanism, weighted fusion is carried out on the correlation enhancement vector and historical dialogue semantic representation, a fusion vector is obtained, a feature vector is generated through a maximum pooling and average pooling combination strategy, clustering is carried out, positive and negative sample pairs are constructed, and semantic similarity loss calculation is carried out. Iteratively training the classification model; and based on the trained classification model, screening question and answer pairs related to the current dialogue question, and generating a structured abstract. Redundant content is remarkably reduced while key semantic information is reserved, and efficient context representation adaptive to large model reasoning is formed.

The invention discloses a table collaborative editing method and system, and relates to the technical field of artificial intelligence, and the method comprises the following steps: during multi-user collaborative editing, receiving user editing operations, packaging each user editing operation into an independent operation sequence, summarizing all the operation sequences, and performing unified operation on the operation sequences, the operation sequence comprises content information and parameter information of user editing behaviors; performing operation conflict detection based on the operation sequence, and if an operation conflict is detected, dividing the conflict into a direct conflict, a structure conflict, a style conflict and a dependency conflict through a conflict classifier; corresponding strategies are matched and executed for different conflict types through the intelligent conflict solver, the strategies comprise the automatic merging strategy, the user selection strategy and the delay processing strategy, and the problems that in the prior art, cell merging intelligent verification cannot be achieved, and conflict solving cannot be achieved during multi-user collaborative editing are solved are solved.

The invention relates to the technical field of computer system performance and security testing, in particular to a computer performance security testing method and system. The method specifically comprises the steps that performance indexes, behavior characteristics and security event data of a target system are obtained in real time and subjected to standardization processing, clustering modeling is conducted on the behavior characteristics subjected to dimension reduction through a Gaussian mixture model so as to establish a normal behavior pattern cluster, and time sequence analysis is conducted on a behavior sequence in combination with a hidden Markov model so as to recognize a system state transition path; adopting a weighted combination strategy to mix and inject a normal load and an aggressive load, and adjusting an abnormal behavior chain through a disturbance function; constructing directed weighted association between the abnormal event and the performance index based on an event-driven causal performance diagram, quantifying a system risk index and triggering an alarm; outputting performance bottlenecks, security vulnerabilities and optimization suggestions, and updating the knowledge base to optimize subsequent models. According to the invention, deep coupling of performance and safety testing, intelligent scheduling of dynamic loads and interpretable positioning of risk paths are realized.

The invention relates to a tail end constant-pressure water supply strategy adjusting method and system based on AI, and belongs to the technical field of secondary water supply.The constant-pressure water supply strategy adjusting method comprises the steps that the flow of a header pipe and real-time health state data of each running pump are periodically collected; according to the header pipe flow, the dynamic demand lift is calculated, and all feasible pump set combination strategies are generated in combination with the dynamic demand lift and the real-time health state data; the score of each candidate strategy is calculated, the candidate strategy with the highest score is used as a pump set scheduling strategy, the pump with the most matched performance is selected from candidate pump sets corresponding to the pump set scheduling strategy to serve as a frequency conversion main guide pump, and the other pumps serve as frequency conversion auxiliary pumps; according to the set minimum service pressure and the dynamic demand lift, the target pressure of a pump outlet is calculated; and according to the target pressure and the actually-collected pump outlet pressure, the rotating speed of the variable-frequency main guide pump is adjusted. The method has the beneficial effects that the power of the pump set is matched with the actual requirement, so that the operation cost is reduced.

The invention discloses a self-service door opening method for a store, and relates to the technical field of self-service stores. The system carries out intelligent matching according to a pre-stored user file, calculates a similarity threshold value through an algorithm, judges that a user is a legal user if the matching degree exceeds a preset value, or rejects access; a dynamic combination strategy is adopted to allow an administrator to configure weights and combination modes of different fields according to the security requirements of the store, so that the user experience is optimized and the privacy leakage risk is reduced on the premise of ensuring the security; a background management system flexibly sets a weight value and combinatorial logic of each verification element, the matching efficiency is dynamically optimized based on a preset similarity threshold, if a comprehensive score after element combination reaches a security level score configured by the system, rapid door opening is automatically authorized, otherwise, secondary verification or manual auditing is triggered, and the safety of the system is ensured. The false rejection rate is reduced, the safety protection capability is improved, the verification modes are automatically switched according to regions or time periods, and the door opening speed is high when the operation is successful.

The invention discloses a system for improving item recommendation accuracy based on space-time diffusion and information bottleneck. The system comprises the following modules: a normalization processing module, a space aggregation module, a time aggregation module, a time diffusion module, a space diffusion module, an information bottleneck module and a prediction module. The system can effectively model an error behavior which is not solved by a current next basket recommendation technology, and through an innovative spatial diffusion module, the system does not depend on strong association among commodities or historical preferences of users any more, but focuses on learning of article spatial combinatorial logic in a shopping basket, so that the system is more intelligent. Therefore, the system can identify the targeted and hidden combination strategy of the user, thereby improving the understanding of the real purchase intention of the user, and achieving the purpose of improving the article recommendation accuracy.

The invention discloses a service combination method based on a colored Petri net and a deep reinforcement learning agent. The method comprises the following steps: S1, constructing a colored Petri net model for manufacturing a service combination system; s2, constructing a deep reinforcement learning environment based on the model, and defining a state space, an action space and a reward function; s3, interacting the intelligent agent with the environment, dynamically generating an action mask by using a Petri net model to constrain an action space, and training the intelligent agent to converge by using a near-end strategy optimization algorithm; and S4, taking the strategy obtained by training as an optimal service combination strategy, and triggering transition in the Petri network to generate a final manufacturing service combination scheme. According to the method, the formalized modeling capability of the colored Petri network and the self-adaptive decision-making capability of deep reinforcement learning are deeply coupled, so that the problem that in a dynamic complex industrial interconnection environment, the formalized modeling capability of the colored Petri network and the self-adaptive decision-making capability of the deep reinforcement learning are influenced is solved; the problems of low decision-making efficiency, poor adaptive ability and opaque optimization process caused by huge decision-making space and complex logic constraints in a manufacturing service combination method are solved.

The invention discloses a cloth warehouse-out scheduling method and system based on multi-strategy cooperation, and the method comprises the steps: constructing a warehouse three-dimensional coordinate model, and building the precise mapping of a warehouse space code and a three-dimensional coordinate; the order priority is quantified by adopting an analytic hierarchy process-fuzzy comprehensive evaluation (AHP-FCE) model, and resource allocation in an idle period and a busy period is adapted through dual-mode scheduling; a shortest order picking path is planned based on a Dijkstra algorithm; based on a reinforcement learning deep Q network (DQN) dynamic decision order combination strategy, invalid movement and resource waste are reduced. The method can remarkably shorten the order processing time, improve the resource utilization rate, balance the personnel efficiency, effectively relieve the operation bottleneck in cloth delivery, and is suitable for various cloth storage scenes.

The present invention relates to the technical field of operating system memory management, and discloses a real-time container memory management method and system for an operating system. By solving the system user combination strategy and the system memory allocation strategy to guide the memory management of the operating system, different minimum management cycles of several system users in the target memory management period are combined, and overall memory allocation is performed for each system user group instead of individual memory allocation for each system user, which can provide redundant memory allocation for each system user, avoid the problem of program running delay caused by calling data from the disk due to insufficient memory allocation accuracy. At the same time, when the memory resources are sufficient, the memory can be released as much as possible for users to use, and when the memory resources are tight, the memory can be utilized maximally, improving the rationality and intelligence of memory management in the multi-user scenario of the operating system, and while meeting the memory requirements of more users as much as possible, improving the fineness of memory allocation.

The embodiment of the invention provides a strategy determination method and device. The method comprises the following steps: acquiring real-time index data and a real-time prediction load of a Serverless function; determining a plurality of target optimization functions according to the real-time index data and the real-time prediction load; the multiple target optimization functions comprise a first target function and a second target function, the first target function is used for minimizing the cold start delay of the function, and the second target function is used for minimizing the resource cost of the function; the candidate solution of each target optimization function indicates a combined strategy of a resource allocation strategy and a cold start strategy; a plurality of target optimization functions are processed by utilizing a multi-target evolutionary algorithm, and a target combination strategy is screened out from numerous candidate solutions by simulating the principle of survival of victims and inferior and continuous optimization of populations in a biological evolution process, so that the resource cost of the functions is reduced on the basis of ensuring cold start low delay of the functions; and the resource allocation efficiency is improved.

The invention provides a page loading method and device and related equipment.The method comprises the steps that a client side responds to loading operation of a page of an application, M loading requests of M services needing to be loaded on the page are obtained, then a merging strategy configured by a user is obtained, N loading requests in the M loading requests are merged according to the merging strategy, and N is a positive integer; a merging request is generated, and N services corresponding to the N requests are provided by the same server side. The client side sends a merging request to the server side to indicate the server side to execute the merging request, then receives an execution result of the merging request returned by the server side, analyzes the execution result into a request result corresponding to each loading request in N loading requests, and sends the request result to the server side; and displaying the request result to the position of the service corresponding to each loading request in the N loading requests in the page. Based on the method provided by the invention, the number of times of interaction between the client and the server can be reduced, and the page loading speed can be improved.

The present application relates to a kind of multi-stage self-defined combination strategy warehouse picking optimization method, device and medium, wherein, method includes collecting order data to be picked and preprocessing generation structured data set;Support user self-defined configuration wave division dimension and weight, complete order clustering and generate picking wave with unique identification;Divide the whole process of picking into four core stages, provide each stage strategy selection and combination interface, generate compatible multi-stage strategy combination scheme;Correlation scheme and wave task are executed in stages, real-time acquisition of field data and dynamically correct abnormal strategy;After completing picking, the core index is counted, report is generated and review output optimization suggestion, the present application solves the problem that traditional picking strategy is poor in adaptability, low efficiency, can be flexibly adapted to multiple warehouse scene, improve picking efficiency and accuracy, reduce resource consumption, and the practical value is remarkable.

The invention discloses an intelligent flight route planning method and system based on big data. The method comprises the steps of multi-source flight route data acquisition, obstacle cluster construction, high-density obstacle area identification, hierarchical planning cost function design, flight planning route parameter optimization and intelligent flight route planning. The invention relates to the technical field of data processing, in particular to an intelligent flight route planning method and system based on big data, according to the scheme, a multi-dimensional high-density obstacle area recognition mechanism is innovatively proposed, a hierarchical planning cost function collaborative technical scheme is designed, the flight safety is enhanced, and the obstacle avoidance efficiency is improved; introduction of a density perception initialization strategy and a double-layer particle ball merging strategy is innovatively proposed, a clustering algorithm is improved, and clustering accuracy and stability are improved. An optimization algorithm is improved by adopting a self-adaptive exploration strategy and a quadratic interpolation strategy, the efficiency of route parameter optimization and the quality of an optimal solution are improved, and accurate optimization of flight planning route parameters is realized.

The invention belongs to the technical field of network security, and discloses a data link layer-oriented industrial control network traffic classification method. Providing a semi-supervised classifier based on a nearest cluster, and constructing the semi-supervised classifier based on a K-Means-based secondary clustering method; the classifier can process data with labels and data without labels at the same time. A combined classification method based on a semi-supervised framework is provided; clustering is performed on industrial control flow data by using an improved K-nearest neighbor density peak value clustering algorithm; and combining the obtained clustering result with the prediction result of the classifier, and adopting a combination strategy to improve the classification accuracy and ensure the reliability of the result. Considering the scarcity of the industrial control traffic data, the invention aims to collect a large amount of data link layer industrial control traffic data based on traffic sessions in real time. Clustering and classification are carried out by taking the traffic session as a basic unit, so that the method not only can capture subtle differences of the traffic of the industrial control network in detail, but also can effectively improve the accuracy of an algorithm.

The present invention relates to an active suppression method for motor bearing voltage based on a software-hardware combination strategy, belonging to the technical field of motor bearing voltage suppression. The active suppression method for motor bearing voltage based on a software-hardware combination strategy includes the following steps: constructing a common-mode equivalent circuit of the internal capacitance of the motor; calculating the bearing voltage division ratio of the motor according to the common-mode equivalent circuit; adopting the AZSPWM strategy for motor control, and injecting voltage at both ends of the bearing by using the principle of the common-mode transformer to cancel the bearing voltage. The invention saves volume, has a simple method, does not need to suppress the common-mode voltage from the source, only needs to suppress the bearing voltage with a small amplitude, the power of the four-leg transistors required is small, and can completely eliminate the bearing voltage and suppress the bearing electro-corrosion phenomenon.

The invention relates to the technical field of natural language processing, and provides a large language model event extraction method fusing label reconstruction and a multi-dimensional instruction set, which comprises the following steps of: 1, reconstructing a label; 2, constructing a multi-dimensional instruction set; constructing two different multi-dimensional instruction sets for the multi-dimensional instruction library by adopting a multi-dimensional layered alternate combination strategy through a second-level instruction architecture and a third-level instruction architecture; 3, fine adjustment of the model; taking the reconstructed document-level event text and the event record sequences in two different formats as input and output of fine tuning respectively, and performing fine tuning on an LLaMA-3. 2-1B model by using a multi-dimensional instruction set and adopting a LoRA technology; 4, event extraction; and extracting events by using a large model for fusing label reconstruction and multi-dimensional instruction set fine tuning. According to the method, the data labeling cost is reduced through label reconstruction, the semantic analysis capability of a large language model on financial field events is enhanced by utilizing a multi-dimensional hierarchical instruction set, and the financial field event extraction performance is improved.

A video recommendation method fusing social information comprises the steps that firstly, a user-video interaction graph and a social graph are constructed through video media platform data, different combination strategies of a diffusion model and graph convolution operation are applied in a denoising priority path and a structure priority path respectively, and collaborative perception and fusion of user embedding and structure information are achieved; secondly, aligning user embedding output by the double-track denoising path by adopting comparative learning, and enhancing robustness; secondly, performing deep coding on the user-video interaction graph and the social graph by applying a double-graph neural network, comprehensively capturing user preferences and embedding videos; and finally, for noise possibly introduced by multi-module fusion, deep denoising is executed to purify user embedding again. According to the method, guidance of structural information is introduced in the diffusion denoising process, and more robust user embedding is obtained by using a comparative learning strategy, so that a structural perception high-fidelity social denoising task is completed, and personalized video recommendation is realized on a video media platform.

The application provides a kind of combination navigation method and system based on dual-antenna heading, comprising: INS and dual-antenna GNSS system are combined together, and utilize the proposed dual-antenna GNSS / INS combination navigation method, so that the system can output continuous stable high-precision carrier attitude information when GNSS signal is disturbed.In the method, a heading combination update strategy is proposed, and the continuous and stable output of the device heading information is completed by parallel execution of the solving process of dual-antenna orientation and the initialization process of IMU.Meanwhile, the sensor information is detected to judge whether the device is stationary, and if it is stationary, the IMU parameter calibration is carried out.According to whether the sensor is initialized, two different precision positioning and orientation results are obtained.The application combines the advantages of dual-antenna GNSS and INS, and applies the dual-antenna combination strategy, so that continuous and effective positioning and orientation results can be obtained without waiting for a long time, and has good practical value.

The invention belongs to the field of distributed version control, and relates to a Git multi-warehouse branch batch statistics and automatic management method and device, computer equipment and a storage medium, and the method comprises the steps: deploying a Shell warehouse branch management script at a terminal, and providing a consistent interaction interface for a user; obtaining a terminal warehouse path and a directory to be scanned under the warehouse path; traversing all Git warehouses under the catalogue to be scanned, acquiring a target branch and a reference branch in each Git warehouse, and comparing whether the target branch is merged with the reference branch or not; displaying the merging state of each warehouse and each target branch, classifying and listing Git warehouses with unmerged branches, and outputting a Git warehouse state statistical result; obtaining Git warehouse branches needing to be merged, configuring a merging strategy, and generating a merging plan; and executing a merging operation, and pushing a merging result to a remote warehouse. And the management efficiency of multi-Git warehouse branch merging can be greatly improved.

The invention discloses a flood storage and detention area optimization scheduling method, and relates to the flood storage scheduling and data processing technology, and the method comprises the steps: determining a plurality of flood storage and detention areas contained in a target drainage basin, and determining the coverage range of each flood storage and detention area; determining a plurality of flood storage heights in the coverage range of each flood storage and detention area, wherein different flood storage heights correspond to corresponding submerging losses; gate parameters of barrage bodies of all the flood storage and detention areas are determined, and gate-inundation loss association rules of the upper-level flood storage and detention areas and the lower-level flood storage and detention areas are established on the basis of the gate parameters; acquiring flood flow data of the target drainage basin; and based on the flood flow data and the gate-inundation loss association rule, generating a gate opening combination strategy to complete optimal scheduling of the flood storage and detention area. According to the method, the operation efficiency of different flood diversion requirements can be improved, the decision-making time when a disaster comes is shortened, the flood diversion effect of the flood storage and detention area is utilized to the maximum extent, and the drainage basin loss is greatly reduced.

The invention discloses a social governance event management method based on multi-source data fusion, and relates to the technical field of social governance and public services. The four modules are respectively a multi-source heterogeneous data fusion and feature standardization module, an event intelligent assignment decision engine module, a repeated event identification and combination strategy module and a whole-process closed-loop management and treatment efficiency quantitative evaluation module, and the multi-source heterogeneous data fusion and feature standardization module comprises feature extraction and feature fusion. According to the social governance event management method based on multi-source data fusion, a heterogeneous event feature vector fusion method based on deep learning is adopted, specifically, a text is vectorized through BERT / Word2Vec, and multi-source heterogeneous data is fused into a unified event feature vector in combination with a weighted summation method, so that information islands formed by multi-source data splitting are avoided, and the social governance event management efficiency is improved. And the integration of the social governance event management data is improved.

The invention discloses a multi-type product combined packaging method and system based on visual quality inspection and related equipment. The method comprises the steps that visual quality inspection is conducted on each product; packaging the plurality of products passing the quality inspection one by one; analyzing the uploaded package combination strategy, and determining all target product types and the quantity specification corresponding to each target product type; and according to each target product type and the corresponding quantity specification, calling each packaged product, and packaging each called product to form a commodity. Therefore, a full-link quality inspection packaging system including product quality inspection screening, packaging and commodity obtaining by combining a plurality of products according to requirements can be formed, and the quality stability and production pertinence of the commodities are ensured. Meanwhile, commodities can be customized and constructed according to a package combination strategy, the overall response speed of multi-type product combination package is improved while the defective rate of production is reduced, and the application universality of the method is improved.

This invention belongs to the field of resource allocation and discloses an emergency resource demand prediction method and apparatus based on WGAN-Stacking. The method includes: acquiring meteorological, geographical, and power grid data; generating fault samples using a Wasserstein generative adversarial network; quantifying the distribution difference between the generated fault samples and real samples based on FID distance and kernel Inception distance to screen target samples whose sample quality reaches a preset threshold, with the quantified distribution difference used to characterize the sample quality of the fault samples; inputting the target samples into a pre-constructed resource demand prediction model to predict the emergency resource demand in different regions; the resource demand prediction model adopts a Stacking ensemble learning framework, using RF, GBDT, and SVM as heterogeneous base learners and FNN as a meta-learner, learning the optimal nonlinear combination strategy between the prediction results of each base learner; and allocating emergency resources according to the prediction results output by the resource demand prediction model.