Multi-agent collaborative intermittent demand forecasting system and its mode classification and model selection method

Abstract

This application discloses a multi-agent collaborative intermittent demand forecasting system and its pattern classification and model selection method. The pattern classification and model selection method includes a pattern classification stage and a model selection stage. In the pattern classification stage, the input high-sensitivity feature set is transformed into a semantic category label set through cluster analysis and semantic inference. In the model selection stage, the semantic category label set is used as the pattern basis. The high-sensitivity feature set and the semantic category label set are integrated. Through a retrieval enhancement generation mechanism, feature and semantic information are fused to output a recommended scheme for the prediction model. When a deviation is detected in the model selection stage, a feedback signal is generated and input into the pattern classification stage. The feedback signal is read first to adjust the execution strategy of pattern classification. This enables the multi-agent collaborative intermittent demand forecasting system to dynamically optimize upstream decisions based on downstream performance, thereby improving the flexibility and reliability of the multi-agent collaborative intermittent demand forecasting system.

Description

Technical Field

[0001] This invention belongs to the field of artificial intelligence and intelligent manufacturing, and in particular relates to a multi-agent collaborative intermittent demand forecasting system and its pattern classification and model selection method. Background Technology

[0002] In the field of intelligent manufacturing, the long service life of large equipment makes spare parts management a prominent challenge for automotive and aerospace companies. Spare parts supply assurance is a crucial link in intelligent manufacturing, directly impacting equipment operating efficiency and downtime losses. Compared to finished product demand, spare parts demand exhibits a distinct long-tail distribution and irregular fluctuations: a wide variety of types with few samples of any single type, often showing periods of zero demand interspersed with sporadic bursts. Demand differences between different factories, equipment, and processes are also significant. This makes demand forecasting not only crucial for inventory optimization and stockout risk control, but also profoundly impacts service levels, operational response efficiency, and maintenance planning, becoming a core link between the operational efficiency and financial performance of intelligent manufacturing.

[0003] Current forecasting research mainly unfolds in two directions: First, traditional time-series forecasting methods, represented by the Croston method and its improved models, exponential smoothing, and (seasonal) ARIMA models. These methods are easy to deploy and perform stably in small-sample scenarios, but they struggle with large-scale automated parameter tuning and are inadequate in adapting to different scenarios and responding to sudden demands. Second, machine learning and ensemble methods based on feature engineering, including gradient boosting techniques, meta-learning methods that automatically select prediction paths from a candidate model library based on feature profiles, and hierarchical probabilistic forecasting models. While these methods improve accuracy under specific conditions, they are still constrained by arduous manual feature engineering, exhibit significant performance fluctuations in scenarios with data distribution shifts or cold starts, and the overall framework still lacks reusability and decision interpretability.

[0004] Currently, there is an urgent need in the industry for a pattern classification and model selection method and system that can provide high reliability and flexibility in intermittent demand forecasting scenarios. Summary of the Invention

[0005] To address the aforementioned deficiencies or improvement needs of existing technologies, this invention provides a multi-agent collaborative intermittent demand forecasting system and its pattern classification and model selection method, which can achieve high reliability and flexibility in intermittent demand forecasting scenarios.

[0006] In some embodiments of this application, through the design of the steps within the pattern classification and model selection method, both the pattern classification agent and the model selection agent execute a dynamic closed-loop mechanism of "perception-decision-execution". In each round of the loop, the current task state is analyzed, the tool to be called is selected through logical reasoning, standardized instructions are generated to drive the tool library to execute, and the cognitive state is updated according to the execution result to enter the next round of decision-making. The input content of each process within the agent is not a fixed template, but is dynamically generated according to the upstream processing results and the current task state. This makes the intermittent demand prediction system of multi-agent collaboration and its pattern classification and model selection method highly flexible, and also improves the overall performance of the multi-agent collaborative system in terms of robustness, execution efficiency and decision transparency.

[0007] To achieve the above objectives, some embodiments of this application provide a pattern classification and model selection method for a multi-agent collaborative intermittent demand forecasting system. The pattern classification and model selection method includes a pattern classification stage and a model selection stage. In the pattern classification stage, the pattern classification agent transforms the input highly sensitive feature set into a semantic category label set through cluster analysis and semantic inference. In the model selection phase, the agent uses the semantic category label set as the pattern basis, integrates the highly sensitive feature set and the semantic category label set, and fuses feature and semantic information through the retrieval enhancement generation mechanism to output the recommendation scheme of the prediction model. Specifically, when a deviation is detected during the model selection phase, a feedback signal is generated and input into the pattern classification phase, and the feedback signal is read first to adjust the execution strategy of pattern classification.

[0008] In some embodiments, the pattern classification stage includes feature preprocessing and preliminary analysis, reasoning-driven clustering strategy selection and iterative optimization, and semantic label generation and structured output steps. In the feature preprocessing and preliminary analysis steps, the input high-sensitivity feature set is standardized and quantitatively diagnosed, and the meta-features of the high-sensitivity feature set are extracted. The meta-features of the high-sensitivity feature set include variance, quantiles, sparsity and correlation density. In the reasoning-driven clustering strategy selection and iterative optimization step, the large language model in the pattern classification step reasons based on the meta-features of the highly sensitive feature set and selects the initial clustering algorithm, performs clustering analysis, and obtains a high-quality clustering structure. After obtaining a stable and high-quality clustering structure, key statistical features are aggregated to form a comprehensive feature vector during the semantic tag generation and structured output process. Based on the comprehensive feature vector and the preset tag generation template, the large language model performs semantic reasoning and generation, produces business semantic tags, and outputs a structured set of semantic category tags.

[0009] In some embodiments, in the pattern classification step, the decision process at each decision step is represented as follows: in, Indicates at time step The reasoning is based on the current state. and historical experience The process of internal analysis and logical deduction using a large language model; Indicates at time step The actions taken were based on reasoning. The conclusion determines the specific operations to be carried out. Indicates at time step The observations were made during the execution of the action. Afterwards, feedback signals or execution results obtained from the outside world, Includes the current highly sensitive feature set, the current clustering results, the preliminary clustering results from the previous round, and a comprehensive evaluation report of the previous round's clustering results. This indicates historical clustering experiences that are similar to the current statistical characteristics.

[0010] In some embodiments, pattern classification also outputs feedback signals to the steps prior to the pattern classification step in a multi-agent collaborative intermittent demand forecasting system, in order to change the execution order of the workflow of the multi-agent collaborative intermittent demand forecasting system.

[0011] In some embodiments, the model selection steps include meta-feature construction and knowledge retrieval, reasoning-driven multi-criteria model analysis and ranking, verification error attribution and feedback generation, generation of recommendations and knowledge base evolution; In the meta-feature construction and knowledge retrieval steps, a meta-feature vector is constructed. Based on the meta-feature vector, accurate semantic retrieval is performed to recall a subset of candidate models. The meta-feature vector integrates statistical, temporal, and deep features of a highly sensitive feature set, semantic label set, and category confidence. In the reasoning-driven multi-criteria model analysis and ranking step, the large language model receives a detailed semantic description of each candidate model subset, and then performs deep matching and multi-criteria joint reasoning to output an interpretively ranked list of candidate models. The detailed semantic description includes the principle, applicable scenarios, parameter requirements, historical performance cases, current meta-features, and semantic category labels. In the verification error attribution and feedback generation steps, the top-ranked candidate models are cross-validated, and error attribution analysis is performed based on the large language model. In the steps of generating recommendations and evolving the knowledge base, an interpretable recommendation report is output, which includes the best model, confidence level, detailed reasons, parameter configuration suggestions, and risk warnings.

[0012] In some embodiments, if the performance of all candidate models fails to meet the threshold requirement or the error pattern points to a systematic upstream defect during the verification error attribution and feedback generation step, a structured and actionable diagnostic report is generated and a feedback signal is issued.

[0013] In some embodiments, during the meta-feature construction and knowledge retrieval steps, an external model semantic knowledge base is integrated through a retrieval enhancement generation mechanism. The external model semantic knowledge base is accurately retrieved, and the subset of candidate models most relevant to the current task is recalled. After the recommendation generation and knowledge base evolution steps, the meta-feature vector, the final recommendation model, the verification error, and the context are stored as a new knowledge unit in the external model semantic knowledge base.

[0014] In some embodiments, at each decision step The decision steps in the model selection process are represented as follows: Among them, system status It includes a highly sensitive feature set and semantic label set, a subset of candidate models, performance metrics of candidate models, error attribution analysis results, and successful model recommendation records under similarity features.

[0015] This application also provides a multi-agent collaborative intermittent demand prediction system, including a data monitoring agent, a feature extraction agent, a pattern classification agent, and a model selection agent. The data monitoring agent transforms the input raw intermittent demand sequence containing mixed noise into a high-quality dataset. The feature extraction agent transforms the high-quality dataset into a highly sensitive feature set. In the pattern classification stage, the agent transforms the input highly sensitive feature set into a semantic category label set through cluster analysis and semantic inference. In the model selection stage, the agent uses the semantic category label set as the pattern basis, integrates the highly sensitive feature set and the semantic category label set, and fuses feature and semantic information through a retrieval enhancement generation mechanism to output a recommended scheme for the prediction model. When a deviation is detected in the model selection stage, a feedback signal is generated and input into the pattern classification stage, and the feedback signal is read first to adjust the execution strategy of pattern classification.

[0016] Some embodiments of this application also provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the prediction method described above.

[0017] Some embodiments of this application specifically involve a multi-agent collaborative intermittent demand forecasting system and its pattern classification and model selection method. The pattern classification agent analyzes the clustering results, generates semantic labels that are easy for business to understand, and classifies complex demand patterns into clear semantic categories. The model selection agent combines a retrieval-enhanced generation (RAG) mechanism with an adaptive feedback loop, calls an external expert knowledge base, and matches the optimal forecasting model for each type of demand. This "divide and conquer-coordinate" architecture design of multi-agent collaborative work provides a systematic solution to cope with the complex uncertainty of intermittent demand forecasting. Each agent constructs a dynamic closed-loop mechanism of "perception-decision-execution": In each round of the loop, the LLM analyzes the current task state, selects the tool to be called through logical reasoning, generates standardized instructions to drive the tool library to execute, and updates the cognitive state according to the execution results to enter the next round of decision-making. This protocol not only ensures the standardization and flexibility of the behavior patterns of each agent, but also provides reliable support and enhances flexibility for system-level coordination of complex tasks. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the workflow of a multi-agent collaborative intermittent demand prediction system provided in some embodiments of this application; Figure 2 for Figure 1 The diagram shows the context-aware adaptive input engineering process of the prediction system for large language models. Figure 3 for Figure 1 The diagram shows a code listing of the adaptive feedback integration mechanism within a multi-agent collaborative intermittent demand forecasting system. Figure 4 for Figure 1 The diagram shows the workflow of the data monitoring agent in a multi-agent collaborative intermittent demand forecasting system. Figure 5 for Figure 1 The diagram shows the workflow of the feature extraction agent in a multi-agent collaborative intermittent demand prediction system. Figure 6 for Figure 5 The diagram shows a code listing illustrating the working process of the feature extraction agent. Figure 7 for Figure 5 The diagram shows a code listing illustrating the adaptive feature selection and fusion process within the intelligent feature extraction system. Figure 8 for Figure 5 The diagram shows the steps of the feature extraction intelligent intracellular adaptive feature selection and fusion process; Figure 9 for Figure 5The diagram shows the steps involved in the feature extraction agent's workflow. Figure 10 for Figure 1 The diagram shows the workflow of a pattern classification agent within a multi-agent collaborative intermittent demand forecasting system. Figure 11 for Figure 10 The diagram shows a code listing for a pattern classification agent. Figure 12 for Figure 10 The diagram shown illustrates the working process of the large language model within the pattern classification intelligent system. Figure 13 for Figure 1 The diagram illustrates the workflow of model agent selection in a multi-agent collaborative intermittent demand forecasting system. Figure 14 A schematic diagram illustrating the performance evaluation of each agent in a multi-agent collaborative intermittent demand prediction system provided in this application embodiment; Figure 15 A schematic diagram of the execution code for a JSON-based structured interaction protocol provided in an embodiment of this application; Figure 16 This is a schematic diagram of an electronic device in some embodiments of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0020] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0021] In the field of intelligent manufacturing, the long service life of large equipment makes spare parts management a prominent challenge for automotive and aerospace companies. Ensuring spare parts supply is a crucial aspect of intelligent manufacturing, directly impacting equipment operating efficiency and downtime losses. While companies often stockpile high-priced inventory to prevent downtime, the extreme sparsity and irregularity of demand still lead to repeated instances of "needing spare parts when needed, or having spare parts but not using them." Compared to finished product demand, spare parts demand exhibits a distinct long-tail distribution and irregular fluctuations: SKUs (Stock Keeping Units, unique identifiers assigned by merchants for refined management and tracking) are numerous, but individual categories have few samples, often resulting in long periods of zero demand interspersed with sporadic bursts. Demand differences between different factories, equipment, and processes are also significant. This makes demand forecasting not only crucial for inventory optimization and stockout risk control but also profoundly impacts service levels, operational response efficiency, and maintenance planning, becoming a core link between the operational efficiency and financial performance of intelligent manufacturing.

[0022] Traditional prediction methods often struggle to balance scalability and interpretability, while directly applying Large Language Models (LLMs) cannot effectively analyze the intrinsic structure of such sparse data.

[0023] like Figure 1 and Figure 2 As shown, this invention provides an LLM-enabled multi-agent collaborative intermittent demand forecasting system, which can be used to predict sparse, intermittent spare parts demand. It includes a data monitoring agent, a feature extraction agent, a pattern classification agent, and a model selection agent. Each agent makes autonomous decisions by comprehensively analyzing upstream output, preset instructions, and historical experience, flexibly calling upon tool modules, and continuously looping a "reasoning-execution-feedback" closed-loop process, forming an end-to-end adaptive intelligent forecasting system (multi-agent workflow) from data monitoring to model optimization. Here, an agent represents an entity within the corresponding system that possesses autonomy, adaptability, and interactivity, capable of perceiving changes in the environment (e.g., through sensors or data input), making judgments and decisions based on its learned knowledge and algorithms, and then executing actions to influence the environment or achieve predetermined goals.

[0024] Specifically, in the embodiments of this application, the data monitoring agent can act as the "quality control center" of the prediction system, responsible for processing the original intermittent demand sequence containing mixed noise. Transform into a clean and well-organized high-quality dataset. Relying on the semantic parsing and context discrimination capabilities of large language models, it can accurately identify anomalies and normal business fluctuations, thereby intelligently scheduling data cleaning tools to ensure the quality of input data.

[0025] The feature extraction agent can act as the "feature engineer" of the prediction system, responsible for normalizing the high-quality data set. Transformed into feature combinations that can keenly capture complex patterns The feature extraction agent uses a large language model as its strategic brain, dynamically analyzes data characteristics through iterative reasoning, autonomously plans the steps and strategies of feature engineering, and flexibly coordinates feature calculation and fusion tools to transform the original requirements into an adaptive representation that integrates statistical features and deep features.

[0026] Pattern classification agents can act as "pattern decoders" within a prediction system, responsible for combining numerical features. Transform into a category of requirement patterns with business semantics Among them, the large language model plays a dual role in clustering strategy selection and semantic generation: first, it matches a suitable clustering algorithm based on the feature structure, and then generates readable labels such as "high-frequency stable consumption type" based on the statistical characteristics of the clustering results, so as to classify complex demand patterns into clear semantic categories and realize the connection between data patterns and business cognition.

[0027] The model selects an agent as the "decision terminal" of the prediction system, responsible for classifying each type of demand pattern. Matching the optimal prediction model Among them, the large language model serves as the reasoning engine that integrates external knowledge. It calls upon the structured model knowledge base through retrieval-enhanced generation technology, combines data features and demand pattern categories to perform multi-dimensional inference, and finally outputs a model recommendation solution with explanations and confidence scores.

[0028] It is understood that, in this embodiment, the general capabilities of the large language model are concretized into four ReAct agents with clearly defined roles and dynamic collaborative operation. These agents make autonomous decisions by comprehensively analyzing upstream outputs, preset instructions, and historical experience, flexibly calling upon specialized modules in a shared tool library, and continuously looping the "reasoning-execution-feedback" closed-loop process. These four types of agents form an end-to-end LLM-enabled multi-agent intermittent demand prediction system, from the initial intermittent demand sequence to the final decision. In this framework, the large language model does not directly perform numerical prediction but undertakes the key functions of core reasoning and intelligent scheduling. Relying on its semantic parsing, logical deduction, and knowledge construction capabilities, the large language model drives the iterative optimization of each agent until the task is accomplished through continuous state analysis, action planning, and effect evaluation. The internal operating mechanisms, tool call links, and collaborative feedback systems of each agent together constitute the dynamic collaborative core of the framework. The large language model provides the entire prediction system with the core driving force for intelligent evolution and autonomous adaptation. By deeply integrating semantic understanding capabilities with statistical methods, a multi-agent collaborative system with autonomous feature extraction, interpretable pattern recognition, and knowledge-guided model selection capabilities was constructed, ultimately establishing a new balance between prediction performance and decision transparency.

[0029] In some embodiments, the prediction method of a multi-agent collaborative intermittent demand prediction system provided in this application may include a data monitoring stage, a feature extraction stage, a pattern classification stage, and a model selection stage. Specifically, the data monitoring agent transforms the input raw intermittent demand sequence containing mixed noise into a high-quality dataset; the feature extraction agent transforms the high-quality dataset into a highly sensitive feature set (or feature combination, as referred to below); the pattern classification agent transforms the input highly sensitive feature set into a semantic category label set through cluster analysis and semantic inference; and the model selection agent uses the semantic category label set as the pattern basis, integrates the highly sensitive feature set and the semantic category label set, and fuses feature and semantic information through a retrieval enhancement generation mechanism to output a recommended scheme for the prediction model. For example, before the data monitoring stage, a modular shared tool library can be established. This modular shared tool library is invoked during the data monitoring, feature extraction, pattern classification, and model selection stages, covering the tools required for each stage. This toolkit integrates full-chain functional modules covering data monitoring, feature extraction, pattern classification, model selection, and performance evaluation. Each tool is encapsulated as a callable unit with a unified interface and independent functionality. Clearly defined functional semantics and calling specifications enable the large language model to accurately identify and schedule tasks. Through this design, the originally complex task of spare parts demand prediction is deconstructed into a standardized sequence of steps that can be planned by the large language model and executed by the toolkit, thus establishing an efficient collaborative mechanism between the general semantic understanding of the large language model and the specialized computational capabilities of the toolkit.

[0030] As shown in Table 1 below, in some examples, the modular shared toolkit may include modules for anomaly monitoring and handling, data preprocessing, data standardization, calculating statistical features, calculating time series features, TSfresh calculating time series features, SBC classification, autoencoder feature extraction, recursive feature elimination, SHAP feature importance analysis, K-means clustering, hierarchical clustering, cluster evaluation, historical storage, LightGBM model prediction, TBS model prediction, RAG TOP-K model selection, and generating operation reports, among other tools.

[0031] Table 1 Overview of Core Multi-Agent Tools In some embodiments, a standardized calling protocol can be built before the data monitoring step. The standardized calling protocol builds a dynamic closed-loop mechanism for perception, decision-making and execution. The calling of the data monitoring agent, feature extraction agent, pattern classification agent and model selection agent and the modular shared tool library follows the standardized calling protocol, and the data monitoring agent, feature extraction agent, pattern classification agent and model selection agent form a whole through the standardized calling protocol.

[0032] Understandably, in this embodiment, a multi-agent collaborative system is constructed, encompassing general scheduling and specialized execution. All agents autonomously collaborate through a standardized invocation protocol, achieving a closed-loop operation from data input to predicted output. Furthermore, the interaction between the multiple agents and the tool library follows a unified standardized invocation protocol, with the large language model playing a core decision-making role. In each iteration, the large language model parses the current task state, selects the tool to be invoked (a specific function) through logical reasoning, generates standardized instructions to drive the tool library's execution, and updates its cognitive state based on the execution results to proceed to the next round of decision-making. This protocol ensures the standardized uniformity of the agents' behavioral patterns and provides reliable support for system-level coordination of complex tasks. In other words, the modular shared tool library and standardized invocation protocol in this application form the technological foundation supporting end-to-end collaboration.

[0033] In summary, by establishing a standardized invocation protocol with a large language model as the decision-making center, we have built a highly efficient and collaborative underlying interaction framework for the four intelligent agents: data monitoring, feature extraction, pattern classification, and model selection. Within this framework, intelligent agents with distinct specializations are organically integrated to jointly construct a complete end-to-end intelligent prediction system driven by a large language model.

[0034] like Figure 1 and Figure 2 As shown, in some embodiments, the scheduling of the workflow (or collaborative process, or simply workflow) of the LLM-enabled multi-agent intermittent demand forecasting method is handled by a state graph execution engine built on the LangGraph framework. This engine positions the large language model as the intelligent decision-making center. Unlike the traditional linear thought chain execution mode, this framework models the forecasting method workflow as a directed state graph: each agent acts as a state node, and the four core agents form an end-to-end information processing link according to the forecasting task logic. Their interaction relationships constitute the state transition path. The large language model dynamically plans the execution order of each node based on the real-time context, improving the flexibility of multi-agent collaboration. This framework adheres to the "divide and conquer" principle, and through the overall scheduling of the state graph engine, it constructs a functionally modular and logically integrated system architecture.

[0035] In some embodiments, the engine may also have a global short-term memory (memory module) into which all agents write processing data and running status in real time, enabling system-level information sharing and backtracking. This design not only retains the output results of each stage but also records execution metadata, including error logs and retries. Crucially, this mechanism supports dynamic path adjustment: for example, when the model selection node detects that the prediction error exceeds the limit, it generates a feedback signal. The engine can then roll back the state to the feature extraction node via a conditional transition path, rather than directly terminating the process, greatly improving system fault tolerance and collaborative efficiency.

[0036] In some embodiments, the multi-agent intermittent demand forecasting system includes an adaptive feedback integration mechanism to form a closed-loop self-optimization capability. This mechanism supports dynamic path adjustment of the forecasting method's workflow. Under this mechanism, feedback signals are no longer simple error reports, but structured signals containing specific problem diagnoses and optimization suggestions. These signals are recorded in the global state. feedback_signal In the field, when the upstream agent is reactivated through the state graph loop mechanism, it will prioritize reading this feedback signal to adjust its execution strategy. In other words, in this embodiment, the input of each process within the data monitoring stage, feature extraction stage, pattern classification, and model selection stage is dynamically generated by the output of the previous process and the task of the current stage.

[0037] For example, such as Figure 1 and Figure 3 As shown, when the model-selected agent detects a significant prediction bias, the system automatically generates a dynamic instruction (feedback signal) to guide the feature extraction agent to add short-term lag features, reactivating both the feature extraction and pattern classification agents. This mechanism allows the system to evolve from a fixed process into a dynamic agent capable of continuous self-optimization based on downstream performance.

[0038] For example, please refer to again Figure 1 and Figure 2 The workflow steps of an LLM-enabled multi-agent intermittent demand forecasting system can be as follows: Please refer to the following: Figure 1 and Figure 4 S1: In the data monitoring phase, as the perception front end of the LLM-enabled multi-agent collaborative intermittent demand forecasting system, relying on the semantic parsing and context discrimination capabilities of the large language model, it executes a closed-loop data cleaning process with semantic awareness. This process cleans the raw, unstructured, noisy, and intermittent demand sequences. Transform into a highly reliable, standardized, and cleaned high-quality dataset. It combines statistical anomaly detection with the semantic reasoning capabilities of a large language model to distinguish between genuine anomalies and normal business fluctuations. Its operational logic adopts a cyclical structure of "perception-reasoning-execution-observation".

[0039] like Figure 4 As shown, Step 1: For example, the data monitoring phase may include a quality scanning and evaluation process. The data monitoring agent can first analyze the input raw data (raw intermittent demand sequence). Perform a rapid quality scan to calculate the proportion of missing values ​​and the proportion of zero values ​​in the original intermittent demand series data, and obtain a basic quality score. : in This represents the number of missing values. This represents the number of duplicate records. This represents the total length of the sequence.

[0040] For example, the preset threshold is If the basic quality score If the original intermittent demand sequence data is deemed to be of substandard quality, then the Step 2 deep anomaly monitoring process will be executed.

[0041] Step 2: The data monitoring stage can include a deep anomaly monitoring process. The data monitoring agent can call the anomaly monitoring and processing tool detect_anomalies_statistical() to combine statistical algorithms with the semantic reasoning capabilities of large language models for joint diagnosis. Specifically, firstly, an improved Isolation Forest algorithm is used for each data point. Calculate anomaly scores Its definition is as follows: in, Indicates sample Expected path length in all isolated trees Its actual path length, This is in the case of a sample size of The normalization constant for the path length.

[0042] For all satisfying Suspected anomalies Among them, the abnormal scoring threshold It can be set to 0.65. This will launch the large language model semantic reasoning engine for confirmation; the input for semantic reasoning is structured context. ,include: front and back The demand value at a given point in time, and the SKU (stock keeping unit, a unique identifier assigned by the merchant to a product for refined management and tracking) of that spare part in the past Relevant business events within the month, such as maintenance work orders, production plan changes, and promotional calendar events. The large language model analyzes this context and outputs confirmation markers. and the corresponding inference confidence. As shown below: In some examples, when the inference confidence of a large language model... i exceeds the threshold Only when this condition is met will the data point be marked as a confirmed outlier.

[0043] Step 3: Dynamic calibration strategy selection and execution. Based on the anomaly types identified in Step 2, an interpolation strategy is dynamically selected. For each identified anomaly, the data monitoring agent selects the optimal correction strategy from the toolkit S. This process is implemented by the large language model by calling the corresponding tools. This decision is controlled by a multi-criteria utility function. here, This represents the score of the improvement effect of strategy s in similar historical scenarios. This indicates its computational efficiency score. The weighting parameters are set as follows: , Prioritize the correction effect. The candidate strategies include linear interpolation, seasonal-trend interpolation based on STL decomposition, and substitution using the mean of similar spare parts. After a strategy is selected, the data intelligence system invokes the appropriate tools to perform the correction and subsequently updates the current data state, denoted as... .

[0044] Step 4: Normalization and Output Process: After completing all anomaly handling, the data monitoring AI will determine the current data status. Performing Min-Max normalization maps the values ​​to the [0, 1] interval, thus generating the final cleaned dataset. The minimum and maximum values ​​used in this normalization will be saved to the global state. This allows for possible inverse transformations in subsequent steps.

[0045] In some embodiments, if the baseline quality score If the data quality is deemed acceptable, the Step 2 deep anomaly monitoring process and the Step 3 dynamic calibration strategy selection and execution process are skipped, and the process directly proceeds to the normalization and output process. Lightweight regularization processing, such as Z-score normalization, is then applied before output. In other words, in this embodiment, the input to each process within the data monitoring agent is dynamically generated by the output of the previous process and the task at the current stage, resulting in greater flexibility. Furthermore, the number of steps executed by the data monitoring agent varies depending on the data type.

[0046] Step 5: In some embodiments, the data monitoring phase also includes closed-loop feedback, where the data monitoring agent is designed to receive feedback. The feedback signal can be transmitted through the global state. The signal is transmitted. Upon receiving such a signal, the data monitoring agent can reassess its quality threshold. This mechanism, which uses anomaly detection parameters to determine whether to initiate a new cleaning cycle (such as a quality scan and evaluation process), enables multi-agent intermittent demand prediction systems to achieve adaptive optimization at the system level. Feedback signals can originate from other stages within the multi-agent intermittent demand prediction system, such as feature extraction, pattern classification, and model selection, to achieve dynamic path adjustment within the mechanism, greatly improving system fault tolerance.

[0047] In this embodiment, the data monitoring agent solves the problem of misjudgment in traditional cleaning methods when dealing with intermittent demands by integrating statistical detection and large model inference.

[0048] S2: As Figure 1 , Figure 5 and Figure 6 As shown, the feature extraction agent is the key transformation module from raw data to a predictive model. During the feature extraction stage, addressing the high-dimensional sparsity and heterogeneous distribution of spare parts demand sequences, the feature extraction agent is built upon a large language model. Using this large language model as the strategy engine, it dynamically analyzes the high-quality dataset through iterative reasoning. Based on the data characteristics, we independently plan the steps and strategies for feature engineering, and flexibly coordinate feature calculation and fusion tools to produce high-quality cleaned datasets. Transform into a highly sensitive feature set with high discriminative power and rich semantics. Specifically: Step 1: In some embodiments, the feature extraction stage may include a context-aware initialization and policy planning process. After the feature extraction agent is activated, it first queries the feedback signal in the global state, such as... feedback_signalFields. If feedback exists from the model selection agent, the large language model will analyze the signal and use it as input to redesign the feature engineering strategy. For example, the feature extraction agent actively responds to external feedback from the pattern classification agent and the model selection agent. For instance, if the feedback signal indicates "insufficient lag features," the large language model can redesign, adding higher-order lag terms in the temporal feature extraction step and re-executing the entire refining process, forming a closed-loop optimization.

[0049] If there is no feedback, the default strategy is executed: first, the demand_pattern_classification() tool is called to classify the cleaned sparse demand sequence based on the SBC classification method. Perform initial pattern assessment, calculate the Average Demand Interval (ADI) and the squared coefficient of variation (CV) of demand. 2 They are categorized into smooth, intermittent, unstable, and blocky demands.

[0050] Step 2: In some embodiments, the feature extraction stage may include a parallel feature extraction tool scheduling process. Based on preliminary analysis or feedback instructions, the large language model formulates a detailed extraction plan and schedules three main types of feature extraction tools: (1) Statistical feature tools (calculate_statistical_features()): calculate window statistics. Set the sliding window to Every day, it calculates statistical characteristics such as mean, variance, skewness, and kurtosis within each window. Simultaneously, it calculates global intermittent indicators, such as the proportion of zero demand. Mean of non-zero demand and standard deviation (2) Temporal feature tool (extract_temporal_patterns()): Deep mining based on the tsfresh library. Key configurations include: (a) Extraction length is (a) The linear trend slope of the sliding window; (b) Perform a Fast Fourier Transform (FFT) and retain the previous... (c) The amplitude and phase of the main frequency components; (d) The intensity indices of the trend, seasonal and residual components are obtained by STL decomposition; (e) The autocorrelation coefficients with lags of 1, 7 and 30 are calculated; (3) Deep feature tool (generate_deep_features_autoencoder()): Train a three-layer fully connected autoencoder. The encoder structure is: input layer (84-dimensional) → ReLU (64-dimensional) → ReLU (32-dimensional) → latent features (16-dimensional). The mean squared error (MSE) is used as the reconstruction loss, and the training is carried out for 50 epochs with a batch size of 32. After training, the 16-dimensional vector output by the encoder is taken as the deep feature; After the above tools are executed in parallel, the generated features are concatenated into the original high-dimensional feature set. Its dimensions Up to several hundred.

[0051] like Figure 7 and Figure 8 As shown, Step 3: In some embodiments, the feature extraction stage may include an adaptive feature selection and fusion process, especially for high-dimensional features. The feature extraction agent then enters the feature refinement stage. The large language model first analyzes... Meta-features, such as mean variance and mean correlation between features, are obtained. Based on this analysis and downstream classification tasks, the large language model selects and executes a hierarchical feature selection strategy from a modular toolkit: ① Quick Filtering: Use the variance threshold filter in the select_features_RFE() tool to remove features with variance less than 1. The characteristics obtained .

[0052] ② Precise Selection: For classification tasks, the `select_features_SHAP()` tool is called first. A basic LightGBM model is fitted to the training set, and the importance ranking of the SHAP values ​​for each feature is calculated. The top-N features are retained. .

[0053] ③ Optional fusion: If there is high collinearity among features, LLM may decide to use principal component analysis (PCA) to fuse features, retaining 95% of the variance component, and finally outputting a low-dimensional, highly sensitive optimized feature set. .

[0054] Step 1: Input and Initialization Phase. The process first receives two key inputs: the raw, unprocessed feature set ( ) and specific downstream task types ( This will determine the direction of feature selection. Subsequently, the system performs an initialization operation, setting the current feature set to be processed as the original feature set, and simultaneously defining a method library containing various mature feature selection algorithms (such as Recursive Feature Emission (RFE), Variance Thresholding (VT), Random Forest (RF), SHAP Value, Lasso Regression, etc.).

[0055] Step 2: State Assessment and Intelligent Decision-Making Core. Before data processing, the system performs state assessment in parallel. It calculates the dimensionality and sparsity of the current feature set and deeply analyzes the task requirements, summarizing this key information into "output features" and passing them to the core "ReAct Decision (LLM Agent)" module. Utilizing the powerful reasoning capabilities of LLM, it analyzes the current state and, through the "action" step, selects the highest-scoring algorithm from the method library as the "best method" for the current situation. )".

[0056] Step 3: Hierarchical Filtering and Validation Feedback Loop. Feature processing employs a hierarchical strategy. First, in the "first stage," a fixed VT (variance threshold) algorithm is used for preliminary coarse filtering to obtain an intermediate feature set. Next, we move to the "second phase," applying the "best approach" selected by LLM to... Perform refined selection to generate candidate feature sets. Then comes the crucial "verification and feedback" stage, where a judgment is made. Does the quality meet the preset threshold? If the quality meets the standards, the process ends, and the final optimized feature set is output. If the quality does not meet the standards, a feedback mechanism is triggered, the "second-best method" is selected as a backup, and the process reverts to the second stage to re-filter until a satisfactory result is obtained.

[0057] like Figure 5 As shown, in some embodiments, the feature extraction agent includes a quality verification loop. During the generation... Then, it uses a simple benchmark model for a quick evaluation on the validation set. If the performance is lower than expected, LLM will re-evaluate the feature selection strategy.

[0058] It is understandable that the feature extraction agent follows the following... Figure 5The defined dynamic workflow coordinates the entire process through a complete ReAct loop: LLM acts as the core inference and planning hub, first performing demand pattern analysis, then scheduling multiple feature computation tools in parallel, and finally executing adaptive feature selection and fusion. This mechanism, through adaptive input engineering strategies, fully leverages LLM's capabilities in context awareness and strategy planning, significantly improving the quality of feature representation for sparse and heterogeneous data, and overcoming the limitations of traditional methods in terms of insufficient feature expressiveness in such scenarios.

[0059] like Figure 9 As shown, after the feature extraction agent starts, the Large Language Model (LLM) first parses the intermittent demand data features and autonomously formulates a feature construction plan. Then, it calls specialized tools such as statistical feature calculation and temporal pattern mining according to the protocol. The LLM dynamically adjusts subsequent strategies based on the feedback from each tool output until a highly discriminative feature set is generated. For example, in the core reasoning stage of the ReAct loop (context-aware initialization and strategy planning), the LLM performs logical deduction based on the current data state. It does not directly handle low-level computation but analyzes data features and combines them with built-in knowledge to formulate the feature construction strategy for the current step. At this stage, the LLM analyzes and determines which mathematical methods can maximize the value of the data, thereby planning a specific action path, i.e., generating an intent plan. During the parallel feature extraction tool scheduling process, the LLM transforms the scheduled features into standardized execution instructions based on a defined strategy, which then interact with the standardized tool library on the right. The LLM precisely selects and calls functional modules from the tool library, such as `Calculate_statistical_features()` for basic statistics, `Extract_temporal_patterns()` for mining temporal patterns, or `Generate_deep_features()` for generating deep features. This process translates abstract semantic planning into concrete computational behavior. During the adaptive feature selection and fusion process, the computational results after tool execution are fed back to the LLM as observations in real time. The LLM receives and parses these feedback results, integrates them with previous contextual information, and updates the current cognitive state. This is a dynamically adjusted process: if the feedback indicates insufficient feature discrimination, the LLM will re-enter the inference stage based on the new state, adjusting the calling strategy or parameters. After multiple rounds of ReAct loops, when the LLM determines that the generated feature set has met the high discrimination requirement or reached the preset termination condition, the loop ends. The system outputs the final fused feature set.

[0060] S3: In some embodiments, such as Figure 1 , Figure 10 and Figure 11 As shown, the pattern classification agent uses the highly sensitive feature set output by the feature extraction agent. As input, its core objective is to transform numerical features into semantic category labels with clear business interpretability through cluster analysis and semantic inference. This provides crucial pattern-based information for the model's agent selection. The pattern classification agent operates under the ReAct model, performing actions at each decision step. Its decision-making process can be formally represented as: in, Indicates at time step The reasoning is based on the current state of the pattern classification agent. and historical experience The process of internal analysis and logical deduction using a large language model; Indicates at time step The actions taken were based on reasoning. The conclusion is that the pattern classification agent determines the specific operation to be performed; Indicates at time step The observations. This was made after the action was carried out. Subsequently, the pattern classification agent obtains feedback information or execution results from the environment or sub-agents.

[0061] Includes the current feature set Current clustering results The preliminary clustering results from the previous round (Initially empty), Comprehensive evaluation report of the previous round of clustering results Including indicators such as the silhouette coefficient and Dunn index, while It is a record of historical clustering experiences that are similar to the current statistical characteristics.

[0062] like Figure 10 and Figure 11 As shown, the execution flow of the pattern classification agent follows a structured iterative loop, sequentially covering three stages: clustering strategy selection, execution and evaluation, and semantic generation. The specific steps are as follows: (1) In some embodiments, the pattern classification stage includes feature preprocessing and preliminary analysis, firstly processing the input high-sensitivity feature set. After standardization, quantitative diagnosis is performed to calculate meta-characteristics such as variance, quantiles, sparsity, and correlation density. .

[0063] (2) In some embodiments, the pattern classification stage includes reasoning-driven clustering strategy selection and iterative optimization: based on the initial state The ReAct loop is initiated. The large language model performs inference and selects the initial clustering algorithm. After execution, the result is... An evaluation report is generated during the observation phase. If the quality does not meet the standards, update the status to [new status]. LLM generation optimization suggestions This triggers a new cycle. This closed-loop mechanism of "evaluation-feedback-replanning" is the core of the pattern classification agent's internal dynamic adaptation, enabling it to continuously optimize the algorithm and parameters based on real-time feedback of clustering results until a stable, high-quality clustering structure is obtained.

[0064] (3) In some embodiments, the pattern classification stage includes semantic label generation and structured output. After the ReAct loop terminates and stable clustering is obtained, the process enters the semantic mapping stage. For each finally determined cluster... The system will aggregate its key statistical features to form a comprehensive feature vector. LLM is based on Based on preset tag generation templates, semantic reasoning and generation are performed to produce outputs such as " High-Frequency, Stable, Growth-Consumable. "and other business semantic tags Ultimately, the pattern classification agent outputs a structured set of semantic labels. and global feedback signal Among them, global feedback signal global feedback signal It can be applied to the data monitoring and / or feature extraction stages. Understandably, when a bias is detected in the model selection stage, the generated feedback signal can also be input into the pattern classification stage, with the feedback signal being prioritized to adjust the execution strategy of pattern classification. Specifically, for example, when a bias is detected in the model selection stage, the generated feedback signal is input into the feature extraction stage. The feature extraction stage outputs a new set of highly sensitive features, which is then input into the pattern classification stage, reactivating the pattern classification agent. In other words, within a multi-agent collaborative intermittent demand prediction system, prioritizing the reading of feedback signals to adjust the execution strategy allows the pattern classification agent and the model selection agent to continuously self-optimize as dynamic agents with high collaborative flexibility.

[0065] These semantically rich labels not only summarize current data patterns but also serve as a crucial hub driving the entire framework to achieve adaptive model selection. Different semantic categories directly and clearly indicate completely different demand behavior patterns, enabling the model selection agent to use these as a basis for differentiated and accurate optimal model retrieval and recommendation, achieving end-to-end adaptive linkage from data pattern recognition to prediction strategy formulation.

[0066] like Figure 12As shown, the pattern classification agent begins its work with a deep understanding and strategy planning of the input highly sensitive feature set. Its built-in large language model calls feature statistical analysis tools through the ReAct architecture to optimize the feature set. Perform analysis and calculate the meta-feature vectors that reflect the characteristics of the data. (Including statistics such as mean, variance, and dimensionality). Based on this, the large language model performs logical reasoning based on meta-features and selects an initial clustering algorithm. This decision-making logic can be expressed as: in, This represents a function that uses LLM for analysis and reasoning to generate strategy definitions. It refers to the optimal clustering algorithm determined after evaluation by large language model reasoning, covering typical methods such as K-Means, DBSCAN, and spectral clustering; This refers to the historical record library, used to store success cases.

[0067] When the initial algorithm The operation produces preliminary clustering results. Subsequently, the pattern classification agent initiates the clustering quality assessment and strategy optimization process: the system automatically calculates internal evaluation indicators such as the silhouette coefficient and the Calinski-Harabasz index, generating the optimal number of clusters K and the corresponding evaluation report. Then, the large language model integrates the report content with the original feature data for analysis, diagnoses the clustering effect, and generates optimization decision suggestions. in, This represents a function defined for evaluating clustering results using LLM inference. If the evaluation... If the result is "True", the system will proceed to the next processing stage; if it fails, the pattern classification agent will optimize accordingly. Adjust the clustering strategy and re-execute the clustering to form a self-optimizing decision-making loop.

[0068] Once the clustering results reach a stable state, each cluster... It will be transformed into a comprehensive statistical feature vector This vector covers the average demand interval. Demand variation coefficient squared Average non-zero demand Zero demand ratio Trend strength These indicators are multi-dimensional and include statistical features such as quartiles, mean, and variance. The core task of LLM is to semantically parse these multi-dimensional statistical features and map them into intuitively understandable business category labels. This semantic generation process can be formally expressed as: in, This represents a function that uses LLM for analysis and reasoning to generate semantic label definitions for each cluster. Indicates the first Semantic tags corresponding to cluster requirements K represents the optimal number of clusters generated by clustering. Specifically, if the statistical characteristics of a certain cluster are as follows: Its demand pattern can be summarized as "short intervals, small fluctuations, stable demand, and a weak growth trend." After comprehensively analyzing these characteristics, LLM will generate a semantic label for it as "high-frequency, stable-growth consumable." Finally, the pattern classification agent outputs a structured set of semantic labels. Each label originates from a deep analysis of multidimensional data patterns and includes key feature descriptions, providing a traceable and rich semantic basis for the subsequent model's decision to select agents.

[0069] S4: As Figure 1 and Figure 13 As shown, the prediction model selects an agent as the decision-making terminal for the multi-agent intermittent demand prediction method and system, responsible for integrating the upstream processing results—the highly sensitive feature set. and semantic tag set By integrating an external knowledge base through a Retrieval Enhancement Generation (RAG) mechanism, the system recommends the optimal prediction model for each demand pattern. Simultaneously, as the core feedback source of the system, it transforms the evaluation results of downstream prediction performance into structured instructions, driving the optimization of the feature extraction agent and the classification agent, forming a globally adaptive closed loop.

[0070] In some embodiments, the overall execution flow of the model selecting the agent is an iterative loop that includes knowledge retrieval, reasoning ranking, and verification feedback. The specific implementation steps can be as follows: Step 1: The model selection stage can include meta-feature construction and knowledge retrieval processes. The model selection agent process begins with the construction of a comprehensive meta-feature vector. This vector fused Statistical, time series, and deep features And category confidence. Based on meta-features. The model selects agents using the RAG mechanism as the front end for knowledge focusing, drawing from an external structured model semantic library. It performs precise semantic retrieval to recall the Top-K most relevant subset of candidate models. This step demonstrates the agent's ability to dynamically access and utilize external knowledge based on the current task context, avoiding the limitations of outdated or insufficient knowledge within large language models, and achieving adaptive expansion of recommendation sources.

[0071] Step 2: Inference-driven multi-criteria model analysis and ranking: Obtaining a subset of candidate models Then, the model selects an agent to enter the core inference loop. The large language model receives a detailed semantic description of each candidate model, including its principles, applicable scenarios, parameter requirements, historical performance examples, current meta-features, and semantic labels. Based on this, the large language model performs deep matching and multi-criteria joint inference. Finally, the large language model outputs an interpretably ranked list of models. Each model is then accompanied by a recommendation reason. This process deeply integrates data features, business semantics, and model knowledge, achieving an intelligent leap from feature matching to semantic-principle matching.

[0072] Step 3: Validation, Error Attribution and Feedback Generation: Perform rigorous cross-validation on the top-ranked models, calculate key indicators such as Mean Absolute Error (MAE), Mean Absolute Scaled Error (MASE), and MIS, and conduct error attribution analysis based on the large language model to identify the weaknesses of candidate models in specific scenarios.

[0073] If the performance of all candidate models fails to meet the threshold requirement, or if the error pattern points to a systematic upstream defect, the model selection agent will initiate a feedback optimization mechanism. It will generate a structured, actionable diagnostic report that clearly identifies the root cause of the problem and provides optimization recommendations. This report is written to the global state via a strict JSON interaction protocol. Feedback signal Once LangGraph detects this signal, it can trigger the backtracking of workflows within each agent or the backtracking of collaborative processes between multiple agents through conditional edges. For example, it can reactivate the feature extraction agent and inject the optimization instruction, thereby initiating a new round of targeted feature engineering optimization. This mechanism allows the model selection agent to transcend the role of a single recommender and become the core engine driving the entire multi-agent intermittent demand prediction system to achieve global, closed-loop, performance-oriented adaptive behavior.

[0074] Step 4: Final Recommendation Generation and Knowledge Base Evolution: After obtaining satisfactory verification results, the results output module will generate the final, interpretable recommendation report. Includes the best model The document includes confidence levels, detailed justifications, parameter configuration recommendations, and risk warnings. It also covers the meta-features of this prediction task. Final Recommendation Model Validation errors and contextual information will be stored as a new knowledge unit in the external model semantic knowledge base. This enables the system to continuously enrich its decision-making basis through continuous learning, achieving long-term adaptive capability where the recommendation strategy evolves with the accumulation of experience.

[0075] like Figure 15 As shown, in some embodiments, to avoid burdening large language models with lengthy original sparse sequences of requirements and high-dimensional features directly input, the feature extraction agent and the pattern classification agent adopt a unified strategy: first, clean the data at the tool layer. Perform in-depth statistical analysis to extract refined meta-features that characterize the essence of the data. The large language model is used as input for semantic summarization. For example, the large language model in the feature extraction agent plans feature engineering paths based on meta-features, while the pattern classification agent selects clustering algorithms based on meta-features. This design allows the large language model to focus its attention on the macroscopic statistical properties that determine data patterns, significantly reducing processing complexity.

[0076] The selection of agents for a model requires selecting from a candidate model library. We recommend suitable prediction models. To address the knowledge overload problem, we introduce Retrieval Augmentation (RAG) technology as a dynamic knowledge filtering front-end, combined with the meta-features output by the pattern classification agent. With semantic tags To construct a model recommendation mechanism for intermittent demand prediction: First, semantic retrieval is used to recall the most relevant Top-K candidate model subsets from the knowledge base. : By using the RAG mechanism to filter a subset of Top-K model candidates, the open-domain model selection problem is essentially converged to a finite and highly relevant decision space. Subsequently, the large language model does not need to process the complete knowledge base, but only needs to be based on the structured candidate model descriptions. The model performs deep comparative reasoning based on its contextual information, ultimately generating the model's recommendation results. This refers to the optimal prediction model ultimately selected for each type of feature. This mechanism ensures that the LLM inference process is always based on refined knowledge units that are highly relevant to the current task, thereby significantly improving the accuracy of model recommendations and the interpretability of decisions.

[0077] These strategies form a unified system within the framework: they are essentially task-context-oriented information preprocessors. Their adaptive nature is reflected in the fact that the input content is not a fixed template, but dynamically generated based on the upstream processing results and the current task state. Through end-to-end input engineering design, we transform the original, complex, sparse, and variable industrial data streams into refined task briefings that conform to the cognitive characteristics of LLM, thereby ensuring the comprehensive performance of the LLM-based multi-agent collaborative prediction system in terms of robustness, execution efficiency, and decision transparency.

[0078] The model selects the agent's complete decision-making process, which follows a strict ReAct paradigm. At each decision step... The ReAct loop for model agent selection can be formally represented as: in, Indicates at time step The reasoning behind this is that the model selects the agent based on the current state. and historical experience The internal analysis and logical derivation process using LLM; Indicates at time step The actions taken were based on reasoning. The conclusion is that the model selects the agent to determine the specific operation to be performed; Indicates at time step The observations. This was made after the action was carried out. Then, the agent receives feedback information or execution results from the environment or sub-agents.

[0079] System status Includes highly sensitive feature sets With semantic tag set Through the RAG mechanism, semantic knowledge bases from external models are obtained. The subset of Top-K candidate models most relevant to the current task, retrieved from the database, along with the performance metrics, error attribution analysis results, and successful model recommendation records under similarity features.

[0080] In some embodiments, this study built an experimental platform based on the DeepSeek-V3 API, and constructed an agent collaborative system by connecting to the official interface and using the LangGraph framework. This model, with its excellent reasoning, contextual understanding, and tool invocation capabilities, became the unified decision-making core for all agents in the system. Regarding parameter settings, the temperature parameter can be adjusted to 0.2 to reduce the randomness in the feature generation and policy planning stages, ensuring stable and reliable output results.

[0081] The coordinated scheduling of the entire system is implemented using the LangGraph framework. Its directed graph structure clearly defines the state transitions and interaction dependencies between sub-agents, thereby realizing the controllable complex task pipeline described in the algorithm flow. The system is deployed on a local computing node, and the hardware uses a 13th generation Intel Core i5-13600KF processor. This configuration can efficiently complete agent scheduling, business logic processing, and local feature calculation in a resource-constrained environment, ensuring the overall operating efficiency and stability of the system.

[0082] In some embodiments, this application selects three publicly available industry datasets with significant differences to systematically verify the comprehensive performance of the multi-agent framework in intermittent demand forecasting. The selection criteria are as follows: (1) Industry representativeness: covering aviation, automobile manufacturing and general manufacturing, corresponding to spare parts demand scenarios with different supply chain characteristics and management models.

[0083] (2) Pattern coverage: The three datasets together cover the spectrum of intermittent demand from extremely sparse to relatively continuous, which can comprehensively test the generalization and adaptability of the framework.

[0084] (3) Universality: The selected datasets are all benchmark data widely used in the field, ensuring that the experimental results are comparable and easy to compare with existing methods.

[0085] The datasets are briefly described below: Dataset 1: Intermittent Aircraft Spare Parts (RAF) Demand: Contains monthly demand for 5,000 types of aircraft spare parts over 84 months. Aircraft spare parts are characterized by high value, long lifespan, and extremely sparse demand. Their extremely high average demand interval (ADI) and coefficient of variation squared (CV²) are typical samples for testing the model's ability to handle highly sparse scenarios.

[0086] Dataset 2: US Automotive Parts Demand: Covers monthly demand for 2,674 parts over 51 months. 1,992 valid sequences were retained after cleaning. Automotive supply chain demand is relatively stable. The mean demand interval (ADI) and coefficient of variation (CV²) of this dataset are significantly lower than those of aviation data, making it suitable for validating the model's performance in scenarios transitioning from intermittent to continuous demand.

[0087] Dataset 3: Spare Parts Sales in the Dutch Manufacturing Industry: Contains weekly sales data for 3,451 products over 150 weeks. Demand for spare parts in the manufacturing industry is influenced by production planning and maintenance cycles, exhibiting moderate intermittency and concentrated bursts. Its demand scale and volatility are widely distributed, making it suitable for testing the framework's scalability in handling highly heterogeneous SKU pools.

[0088] To clearly present the data characteristics, Tables 2 and 3 summarize the non-zero demand mean for each dataset ( Size_mean ), standard deviation ( Size_stdCore statistics include the coefficient of variation squared (CV²) and the average demand interval (ADI).

[0089] Table 2 Core statistics of dataset 1 distributed Size_mean Size_std CV2 ADI Minimum value 1.00 0.00 0.00 3.47 Upper quartiles 1.56 0.81 0.18 7.40 median 3.83 3.06 0.46 9.38 lower quartiles 11.33 9.35 0.90 11.14 Maximum value 668.00 874.42 11.88 26.67 Table 3 shows the core statistics of dataset 2. distributed Size_mean Size_std CV2 ADI Minimum value 1 0 0 1.34 Upper quartiles 1.38 0.55 0.14 2.43 median 1.67 0.92 0.28 3.64 lower quartiles 2.12 1.36 0.43 6.38 Maximum value 15 21.14 4.32 10.2 Table 4 shows the core statistics of dataset 3. distributed Size_mean Size_std CV2 ADI Minimum value 0.31 0 0 1.01 Upper quartiles 3.66 2.97 0.42 3.85 median 8.99 7.84 0.78 7.89 lower quartiles 22.91 21.76 1.36 15 Maximum value 1132.21 853.5 14.99 25 A comprehensive comparison of the above statistical characteristics reveals the following: In terms of demand scale distribution, both the RAF and Dutch datasets exhibit significant long-tail characteristics, with a small number of highly heterogeneous SKUs dominating the demand, while the US automotive dataset shows a relatively uniform distribution. Regarding intermittency and volatility, the RAF dataset has the longest demand intervals and the greatest volatility, representing an extremely intermittent type; the Dutch dataset falls in the middle; and the US automotive dataset is the most stable. These three datasets together constitute a complete demand spectrum ranging from stable and continuous to highly intermittent, providing a sufficient data foundation for system validation of the prediction accuracy, interpretability, and stability of the multi-agent framework in different scenarios.

[0090] like Figure 14 As shown, the radar chart constructed using seven core metrics visually presents the performance of each module on the three datasets, covering anomaly detection recall, data consistency score, feature importance score, clustering algorithm selection accuracy, semantic label evaluation accuracy, Top-1 recommendation accuracy, and recommendation confidence.

[0091] The classification agent is responsible for transforming demand patterns into actionable semantic labels. Experiments show that its clustering algorithm selection accuracy exceeds 80% on three datasets (85.2% for US automotive data), indicating its ability to intelligently match clustering methods based on data characteristics; the semantic label evaluation accuracy reaches 88.3% (RAF), 92.1% (US automotive), and 85.7% (Netherlands-made), respectively. These labels, generated by a large language model based on a three-layer semantic framework of frequency-stability-scale, balance machine readability and business interpretability, effectively improving the acceptability of the model output in real-world scenarios.

[0092] The model selection agent, serving as the central decision-making framework, consistently achieved a Top-1 recommendation accuracy of 76%-79% across three datasets (79.1% being the best on the US automotive dataset), with recommendation confidence remaining high at 80%-85%, demonstrating its ability to consistently provide reliable recommendations in scenarios with fluctuating and intermittent demand. This module also exhibited an average feedback optimization success rate of 17% (18.5% on the Dutch manufacturing dataset), indicating the system's ability to automatically correct suboptimal decisions. Through a retrieval-enhanced generation mechanism, this agent not only outputs model recommendations but also provides traceable reasoning chains and attribution analysis, significantly enhancing decision transparency.

[0093] Analysis of both radar charts and fine-grained indicators reveals that the four agents form an organic whole through dynamic collaboration: the pattern classification agent performs semantic parsing of patterns, while the model selection agent delivers accurate and reliable decision outputs. This multi-agent "divide and conquer-coordinate" architecture provides a systematic solution to address the complex uncertainties of intermittent demand forecasting.

[0094] The transparency achieved by forcibly generating explicit reasoning trajectories in this embodiment becomes a driving force for system robustness: the process of converting numerical clustering into business semantic labels forces the system to establish logical decision-making basis; and the closed-loop design of downstream agents guiding upstream optimization through feedback mechanisms is a self-evolution capability that black-box models cannot achieve. This logic-driven transparent decision-making not only enhances business trust but also gives the system the vitality of self-correction.

[0095] like Figure 16 As shown, in some embodiments, this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the numerical simulation, analysis, or control steps in any of the methods described above.

[0096] At the hardware level, this electronic device includes a processor, internal bus, network interface, memory, and non-volatile storage, and may also include other hardware required for the business logic. The processor reads the corresponding computer program from the non-volatile storage into memory and then executes it to achieve the above-mentioned functions. Figure 1 This specification does not exclude other implementation methods besides software implementation, such as logic devices or a combination of hardware and software. In other words, the execution entity of the following processing flow is not limited to individual logic units; it can also be hardware or logic devices. It is understood that by simply performing some logic programming on the method flow using a hardware description language and programming it into an integrated circuit, the hardware circuit implementing the logic method flow can be obtained.

[0097] The above method can be implemented by a controller in any suitable manner. For example, the controller can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, microcontrollers. A memory controller can also be implemented as part of the control logic of a memory. It is understood that, in addition to implementing the controller in purely computer-readable program code form, the same functionality can be achieved by logically programming the method steps to make the controller take the form of logic gates, switches, ASICs, programmable logic controllers, and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the means included therein for implementing various functions can also be considered as structures within the hardware component. Or even, the means for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0098] The systems, devices, modules, or units in the above embodiments can be implemented by computer chips or entities, or by products with certain functions.

[0099] For ease of description, the above devices are described in terms of function, divided into various modules. Of course, in implementing this specification, the functions of each module can be implemented in one or more software and / or hardware.

[0100] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this specification may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0101] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1A device that provides the functions specified in one or more boxes.

[0102] In some embodiments, this application may also provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the prediction method described above.

[0103] In some embodiments, these computer program instructions may also be stored in a computer-readable storage medium capable of directing a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means implemented in a process. Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0104] In some embodiments, these computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0105] In some embodiments, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0106] In some embodiments, memory may include non-persistent memory in a computer-readable medium, such as random access memory (RAM) and / or non-volatile memory, like read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0107] In some embodiments, computer-readable media, including permanent and non-permanent, removable and non-removable media, can be used to store information by any method or technology. The information can be computer-readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0108] Those skilled in the art will readily understand that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for pattern classification and model selection in a multi-agent collaborative intermittent demand forecasting system, characterized in that, The pattern classification and model selection method includes a pattern classification stage and a model selection stage. In the pattern classification stage, the pattern classification agent transforms the input high-sensitivity feature set into a semantic category label set through cluster analysis and semantic inference. The model selection agent, in the model selection stage, uses the semantic category label set as the pattern basis, integrates the highly sensitive feature set and the semantic category label set, and fuses feature and semantic information through a retrieval enhancement generation mechanism to output the recommendation scheme of the prediction model. Specifically, when a deviation is detected in the model selection stage, a feedback signal is generated and input into the pattern classification stage, and the feedback signal is read first to adjust the execution strategy of the pattern classification.

2. The method for pattern classification and model selection in a multi-agent collaborative intermittent demand forecasting system according to claim 1, characterized in that, The pattern classification stage includes feature preprocessing and preliminary analysis, reasoning-driven clustering strategy selection and iterative optimization, and semantic label generation and structured output steps. In the feature preprocessing and preliminary analysis steps, the input high-sensitivity feature set is standardized and quantitatively diagnosed, and the meta-features of the high-sensitivity feature set are extracted. The meta-features of the high-sensitivity feature set include variance, quantiles, sparsity and correlation density. In the inference-driven clustering strategy selection and iterative optimization step, the large language model in the pattern classification step infers based on the meta-features of the highly sensitive feature set and selects an initial clustering algorithm, performs clustering analysis, until a high-quality clustering structure is obtained. After obtaining a stable high-quality clustering structure, during the semantic tag generation and structured output process, key statistical features are aggregated to form a comprehensive feature vector. Based on the comprehensive feature vector and the preset tag generation template, the large language model performs semantic reasoning and generation to produce business semantic tags and output a structured set of semantic category tags.

3. The method for pattern classification and model selection in a multi-agent collaborative intermittent demand forecasting system according to claim 2, characterized in that, In the pattern classification step, the decision-making process at each decision step is represented as follows: in, Indicates at time step The reasoning is based on the current state. and historical experience The internal analysis and logical deduction process using the large language model, Indicates at time step The actions taken were based on reasoning. The conclusion determines the specific operations to be carried out. Indicates at time step The observations were made during the execution of the action. Subsequently, the feedback signals or execution results obtained from the outside world, It includes the currently described high-sensitivity feature set, the current clustering results, the preliminary clustering results from the previous round, and a comprehensive evaluation report of the previous round's clustering results. This indicates historical clustering experience used to record statistical characteristics similar to the current features.

4. The method for pattern classification and model selection of a multi-agent collaborative intermittent demand forecasting system according to claim 3, characterized in that, The pattern classification also outputs feedback signals to the steps prior to the pattern classification step of the multi-agent collaborative intermittent demand forecasting system, so as to change the execution order of the workflow of the multi-agent collaborative intermittent demand forecasting system.

5. A method for pattern classification and model selection in a multi-agent collaborative intermittent demand forecasting system according to any one of claims 1 to 4, characterized in that, The model selection steps include meta-feature construction and knowledge retrieval, reasoning-driven multi-criteria model analysis and ranking, verification error attribution and feedback generation, generation of recommendations and knowledge base evolution; In the meta-feature construction and knowledge retrieval steps, a meta-feature vector is constructed, and based on the meta-feature vector, a precise semantic retrieval is performed to recall a subset of candidate models. The meta-feature vector integrates the statistical, temporal, and deep features of the high-sensitivity feature set, the semantic label set, and the category confidence. In the reasoning-driven multi-criteria model analysis and ranking step, the large language model receives a detailed semantic description of each subset of candidate models, and then performs deep matching and multi-criteria joint reasoning to output a list of candidate models that has been interpreted and ranked. The detailed semantic description includes the principle, applicable scenarios, parameter requirements, historical performance cases, current meta-features, and semantic category labels. In the verification error attribution and feedback generation step, the candidate models ranked first are cross-validated, and error attribution analysis is performed based on the large language model. In the recommendation generation and knowledge base evolution steps, an interpretable recommendation report is output, which includes the best model, confidence level, detailed reasons, parameter configuration suggestions, and risk warnings.

6. The method for pattern classification and model selection in a multi-agent collaborative intermittent demand forecasting system according to claim 5, characterized in that, If, during the verification error attribution and feedback generation step, the performance of all the candidate models fails to meet the threshold requirement, or the error pattern points to a systematic upstream defect, a structured and actionable diagnostic report is generated, and the feedback signal is issued.

7. The method for pattern classification and model selection in a multi-agent collaborative intermittent demand forecasting system according to claim 6, characterized in that, In the meta-feature construction and knowledge retrieval steps, an external model semantic knowledge base is integrated through a retrieval enhancement generation mechanism. The external model semantic knowledge base is accurately retrieved to recall the subset of candidate models most relevant to the current task. After the recommendation generation and knowledge base evolution steps, the meta-feature vector, the final recommendation model, the verification error, and the context are stored as a new knowledge unit in the external model semantic knowledge base.

8. The method for pattern classification and model selection of a multi-agent collaborative intermittent demand forecasting system according to claim 7, characterized in that, At every decision step The decision steps in the model selection phase are represented as follows: Among them, system status It includes the highly sensitive feature set and the semantic label set, the candidate model subset, the performance indicators of the candidate models, the error attribution analysis results, and the successful model recommendation records under similarity features.

9. A multi-agent collaborative intermittent demand forecasting system, characterized in that, This includes data monitoring agents, feature extraction agents, pattern classification agents, and model selection agents; The data monitoring agent is used to transform the input raw intermittent demand sequence containing mixed noise into a high-quality dataset; The feature extraction agent is used to transform the high-quality dataset into a highly sensitive feature set; In the pattern classification stage, the intelligent agent transforms the input highly sensitive feature set into a semantic category label set through cluster analysis and semantic inference. The model selection agent uses the semantic category label set as the pattern basis during the model selection stage, integrates the high-sensitivity feature set and the semantic category label set, and fuses feature and semantic information through a retrieval enhancement generation mechanism to output the recommendation scheme of the prediction model. Specifically, when a deviation is detected in the model selection stage, a feedback signal is generated and input into the pattern classification stage, and the feedback signal is read first to adjust the execution strategy of the pattern classification.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the pattern classification and model selection method as described in any one of claims 1-8.

Images