Dynamic rule-based multimodal inventory resonance management method, medium and equipment

Abstract

The present invention discloses a multimodal transport inventory resonance management method, medium and equipment based on dynamic rules. The method includes: classifying and collecting raw data and preprocessing it into standardized intermodal transport data; constructing a transport time prediction model based on the preprocessed second transport state parameter to extract dynamic transport characteristics, and performing spectral analysis on the second inventory level parameter to obtain inventory cycle fluctuation characteristics; extracting the resonance coupling characteristics of the second transport state parameter and the second inventory level parameter through time-frequency correlation analysis and establishing a feature coding library; constructing a mapping relationship network between transport mode switching rules and inventory adjustment strategies based on the above characteristics; establishing a resonance control parameter conversion relationship between a baseline transport scenario and a target transport scenario; and finally constructing a multidimensional resonance management knowledge base. The present invention realizes the coordinated control of transport delays and inventory fluctuations under multimodal transport, effectively suppresses the supply chain resonance effect, and improves the system's responsiveness and stability.

Description

Technical Field

[0001] The present invention relates to the technical field of supply chain collaborative optimization, and in particular to a multimodal transport inventory resonance management method, medium and equipment based on dynamic rules. Background Art

[0002] In modern supply chain management, multimodal transport is widely used for bulk cargo transportation due to its flexibility and cost advantages. However, this transportation model involves the coordination of multiple modes, including road, rail, and water transportation. Its dynamic and complex nature presents significant challenges for inventory management. Traditional methods often use static rules for inventory replenishment and transportation scheduling, such as setting safety stocks based on historical average transportation times or implementing fixed replenishment cycles. While simple and feasible, this approach struggles to adapt to operational uncertainties such as fluctuating transportation times and volatile demand.

[0003] Existing research has made some progress in transportation optimization and inventory control, but the two are often treated as independent problems. Transportation scheduling models usually focus on route selection and cost minimization, while inventory models focus on demand forecasting and replenishment strategies. This separation results in the system lacking an effective coordinated response mechanism when faced with sudden delays or surges in demand. For example, when a transportation route is interrupted, traditional methods find it difficult to adjust inventory strategies in a timely manner to cope with delivery delays, which may trigger a chain reaction. In addition, the differences in data collection standards and timeliness of different transportation modes further increase the difficulty of overall collaborative optimization. How to achieve dynamic matching of transportation resources and inventory levels has become a key issue in improving the responsiveness of the supply chain. Summary of the Invention

[0004] In view of the above problems, the present invention provides a multimodal transport inventory resonance management method, medium and equipment based on dynamic rules. Through time-frequency correlation analysis of transportation and inventory and dynamic strategy mapping, it realizes supply chain collaborative optimization and solves the problem of mutual amplification of transportation delays and inventory fluctuations in multimodal transport.

[0005] To achieve the above objectives, in a first aspect, the present application provides a multimodal transport inventory resonance management method based on dynamic rules, comprising:

[0006] Collecting raw data according to the transportation mode classification, the raw data including a first transportation status parameter, a first inventory level parameter, and a first demand fluctuation parameter;

[0007] Preprocessing the original data to obtain standardized intermodal transport data, the standardized intermodal transport data including a second transport state parameter, a second inventory level parameter, and a second demand fluctuation parameter after time series alignment;

[0008] A transportation timeliness prediction model is constructed based on the second transportation status parameter to extract dynamic transportation characteristics, including the delay probability distribution and switching response time of each transportation mode. In addition, a spectral analysis is performed on the second inventory level parameter to obtain inventory cycle fluctuation characteristics.

[0009] Performing a time-frequency correlation analysis on the second transportation state parameter and the second inventory level parameter to extract resonance coupling features, the resonance coupling features including a transportation delay propagation coefficient and a demand amplification factor, and establishing a resonance feature encoding library containing typical resonance modes, the typical resonance modes including at least one of a bullwhip effect, seasonal fluctuations, and sudden interruptions;

[0010] Based on dynamic transportation characteristics, inventory cycle fluctuation characteristics and resonance coupling characteristics, a mapping relationship network between transportation mode switching rules and inventory adjustment strategies is constructed;

[0011] Based on the dynamic transport characteristics and resonance coupling characteristics, the resonance control parameter conversion relationship between the baseline transport scenario and the target transport scenario is established;

[0012] Dynamic transportation characteristics, inventory cycle fluctuation characteristics, resonance feature coding library, mapping relationship network and resonance control parameter conversion relationship are dynamically integrated to construct a multidimensional resonance management knowledge base including transportation characteristics, inventory characteristics, resonance characteristics and control strategies.

[0013] Furthermore, the original data is preprocessed to obtain standardized intermodal transport data. The standardized intermodal transport data includes the second transport status parameter, the second inventory level parameter, and the second demand fluctuation parameter after time series alignment, including:

[0014] Selecting transport node data that meets a preset integrity threshold from the first transport status parameter as the second transport status parameter, where the preset integrity threshold includes a continuity index of GPS positioning and a sampling rate index of the sensor;

[0015] Extracting, from the first inventory level parameter, inventory records that match the collection period of the second transportation status parameter as the second inventory level parameter, wherein the time deviation of the collection period during matching does not exceed a preset inventory synchronization window;

[0016] Selecting order data corresponding to the second transport status parameter collection period from the first demand fluctuation parameter as the second demand fluctuation parameter;

[0017] Establishing a transport-inventory-demand association relationship between the second transport status parameter, the second inventory level parameter, and the second demand fluctuation parameter, and generating a unified transport batch identifier;

[0018] Extract multimodal transport features from the second transport status parameter to eliminate the heterogeneity of data collected by different transport modes;

[0019] Perform outlier correction on the second inventory level parameter to retain the inventory change characteristics within the effective fluctuation range;

[0020] Generate standardized intermodal data.

[0021] Furthermore, a transportation time efficiency prediction model is constructed based on the second transportation state parameter to extract dynamic transportation characteristics. The dynamic transportation characteristics include the delay probability distribution and switching response time of each transportation mode, including:

[0022] Construct a transportation time efficiency prediction model. Select the time efficiency data of multiple transportation modes from the second transportation status parameter as the training input of the transportation time efficiency prediction model, and generate the time efficiency feature vector of each transportation mode. The time efficiency feature vector includes the historical delay rate and real-time operation status.

[0023] A road transport delay prediction sub-model is constructed based on the time efficiency feature vector to extract the road transport delay probability distribution, which includes the average delay duration and delay frequency of different road sections.

[0024] A railway transport delay prediction sub-model is constructed based on the timeliness feature vector to extract the railway transport delay probability distribution, which includes the fluctuation range of the on-time rate of each train and the probability of scheduling anomalies.

[0025] A water transport delay prediction sub-model is constructed based on the time-efficiency feature vector to extract the probability distribution of water transport delay, which includes the port operation delay time and the waterway congestion coefficient.

[0026] Calculate the switching response time between road transport, rail transport and water transport, including the transfer connection time from road to rail, the loading and unloading waiting time from rail to water transport, and the start delay time of emergency transport;

[0027] The delay probability distributions of road transport, railway transport and water transport are integrated with the switching response time to generate dynamic transport features.

[0028] Furthermore, a spectrum analysis is performed on the second inventory level parameter to obtain inventory cycle fluctuation characteristics, including:

[0029] Extracting multi-time scale inventory change series from the second inventory level parameter to obtain original inventory fluctuation data, where the multi-time scales include daily fluctuation cycle, weekly fluctuation cycle and seasonal fluctuation cycle;

[0030] Perform wavelet transform on the original inventory fluctuation data to decompose it into inventory fluctuation components of different frequency bands. The inventory fluctuation components include high-frequency random fluctuation components, medium-frequency business cycle components, and low-frequency trend components.

[0031] Perform extreme value statistical analysis on high-frequency random fluctuation components to identify sudden fluctuation events, calculate the probability density function and autocorrelation characteristics of sudden fluctuation events, and generate sudden fluctuation characteristic parameters;

[0032] Based on the sudden fluctuation characteristic parameters, an inventory abnormal fluctuation warning model is constructed. The abnormal fluctuation warning model includes fluctuation thresholds and warning response rules.

[0033] Identify typical inventory fluctuation patterns from medium-frequency business cycle components, including purchase cycle fluctuations, sales cycle fluctuations, and replenishment cycle fluctuations, and embed sudden fluctuation characteristic parameters as correction factors into typical inventory fluctuation patterns;

[0034] Extract trend characteristic parameters based on low-frequency trend components, including seasonal index and long-term growth rate after correction for sudden fluctuations;

[0035] The inventory fluctuation components, typical inventory fluctuation patterns, sudden fluctuation characteristic parameters and trend characteristic parameters are integrated to generate a multi-dimensional inventory cycle fluctuation characteristic matrix containing normal fluctuations and abnormal fluctuations;

[0036] The multi-dimensional inventory cycle fluctuation characteristic matrix is ​​standardized to obtain the standardized inventory cycle fluctuation characteristics.

[0037] Furthermore, a time-frequency correlation analysis is performed on the second transportation state parameter and the second inventory level parameter to extract resonance coupling features. The resonance coupling features include a transportation delay propagation coefficient and a demand amplification factor. A resonance feature encoding library containing typical resonance modes is established. The typical resonance modes include at least one of a bullwhip effect, seasonal fluctuations, and sudden interruptions, including:

[0038] Time-align the transportation delay event sequence in the second transportation status parameter with the inventory fluctuation sequence in the second inventory level parameter to establish a delay-fluctuation correlation time axis;

[0039] Perform cross-spectral analysis on the delay-fluctuation correlation time axis, calculate the coherent spectral density of transportation delay and inventory fluctuation, and extract the transportation delay propagation coefficient, which represents the inventory fluctuation amplitude caused by unit transportation delay;

[0040] Perform Granger causality analysis on the second demand fluctuation parameter and the second inventory level parameter to determine the transmission delay and amplification factor from demand fluctuation to inventory fluctuation, and generate a demand amplification factor.

[0041] Identify typical resonance patterns from the latency-fluctuation correlation timeline, including bullwhip effects, seasonal fluctuations, and sudden disruptions;

[0042] Parameterized modeling of typical resonance modes is performed to generate resonance feature encoding vectors, which contain the combined characteristics of transport delay propagation coefficients and demand amplification factors.

[0043] The resonance feature encoding vector is associated with the transport batch identifier and stored to form a structured resonance feature;

[0044] and, constructing a resonance intensity prediction model based on the resonance feature encoding vector to generate a resonance risk level as a supplementary resonance feature;

[0045] The supplementary resonance features and structured resonance features are organized into a resonance feature coding library.

[0046] Furthermore, based on the dynamic transportation characteristics, inventory cycle fluctuation characteristics, and resonance coupling characteristics, a mapping relationship network between transportation mode switching rules and inventory adjustment strategies is constructed, including:

[0047] Extracting timeliness risk indicators for each mode of transport from dynamic transport characteristics. These indicators include delay probability distribution curves and switching response time thresholds.

[0048] Analyze the key fluctuation parameters in inventory cycle fluctuation characteristics, including safety stock threshold and replenishment cycle sensitivity;

[0049] Determining resonance control parameters based on resonance coupling characteristics, the resonance control parameters including a transport delay propagation coefficient critical value and a demand amplification factor tolerance;

[0050] Conduct multi-dimensional correlation analysis on timeliness risk indicators, key fluctuation parameters and resonance control parameters to establish a three-dimensional decision space of transportation-inventory-resonance;

[0051] A transport mode switching rule set is constructed in the three-dimensional decision space of transport-inventory-resonance. The transport mode switching rule set includes the transport mode priority sorting and switching trigger conditions under different risk levels.

[0052] Based on the resonance control parameters, an inventory adjustment strategy matrix is ​​generated in the transportation-inventory-resonance three-dimensional decision space. The inventory adjustment strategy matrix includes a dynamic safety stock calculation formula and an emergency replenishment trigger mechanism.

[0053] Co-optimize the transportation mode switching rule set and the inventory adjustment strategy matrix to generate a strategy feature coding library containing typical scenario response solutions;

[0054] Also, conduct visual modeling of the strategy feature coding library and construct a strategy decision tree as a supplementary control basis;

[0055] The strategy decision tree, transportation mode switching rule set and inventory adjustment strategy matrix are integrated to form a mapping relationship network of transportation-inventory linkage.

[0056] Furthermore, based on the dynamic transport characteristics and the resonance coupling characteristics, a resonance control parameter conversion relationship between the baseline transport scenario and the target transport scenario is established, including:

[0057] Extracting baseline transport time efficiency characteristic parameters of the baseline transport scenario and target transport time efficiency characteristic parameters of the target transport scenario from the dynamic transport characteristics, wherein the baseline transport time efficiency characteristic parameters include baseline delay probability distribution and baseline switching response time, and the target transport time efficiency characteristic parameters include target delay probability distribution and target switching response time;

[0058] Extracting baseline resonance characteristic parameters of the baseline transport scenario and target resonance characteristic parameters of the target transport scenario from the resonance coupling characteristics, the baseline resonance characteristic parameters including a baseline delay propagation coefficient and a baseline demand amplification factor, and the target resonance characteristic parameters including a target delay propagation coefficient and a target demand amplification factor;

[0059] Compare and analyze the baseline transport time characteristic parameters with the target transport time characteristic parameters, and calculate the equivalent conversion coefficient of the transport delay probability and the scenario adjustment factor of the switching response time;

[0060] Matching and calibrating the reference resonance characteristic parameters with the target resonance characteristic parameters to determine the scenario scaling ratio of the delay propagation coefficient and the scenario correction parameters of the demand amplification factor;

[0061] A resonance control parameter conversion model is constructed based on the equivalent conversion coefficient, scene adjustment factor, scene scaling ratio, and scene correction parameter. The resonance control parameter conversion model includes a scene conversion algorithm for transportation timeliness characteristics and a scene adaptation rule for resonance characteristics.

[0062] The characteristic parameters of the benchmark transport scenario are input into the resonance control parameter conversion model for parameter conversion, and the optimized control parameters of the target transport scenario are output;

[0063] Verify the scenario applicability of the optimized control parameters and generate conversion relationship correction coefficients including scenario difference compensation terms;

[0064] A resonance control parameter conversion relationship is generated according to the resonance control parameter conversion model and the conversion relationship correction coefficient.

[0065] Furthermore, dynamic transportation characteristics, inventory cycle fluctuation characteristics, resonance feature encoding library, mapping relationship network and resonance control parameter conversion relationship are dynamically integrated to construct a multi-dimensional resonance management knowledge base containing transportation characteristics, inventory characteristics, resonance characteristics and control strategies, including:

[0066] Extracting multimodal transport characteristic data from dynamic transport characteristics, the multimodal transport characteristic data including road transport characteristic parameters, railway transport characteristic parameters and water transport characteristic parameters;

[0067] Extracting multi-time scale inventory characteristic data from inventory cycle fluctuation characteristics, the multi-time scale inventory characteristic data includes daily fluctuation characteristic parameters, weekly fluctuation characteristic parameters and seasonal fluctuation characteristic parameters;

[0068] Extracting typical resonance mode feature data from a resonance feature coding library, wherein the typical resonance mode feature data includes bullwhip effect feature parameters, seasonal resonance feature parameters and sudden interruption conduction feature parameters;

[0069] Extracting transportation-inventory linkage strategy data from the mapping relationship network, the transportation-inventory linkage strategy data including transportation mode switching rule parameters and inventory adjustment strategy parameters;

[0070] Extracting multi-scenario conversion parameter data from the resonance control parameter conversion relationship, the multi-scenario conversion parameter data including reference scenario feature parameters and target scenario optimization parameters;

[0071] The multimodal transport feature data, multi-time scale inventory feature data, typical resonance mode feature data, transport-inventory linkage strategy data and multi-scenario conversion parameter data are fused to generate a multi-dimensional resonance feature matrix.

[0072] The knowledge base storage structure is constructed based on the multi-dimensional resonance feature matrix. The knowledge base storage structure includes:

[0073] Transport characteristic storage module, used to store and manage multimodal transport characteristic data;

[0074] Inventory feature storage module, used to store and manage multi-time scale inventory feature data;

[0075] A resonance characteristic storage module, used for storing and managing characteristic data of typical resonance modes;

[0076] Policy rule storage module, used to store and manage transportation-inventory linkage policy data;

[0077] A scene conversion storage module is used to store and manage multi-scene conversion parameter data;

[0078] Optimize the index of the knowledge base storage structure, establish a fast retrieval mechanism for feature parameters and a dynamic call interface for policy rules;

[0079] The multidimensional resonance feature matrix is ​​integrated with the knowledge base storage structure to form a multidimensional resonance management knowledge base including transportation feature dimension, inventory feature dimension, resonance feature dimension and control strategy dimension.

[0080] In a second aspect, the present invention further provides a computer-readable storage medium storing computer program instructions, which implement the method described in the first aspect when executed by a processor.

[0081] In a third aspect, the present invention further provides an electronic device comprising a memory and a processor, wherein the memory is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method described in the first aspect.

[0082] Different from the existing technology, the above technical solution provides a multimodal transport inventory resonance management method, medium and equipment based on dynamic rules. The method includes: classifying and collecting the first transportation state parameter, the first inventory level parameter and the first demand fluctuation parameter and preprocessing them into standardized intermodal transport data; constructing a transportation time prediction model based on the second transportation state parameter to extract dynamic transportation characteristics, and at the same time performing spectral analysis on the second inventory level parameter to obtain inventory cycle fluctuation characteristics; extracting the resonance coupling characteristics of the second transportation state parameter and the second inventory level parameter through time-frequency correlation analysis and establishing a feature coding library; constructing a mapping relationship network between transportation mode switching rules and inventory adjustment strategies based on the above characteristics; establishing a resonance control parameter conversion relationship between the baseline transportation scenario and the target transportation scenario; and finally integrating multidimensional features and rules to construct a multidimensional resonance management knowledge base. The present invention realizes the coordinated control of transportation delays and inventory fluctuations under multimodal transport through dynamic coupling analysis of transportation and inventory, effectively suppresses the resonance effect of the supply chain, and improves the system's responsiveness and stability.

[0083] The above-mentioned records related to the content of the invention are only an overview of the technical solution of this application. In order to enable ordinary technicians in this field to understand the technical solution of this application more clearly, and then implement it according to the text of the specification and the contents recorded in the drawings, and to make the above-mentioned purposes and other purposes, features and advantages of this application easier to understand, the following is an explanation in combination with the specific implementation methods and drawings of this application. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] The drawings are only used to illustrate the principles, implementation methods, applications, features and effects of the specific embodiments of the present invention and other related contents, and are not to be considered as limiting the present application.

[0085] In the drawings of the specification:

[0086] Figure 1 A method step diagram showing steps S101 to S107 of the method described in the specific embodiment;

[0087] Figure 2 A method step diagram showing steps S201 to S207 of the method described in the specific embodiment;

[0088] Figure 3 A method step diagram showing steps S301 to S306 of the method described in the specific embodiment;

[0089] Figure 4 A method step diagram showing steps S401 to S408 of the method described in the specific embodiment;

[0090] Figure 5 Schematic diagram of the structure of the electronic device described in the specific implementation.

[0091] The reference numerals in the above drawings are described as follows:

[0092] 1. Electronic equipment;

[0093] 11. Memory;

[0094] 12. Processor. DETAILED DESCRIPTION

[0095] In order to explain in detail the possible application scenarios, technical principles, specific solutions that can be implemented, and the purpose and effects of this application, the following is a detailed description of the specific embodiments listed in conjunction with the accompanying drawings. The embodiments described herein are only used to more clearly illustrate the technical solutions of this application and are therefore only examples and are not intended to limit the scope of protection of this application.

[0096] References to "embodiments" herein mean that the specific features, structures, or characteristics described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the word "embodiment" in various places in the specification does not necessarily refer to the same embodiment, nor does it particularly limit its independence or relevance to other embodiments. In principle, in this application, as long as there are no technical contradictions or conflicts, the various technical features mentioned in the embodiments can be combined in any manner to form a corresponding implementable technical solution.

[0097] Unless otherwise defined, the technical terms used herein have the same meanings as those generally understood by those skilled in the art to which this application belongs; the use of relevant terms herein is only for describing specific embodiments and is not intended to limit this application.

[0098] In the description of this application, the term "and / or" is used to describe a logical relationship between objects, indicating that three relationships can exist. For example, A and / or B means: A exists, B exists, and both A and B exist. In addition, the character " / " in this document generally indicates that the objects before and after are in a logical "or" relationship.

[0099] In this application, terms such as "first" and "second" are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual quantity, priority or sequence relationship between these entities or operations.

[0100] Without further limitations, in this application, the words "include", "comprise", "have" or other similar open-ended expressions used in sentences are intended to cover non-exclusive inclusion. These expressions do not exclude the presence of additional elements in the process, method or product that includes the elements, so that the process, method or product that includes a series of elements may include not only those defined elements, but also other elements that are not explicitly listed, or also include elements inherent to such process, method or product.

[0101] Consistent with the understanding in the Examination Guidelines, in this application, expressions such as "greater than," "less than," and "exceed" are understood to exclude the number itself; expressions such as "above," "below," and "within" are understood to include the number itself. Furthermore, in the description of the embodiments of this application, "multiple" means two or more (including two), and similar expressions related to "multiple," such as "multiple groups" and "multiple times," are also understood in this manner, unless otherwise specifically defined.

[0102] See also Figure 1 In a first aspect, this embodiment provides a multimodal transport inventory resonance management method based on dynamic rules, comprising:

[0103] S101. Collecting raw data by transportation mode, where the raw data includes a first transportation status parameter, a first inventory level parameter, and a first demand fluctuation parameter;

[0104] S102: Preprocess the original data to obtain standardized intermodal transport data, where the standardized intermodal transport data includes a second transport status parameter, a second inventory level parameter, and a second demand fluctuation parameter after time series alignment.

[0105] S103: Build a transportation timeliness prediction model based on the second transportation status parameter, extract dynamic transportation characteristics, including delay probability distribution and switching response time of each transportation mode, and perform spectral analysis on the second inventory level parameter to obtain inventory cycle fluctuation characteristics;

[0106] S104. Performing a time-frequency correlation analysis on the second transportation state parameter and the second inventory level parameter to extract resonance coupling features, where the resonance coupling features include a transportation delay propagation coefficient and a demand amplification factor, and establishing a resonance feature encoding library containing typical resonance modes, where the typical resonance modes include at least one of a bullwhip effect, seasonal fluctuations, and sudden interruptions.

[0107] S105. Construct a mapping relationship network between transportation mode switching rules and inventory adjustment strategies based on dynamic transportation characteristics, inventory cycle fluctuation characteristics, and resonance coupling characteristics;

[0108] S106. Establishing a resonance control parameter conversion relationship between a reference transport scenario and a target transport scenario based on the dynamic transport characteristics and the resonance coupling characteristics;

[0109] S107. Dynamically integrate the dynamic transportation characteristics, inventory cycle fluctuation characteristics, resonance feature coding library, mapping relationship network and resonance control parameter conversion relationship to build a multi-dimensional resonance management knowledge base including transportation characteristics, inventory characteristics, resonance characteristics and control strategies.

[0110] In step S101, raw data forms the fundamental input for multimodal inventory resonance management. Preferably, the first transport status parameter, including real-time location, speed, and equipment status information for each mode of transport, is collected through the transport management system. The first inventory level parameter, recording inventory quantity and turnover cycle at each node, is obtained through the warehouse management system. The first demand fluctuation parameter, collected through the sales system, reflects the dynamic characteristics of market demand. This data is synchronized with the business system in real time via IoT devices.

[0111] Furthermore, the raw data collection process adopts an adaptive sampling strategy, setting differentiated collection frequencies for different modes of transportation, which can both ensure data timeliness and reduce system load; the collection of the first demand fluctuation parameter pays special attention to the increase in data density during promotional activities and seasonal transitions, and dynamically adjusts the collection granularity through a sliding time window.

[0112] In step S102, standardized intermodal data is structured through data cleansing and time series alignment. The second transport status parameter is the result of removing outliers and normalizing the original transport data; the second inventory level parameter undergoes inventory dimension unification and missing value filling; and the second demand fluctuation parameter undergoes seasonal adjustment and smoothing. Time series alignment is achieved through unified timestamps, ensuring comparability of data across different systems.

[0113] Furthermore, the standardization process introduces a mechanism to retain the characteristics of the transport mode. In addition to the conventional standardization of the second transport state parameters, it also retains the key indicator dimensions unique to each transport mode, such as the route weather impact factor of air transport and the real-time traffic congestion index of road transport. The time series alignment adopts a dynamic time warping algorithm, which can effectively deal with the natural time scale differences between different transport modes, especially for the time series matching problems of long-cycle transport modes such as sea transport and short-cycle transport such as express delivery.

[0114] In step S103, the transportation time efficiency prediction model is constructed using the time series analysis method. Preferably, the delay probability distribution in the dynamic transportation characteristics is obtained through statistics of historical transportation delay events to characterize the reliability of each transportation route; the switching response time is calculated through transportation mode conversion cases to reflect the efficiency of multimodal transport connection; the inventory cycle fluctuation characteristics are extracted through Fourier transform to identify the periodic laws of inventory changes, including the main fluctuation cycles and their amplitude characteristics.

[0115] In step S104, time-frequency correlation analysis is performed using wavelet transforms. The transport delay propagation coefficient quantifies the transmission strength of transport delays on inventory fluctuations, and the demand amplification factor calculates the degree of distortion of demand signals within the supply chain. A resonance feature encoding library is constructed using a pattern recognition algorithm. The bullwhip effect pattern identifies the characteristics of the demand signal's progressive amplification, the seasonal fluctuation pattern extracts periodic inventory resonance patterns, and the sudden interruption pattern records the impact characteristics of abnormal events.

[0116] In step S105, a mapping network is constructed using a decision tree algorithm. This algorithm associates the delay probability in dynamic transportation characteristics with the inventory safety threshold. Transportation mode priority rules are generated based on the switching response time, and a replenishment timing strategy is formulated based on the inventory cycle fluctuation characteristics. Nodes in the mapping network represent decision rules, and edges represent conditional jump relationships.

[0117] Furthermore, the construction of the mapping relationship network can introduce a fuzzy reasoning mechanism to handle the nonlinear relationship between dynamic transportation characteristics and inventory cycle fluctuation characteristics; the decision rules in the mapping relationship network can set a dynamic confidence threshold, and automatically trigger the manual review process when the credibility of the input feature is lower than the threshold; in particular, for special transportation categories such as cold chain, the mapping relationship network will load preset temperature control constraints to ensure that the generated strategy meets the requirements for item preservation.

[0118] In step S106, a resonance control parameter conversion relationship is established through scenario comparison analysis. The baseline transport scenario refers to a typical historical operating state, while the target transport scenario refers to current or predicted transport conditions. This conversion relationship includes a transport capacity conversion factor and an inventory adjustment factor, which are used to adapt the historical optimization strategy to the new scenario. The conversion process considers regional factors such as transport infrastructure variance, such as port throughput and railway branch line coverage, and automatically adjusts the magnitude of parameter conversion. Furthermore, an analogical reasoning model is initiated for newly emerging transport scenarios, retrieving the most similar known model from the feature encoding library as the basis for conversion.

[0119] In step S107, the multi-dimensional resonance management knowledge base is implemented using a graph database structure, the transportation feature stores the timeliness prediction model parameters, the inventory feature records the spectrum analysis results, the resonance feature saves the coding pattern library, and the control strategy maintains the mapping relationship network and parameter conversion rules.

[0120] Preferably, the multidimensional resonance management knowledge base adopts a federated learning architecture, allowing sub-centers in different regions to share resonance pattern characteristics while protecting data privacy; the update mechanism of the multidimensional resonance management knowledge base can be set up with double verification, and all new patterns must pass historical data backtracking verification and real-time business environment testing before they are officially adopted; for major emergencies such as natural disasters, the system provides a quick entry channel for emergency modes, but it will be marked as pending verification and its scope of application will be limited.

[0121] This implementation eliminates discrepancies in multi-source data through standardization, uses time-frequency analysis to reveal implicit connections between transportation and inventory, and identifies typical resonance patterns based on feature encoding, ultimately building a dynamically adjustable decision-making knowledge base. As the transportation environment changes, the system can quickly match historical resonance patterns and generate adaptive inventory strategies through parameter conversion, achieving coordinated optimization of transportation resources and inventory levels.

[0122] See also Figure 2 In some embodiments, the raw data is preprocessed to obtain standardized intermodal transport data, which includes the second transport status parameter, the second inventory level parameter, and the second demand fluctuation parameter after time alignment, including:

[0123] S201, selecting transport node data that meets a preset integrity threshold from the first transport status parameter as a second transport status parameter, where the preset integrity threshold includes a continuity index of GPS positioning and a sampling rate index of a sensor;

[0124] S202: extracting, from the first inventory level parameter, inventory records that match a collection period of the second transportation status parameter as second inventory level parameters, wherein a time deviation of the collection period during matching does not exceed a preset inventory synchronization window;

[0125] S203: Selecting order data corresponding to a second transport status parameter collection period from the first demand fluctuation parameter as a second demand fluctuation parameter;

[0126] S204: Establish a transport-inventory-demand association relationship between the second transport status parameter, the second inventory level parameter, and the second demand fluctuation parameter, and generate a unified transport batch identifier;

[0127] S205. Extract multimodal transport features from the second transport status parameter to eliminate heterogeneity of data collected by different transport modes;

[0128] S206. Perform outlier correction processing on the second inventory level parameter to retain inventory change characteristics within the effective fluctuation range;

[0129] S207: Generate standardized intermodal transport data.

[0130] In step S201, a preset integrity threshold is used to ensure the reliability of transport node data. The continuity index of GPS positioning is determined by analyzing the spatiotemporal distribution of trajectory points, taking into account the differences in the characteristics of different modes of transportation, such as road and air transport. The sensor sampling rate index can be dynamically adjusted based on the transportation environment, automatically increasing the frequency of key parameter collection in high-risk transport sections. Only transport node data that meets this threshold is used as the second transport status parameter for subsequent processing.

[0131] In step S202, the preset inventory synchronization window is a key parameter for coordinating transportation and inventory timing. Its time deviation range can be dynamically adjusted based on the characteristics of the transportation vehicle. For special scenarios such as temperature-controlled transportation, the system will appropriately narrow the time window to ensure strict synchronization of environmental parameters. Preferably, the matching process uses a bidirectional time sliding algorithm, establishing an association when the inventory record timestamp falls within the time extension range of the transportation collection period.

[0132] In step S203, the selection of order data is achieved through the transport batch association mechanism. The system automatically matches the order delivery time with the transport vehicle loading time window. Furthermore, for special business scenarios such as pre-sale orders, an association relationship is established based on the estimated delivery time to ensure the integrity of demand fluctuation parameters.

[0133] In step S204, the transport-inventory-demand association is established using a composite-coded transport batch identifier, which includes the transport vehicle number, carrier code, and scheduled departure timestamp. When establishing the association, the transport route is automatically verified to ensure that it passes through the geofenced area where the inventory node is located, ensuring the accuracy of the data association.

[0134] In step S205, multimodal transport feature extraction includes two stages: data standardization and feature alignment. In particular, the record of transport mode conversion features is added to fully describe the connection efficiency between different transport stages. Furthermore, for cross-border transport scenarios, special process indicators such as customs clearance time are automatically identified and retained.

[0135] In step S206, outlier correction is performed using a dynamic threshold method, taking into account the characteristics of the inventory categories to implement differentiated processing. Preferably, a relatively loose correction threshold is used for fast-moving consumer goods, while a more stringent standard is used for precision instruments, eliminating outliers while retaining the true business fluctuation characteristics.

[0136] In step S207, preferably, the generation of standardized intermodal transport data includes a data quality marking step, adding an integrity level identifier for each parameter, and the subsequent analysis module can automatically adjust the data processing strategy according to these identifiers to ensure the reliability of the analysis results.

[0137] This embodiment achieves standardized conversion of multi-source heterogeneous data by establishing refined data processing rules and quality control mechanisms. The composite coding scheme for transport batch identification and the dynamic quality tagging system provide a high-quality data foundation for subsequent time-frequency correlation analysis. Each processing step adheres to unified specifications while adapting to the specific needs of different transport scenarios. The resulting standardized intermodal data fully preserves the business correlation characteristics between transport, inventory, and demand, making it particularly suitable for data analysis in complex multimodal transport environments and providing reliable data support for transport resource allocation and inventory optimization.

[0138] See also Figure 3 In some embodiments, a transportation time efficiency prediction model is constructed based on the second transportation status parameter to extract dynamic transportation characteristics, which include the delay probability distribution and switching response time of each transportation mode, including:

[0139] S301: Construct a transportation time efficiency prediction model, select time efficiency data of multiple transportation modes from the second transportation status parameter as training input of the transportation time efficiency prediction model, and generate a time efficiency feature vector for each transportation mode, wherein the time efficiency feature vector includes a historical delay rate and a real-time operation status;

[0140] S302: construct a highway transport delay prediction sub-model based on the timeliness feature vector and extract the highway transport delay probability distribution, which includes the average delay duration and delay frequency of different road sections;

[0141] S303. Construct a railway transportation delay prediction sub-model based on the timeliness feature vector and extract the railway transportation delay probability distribution, which includes the on-time rate fluctuation range and scheduling abnormality probability of each train.

[0142] S304. Constructing a water transport delay prediction sub-model based on the time efficiency feature vector and extracting a water transport delay probability distribution, where the water transport delay probability distribution includes the port operation delay time and the waterway congestion coefficient;

[0143] S305. Calculate the switching response time between road transport, rail transport, and water transport, including the transfer connection time from road to rail, the loading and unloading waiting time from rail to water, and the start delay time of emergency transport;

[0144] S306: Fusing the road transport delay probability distribution, the railway transport delay probability distribution, the water transport delay probability distribution, and the switching response time to generate dynamic transport features.

[0145] In step S301, the transportation timeliness prediction model is constructed using a multimodal feature fusion approach. The timeliness feature vector incorporates not only historical delay rates and real-time operational status, but also vehicle health indicators (such as vehicle / vessel / crew maintenance records) and environmental factors (such as weather and traffic control). Training input utilizes a sliding time window mechanism to ensure the model captures both short-term fluctuations and long-term trends in transportation timeliness. Furthermore, a federated learning framework can be used for model training, enabling transport companies to share model parameters while protecting data privacy, thereby improving the generalization of predictions.

[0146] In step S302, the highway transport delay prediction sub-model uses a spatiotemporal graph neural network, combining road network topology and real-time traffic flow data, to dynamically calculate delay risks for different road sections. To improve prediction accuracy, average delay duration and delay frequency are statistically analyzed by time period and road section combination. For example, minute-level predictions are used for peak periods, while hour-level predictions are used for normal periods. Preferably, the model automatically triggers a high-frequency dynamic correction mechanism for sections experiencing extreme weather or accidents to ensure real-time prediction results.

[0147] In step S303, the railway transportation delay prediction sub-model preferably utilizes an LSTM combined with an Attention mechanism, focusing on analyzing the impact of train timetable conflicts and dispatch instruction delays on delays. The fluctuation range of on-time performance is modeled by train type, and the probability of dispatch anomalies is dynamically adjusted based on historical dispatch logs and real-time signal system status. Preferably, a pre-configured rule library is loaded to correct prediction deviations for special scenarios such as temporary train additions or line construction, ensuring the reliability of the results.

[0148] In step S304, the waterway transport delay prediction sub-model preferably incorporates port operation simulation data to quantify the impact of berth allocation efficiency and loading and unloading equipment availability on delays. The waterway congestion coefficient can be calculated through cluster analysis of AIS vessel trajectories, distinguishing between normal congestion (e.g., narrow waterways) and sudden congestion (e.g., navigation closures due to accidents). Furthermore, the waterway transport delay prediction sub-model supports adaptive weighting for tidal cycles and seasonal cargo flow peaks (e.g., the peak season for bulk commodities), further enhancing the robustness of the prediction.

[0149] In step S305, the handover response time is calculated using a multi-scenario benchmarking approach. The road-to-rail transfer connection time is dynamically adjusted based on historical operational data from the transfer station. The loading and unloading waiting time from rail to waterway is optimized using port tide tables. The start-up delay time for emergency transport is calculated based on a historical emergency case database and supports manual correction based on experience. Expert review is triggered when the automated calculation results deviate from actual operational data.

[0150] In step S306, feature fusion preferably employs a graph attention network to construct a correlation map between transport mode, switching node, and delay risk, dynamically adjusting the weights of each feature. The fused dynamic transport feature output includes a delay risk heat map and switching strategy priority scores, providing visual support for multimodal transport decision-making.

[0151] This embodiment generates a time-sensitive feature vector based on standardized transport status parameters, encompassing historical delay rates and real-time operational status. It then constructs delay prediction sub-models for road, rail, and waterway transport, extracting the unique delay probability distribution characteristics of each mode of transport. It also calculates the switching response time between different modes of transport, ultimately generating comprehensive dynamic transport features through feature fusion. By constructing a time-sensitive prediction model that coordinates multiple modes of transport, this embodiment achieves accurate prediction of delay risks for road, rail, and waterway transport, as well as intelligent calculation of switching response times, providing real-time, reliable dynamic transport feature support for multimodal transport decision-making.

[0152] See also Figure 4 In some embodiments, performing spectral analysis on the second inventory level parameter to obtain inventory cycle fluctuation characteristics includes:

[0153] S401. Extracting a multi-time-scale inventory change sequence from the second inventory level parameter to obtain original inventory fluctuation data, where the multi-time-scale includes a daily fluctuation cycle, a weekly fluctuation cycle, and a seasonal fluctuation cycle;

[0154] S402. Perform wavelet transform on the original inventory fluctuation data to decompose it into inventory fluctuation components of different frequency bands. The inventory fluctuation components include a high-frequency random fluctuation component, a medium-frequency business cycle component, and a low-frequency trend component.

[0155] S403, performing extreme value statistical analysis on the high-frequency random fluctuation component, identifying sudden fluctuation events, and calculating the probability density function and autocorrelation characteristics of the sudden fluctuation events to generate sudden fluctuation characteristic parameters;

[0156] S404. Based on the sudden fluctuation characteristic parameters, construct an inventory abnormal fluctuation warning model, which includes a fluctuation threshold and warning response rules;

[0157] S405. Identify typical inventory fluctuation patterns from the medium-frequency business cycle components, including purchase cycle fluctuations, sales cycle fluctuations, and replenishment cycle fluctuations, and embed sudden fluctuation characteristic parameters as correction factors into the typical inventory fluctuation patterns;

[0158] S406, extracting trend characteristic parameters based on the low-frequency trend component, the trend characteristic parameters including the seasonal index and the long-term growth rate corrected by sudden fluctuations;

[0159] S407: Fusing inventory fluctuation components, typical inventory fluctuation patterns, sudden fluctuation characteristic parameters, and trend characteristic parameters to generate a multi-dimensional inventory cycle fluctuation characteristic matrix containing normal fluctuations and abnormal fluctuations;

[0160] S408. Standardize the multi-dimensional inventory cycle fluctuation characteristic matrix to obtain standardized inventory cycle fluctuation characteristics.

[0161] In step S401, multi-timescale inventory change series are acquired using a sliding window sampling technique. Preferably, daily fluctuation cycles utilize 15-minute granularity data to capture real-time inventory changes. Weekly fluctuation cycles are associated with enterprise schedules and logistics distribution cycles. Seasonal fluctuation cycles are annotated based on industry sales calendars and climate characteristics. Furthermore, time series alignment techniques are employed to eliminate time zone differences and inconsistent collection frequencies across different data sources.

[0162] In step S402, the wavelet transform uses Morlet basis functions, and adaptive frequency band segmentation is achieved by adjusting the scale parameter: high-frequency random fluctuation components capture transient fluctuations such as sales promotions, medium-frequency business cycle components reflect the regular business rhythm, and low-frequency trend components correspond to strategic inventory adjustments. Specifically, the decomposed components are energy-normalized to eliminate dimensionality effects.

[0163] In step S403, the extreme value statistical analysis preferably employs a POT model, using a dynamic threshold method to isolate abnormal fluctuations. Its probability density function is fitted to a generalized Pareto distribution, and the autocorrelation characteristics are obtained by calculating the partial autocorrelation coefficient of the lag order. Preferably, a composite event marking mechanism is triggered for continuous abnormal fluctuations lasting more than 24 hours.

[0164] In step S404, the fluctuation threshold is dynamically calculated using quantile regression. The warning response rules include a three-level response mechanism: Level 1 triggers automatic replenishment strategy adjustments, Level 2 initiates safety stock verification, and Level 3 utilizes the transportation timeliness prediction model to generate a contingency plan. The fluctuation threshold update cycle is synchronized with the inventory count cycle.

[0165] In step S405, a dynamic time warping algorithm is preferably used to identify typical inventory fluctuation patterns. Purchasing cycle fluctuations are modeled in association with switching response time. Sales cycle fluctuations are phase-calibrated using a second demand fluctuation parameter. Replenishment cycle fluctuations are corrected by incorporating a transport delay propagation coefficient. When the burst fluctuation characteristic parameter is used as a correction factor, its embedding weight is determined using a Bayesian probability weighting method.

[0166] In step S406, preferably, the trend characteristic parameter extraction process adopts the Holt-Winters three-parameter exponential smoothing method, the seasonal index is synchronized with the resonance coupling characteristic after the sudden fluctuation correction, and the long-term growth rate parameter is used to update the effective fluctuation range.

[0167] In step S407, feature fusion preferably employs tensor concatenation technology to perform a multi-dimensional integration of the frequency domain features of inventory fluctuation components, the time domain features of typical inventory fluctuation patterns, and the anomaly detection results of sudden fluctuation feature parameters. The resulting multi-dimensional inventory cycle fluctuation feature matrix has row dimensions corresponding to the preset inventory synchronization window, and column dimensions containing transportation features, inventory features, resonance features, and control strategies.

[0168] In step S408 , the RobustScaler method may be used for standardization processing to eliminate the differences in different feature dimensions.

[0169] This implementation decomposes raw inventory fluctuation data into high-frequency random fluctuations, medium-frequency business cycles, and low-frequency trend components. Extreme value statistical analysis is used to identify sudden fluctuation characteristics and establish a three-level early warning mechanism. A dynamic time warping algorithm is then used to extract typical business fluctuation patterns and embed sudden fluctuation characteristics as correction factors, ultimately generating a standardized multidimensional feature matrix. This implementation utilizes multi-scale spectral analysis technology to achieve precise decomposition and anomaly detection of inventory fluctuations. It intelligently integrates daily, weekly, and seasonal fluctuation patterns with sudden abnormal events to generate standardized inventory cycle fluctuation characteristics, providing multi-dimensional inventory feature support for dynamic supply chain decision-making.

[0170] In some embodiments, a time-frequency correlation analysis is performed on the second transportation state parameter and the second inventory level parameter to extract resonance coupling features, where the resonance coupling features include a transportation delay propagation coefficient and a demand amplification factor. A resonance feature encoding library containing typical resonance modes is established, where the typical resonance modes include at least one of a bullwhip effect, seasonal fluctuations, and sudden interruptions, including:

[0171] Time-align the transportation delay event sequence in the second transportation status parameter with the inventory fluctuation sequence in the second inventory level parameter to establish a delay-fluctuation correlation time axis;

[0172] Perform cross-spectral analysis on the delay-fluctuation correlation time axis, calculate the coherent spectral density of transportation delay and inventory fluctuation, and extract the transportation delay propagation coefficient, which represents the inventory fluctuation amplitude caused by unit transportation delay;

[0173] Perform Granger causality analysis on the second demand fluctuation parameter and the second inventory level parameter to determine the transmission delay and amplification factor from demand fluctuation to inventory fluctuation, and generate a demand amplification factor.

[0174] Identify typical resonance patterns from the latency-fluctuation correlation timeline, including bullwhip effects, seasonal fluctuations, and sudden disruptions;

[0175] Parameterized modeling of typical resonance modes is performed to generate resonance feature encoding vectors, which contain the combined characteristics of transport delay propagation coefficients and demand amplification factors.

[0176] The resonance feature encoding vector is associated with the transport batch identifier and stored to form a structured resonance feature;

[0177] and, constructing a resonance intensity prediction model based on the resonance feature encoding vector to generate a resonance risk level as a supplementary resonance feature;

[0178] The supplementary resonance features and structured resonance features are organized into a resonance feature coding library.

[0179] In this embodiment, the transportation delay event sequence refers to a time-stamped collection of delay events extracted from the second transportation status parameter, calculated using real-time location deviations and planned arrival times recorded by the transportation management system. The inventory fluctuation sequence, on the other hand, refers to the inventory level change trajectory with the same time base, extracted from the second inventory level parameter. The delay-fluctuation correlation timeline is precisely aligned using a dynamic time warping algorithm, specifically accounting for differences in data collection frequencies across different transportation modes to ensure accurate temporal correlation.

[0180] Cross-spectral analysis is used to quantify the frequency-domain correlation between shipping delays and inventory fluctuations. The coherent spectral density, calculated using a fast Fourier transform, reflects the strength of the energy coupling between the two at specific frequencies. The shipping delay propagation coefficient, a key resonance characteristic, is calculated through time-domain regression analysis and frequency-domain energy integration, ultimately representing the normalized inventory fluctuation amplitude caused by shipping delays per unit time. The demand amplification factor uses Granger causality testing to determine the lag order and amplification factor of the demand signal's impact on inventory. Its calculation utilizes a sliding window mechanism to adapt to varying transmission characteristics in different business scenarios.

[0181] Typically, a pattern matching algorithm is used to identify typical resonance patterns. Bullwhip effect patterns are identified by detecting the progressive amplification of demand signals across the supply chain hierarchy; seasonal fluctuation patterns are identified through spectrum peak detection combined with business calendar verification; and sudden outage patterns are identified using anomaly detection algorithms. Resonance feature encoding vectors are generated through parametric modeling. These encoding vectors not only incorporate fundamental features such as the transport delay propagation coefficient and demand amplification factor, but also capture their interactive effects under different delay combinations, forming a multidimensional feature representation.

[0182] The resonance intensity prediction model can be constructed using a gradient boosting decision tree, with the input being a resonance feature encoding vector and the output being a probability score reflecting the resonance risk level. Preferably, this model is trained using historical resonance event samples to predict the risk level of new transportation-inventory resonance patterns. The resulting resonance feature encoding library is stored in a graph database structure, with the transport batch identifier as the primary key, enabling efficient associative queries with transport and inventory features.

[0183] This embodiment combines time-frequency correlation analysis with pattern recognition technology to achieve systematic extraction and structured expression of transportation-inventory resonance features, providing key resonance risk warning capabilities for subsequent multimodal transport decisions.

[0184] In some embodiments, a mapping relationship network between transportation mode switching rules and inventory adjustment strategies is constructed based on dynamic transportation characteristics, inventory cycle fluctuation characteristics, and resonance coupling characteristics, including:

[0185] Extracting timeliness risk indicators for each mode of transport from dynamic transport characteristics. These indicators include delay probability distribution curves and switching response time thresholds.

[0186] Analyze the key fluctuation parameters in inventory cycle fluctuation characteristics, including safety stock threshold and replenishment cycle sensitivity;

[0187] Determining resonance control parameters based on resonance coupling characteristics, the resonance control parameters including a transport delay propagation coefficient critical value and a demand amplification factor tolerance;

[0188] Conduct multi-dimensional correlation analysis on timeliness risk indicators, key fluctuation parameters and resonance control parameters to establish a three-dimensional decision space of transportation-inventory-resonance;

[0189] A transport mode switching rule set is constructed in the three-dimensional decision space of transport-inventory-resonance. The transport mode switching rule set includes the transport mode priority sorting and switching trigger conditions under different risk levels.

[0190] Based on the resonance control parameters, an inventory adjustment strategy matrix is ​​generated in the transportation-inventory-resonance three-dimensional decision space. The inventory adjustment strategy matrix includes a dynamic safety stock calculation formula and an emergency replenishment trigger mechanism.

[0191] Co-optimize the transportation mode switching rule set and the inventory adjustment strategy matrix to generate a strategy feature coding library containing typical scenario response solutions;

[0192] Also, conduct visual modeling of the strategy feature coding library and construct a strategy decision tree as a supplementary control basis;

[0193] The strategy decision tree, transportation mode switching rule set and inventory adjustment strategy matrix are integrated to form a mapping relationship network of transportation-inventory linkage.

[0194] In this embodiment, the timeliness risk index is a parameter extracted from dynamic transportation characteristics to quantify transportation reliability. The delay probability distribution curve is generated by statistically analyzing historical transportation delay events, reflecting the delay risk of different transportation modes during specific time periods. The switching response time threshold is determined based on historical case studies of transportation mode switching, reflecting the time sensitivity of intermodal transport connections. The safety stock threshold is determined through extreme value analysis of inventory cycle fluctuations, and the replenishment cycle sensitivity reflects the responsiveness of inventory levels to changes in replenishment time.

[0195] Resonance control parameters serve as key indicators for coordinating transportation and inventory. The critical value of the transportation delay propagation coefficient, determined through time-frequency correlation analysis, is used to determine whether transportation delays trigger significant inventory fluctuations. The demand amplification factor tolerance is set based on supply chain stability requirements to control the upper limit of demand signal distortion during transmission. The three-dimensional decision space of transportation, inventory, and resonance is constructed using multidimensional feature vectors, whose dimensional weights are dynamically adjusted based on business scenarios, enabling strategy adaptation in different transportation environments.

[0196] Optimally, the transport mode switching rule set is generated using a fuzzy inference mechanism. The transport mode priority ranking considers the balance between delay probability and switching time, and the switching trigger conditions are linked to the resonance risk level and inventory status. The dynamic safety stock calculation formula in the inventory adjustment strategy matrix incorporates the transport delay propagation coefficient as a correction factor, and the emergency replenishment trigger mechanism is linked in real time to the demand amplification factor. A policy feature encoding library is constructed using case-based reasoning methods to store optimized policy combinations for typical scenarios.

[0197] The visual modeling process utilizes a decision tree algorithm to transform complex, multidimensional relationships into interpretable rule nodes, each of which contains transport mode selection criteria and corresponding inventory adjustments. The resulting mapping network integrates policy elements through a graph structure, supporting real-time policy retrieval and dynamic optimization as transport status changes. By constructing a three-dimensional decision space and a policy encoding library, this implementation achieves intelligent coordination between transport mode switching and inventory adjustments, effectively enhancing the multimodal transport system's ability to cope with complex fluctuations.

[0198] In some embodiments, establishing a resonance control parameter conversion relationship between a reference transport scenario and a target transport scenario based on the dynamic transport characteristics and the resonance coupling characteristics includes:

[0199] Extracting baseline transport time efficiency characteristic parameters of the baseline transport scenario and target transport time efficiency characteristic parameters of the target transport scenario from the dynamic transport characteristics, wherein the baseline transport time efficiency characteristic parameters include baseline delay probability distribution and baseline switching response time, and the target transport time efficiency characteristic parameters include target delay probability distribution and target switching response time;

[0200] Extracting baseline resonance characteristic parameters of the baseline transport scenario and target resonance characteristic parameters of the target transport scenario from the resonance coupling characteristics, the baseline resonance characteristic parameters including a baseline delay propagation coefficient and a baseline demand amplification factor, and the target resonance characteristic parameters including a target delay propagation coefficient and a target demand amplification factor;

[0201] Compare and analyze the baseline transport time characteristic parameters with the target transport time characteristic parameters, and calculate the equivalent conversion coefficient of the transport delay probability and the scenario adjustment factor of the switching response time;

[0202] Matching and calibrating the reference resonance characteristic parameters with the target resonance characteristic parameters to determine the scenario scaling ratio of the delay propagation coefficient and the scenario correction parameters of the demand amplification factor;

[0203] A resonance control parameter conversion model is constructed based on the equivalent conversion coefficient, scene adjustment factor, scene scaling ratio, and scene correction parameter. The resonance control parameter conversion model includes a scene conversion algorithm for transportation timeliness characteristics and a scene adaptation rule for resonance characteristics.

[0204] The characteristic parameters of the benchmark transport scenario are input into the resonance control parameter conversion model for parameter conversion, and the optimized control parameters of the target transport scenario are output;

[0205] Verify the scenario applicability of the optimized control parameters and generate conversion relationship correction coefficients including scenario difference compensation terms;

[0206] A resonance control parameter conversion relationship is generated according to the resonance control parameter conversion model and the conversion relationship correction coefficient.

[0207] In this embodiment, the baseline transport scenario is a typical transport environment with sufficient historical data. Its baseline delay probability distribution is constructed by statistically analyzing historical delay events, reflecting the inherent transport risk characteristics of this scenario. The target transport scenario is a new transport environment to be evaluated, and its target delay probability distribution is generated using real-time monitoring data or scenario simulation. Both the baseline and target switching response times are calculated using historical operational records of transport mode conversions, representing the efficiency of intermodal transport connections under different scenarios.

[0208] The delay propagation coefficient within the baseline resonance characteristic parameters is derived through historical transportation-inventory correlation analysis, reflecting the transmission intensity of transportation delays to inventory fluctuations under the baseline scenario. The target delay propagation coefficient is preliminarily estimated through scenario similarity matching. The demand amplification factor is determined through supply chain demand transmission analysis. The difference between the baseline and target values ​​reflects the degree of demand signal distortion under different scenarios.

[0209] Optimally, the equivalent conversion coefficient for the transport delay probability is calculated using a probability distribution matching algorithm to quantitatively compare risk levels across different scenarios. The scenario adjustment factor is determined through differential analysis of transport operational characteristics. The scenario scaling ratio for the delay propagation coefficient is derived based on the coupling relationship between transport timeliness and resonance characteristics, while the scenario correction parameter for the demand amplification factor takes into account the adaptive adjustment of the supply chain structure.

[0210] The resonance control parameter conversion model is constructed using a multi-layer perceptron. Its input layer receives baseline scenario features, the hidden layer performs feature conversion operations, and the output layer generates target scenario parameters. Scenario applicability is verified through backpropagation error analysis, and a conversion relationship correction coefficient is used to compensate for scenario-specific features not captured by the model. The final conversion relationship is expressed as a parameter mapping matrix, enabling rapid migration of control strategies between different transportation scenarios and effectively addressing the difficulty in quickly and accurately setting resonance control parameters in new scenarios.

[0211] This embodiment achieves rapid migration and precise adaptation of control strategies under different transportation environments by establishing a resonant control parameter conversion relationship between the baseline and target transportation scenarios, effectively improving the adaptability and risk control level of the multimodal transport system to new scenarios.

[0212] In some embodiments, dynamic transportation characteristics, inventory cycle fluctuation characteristics, resonance characteristic coding library, mapping relationship network and resonance control parameter conversion relationship are dynamically integrated to construct a multi-dimensional resonance management knowledge base containing transportation characteristics, inventory characteristics, resonance characteristics and control strategies, including:

[0213] Extracting multimodal transport characteristic data from dynamic transport characteristics, the multimodal transport characteristic data including road transport characteristic parameters, railway transport characteristic parameters and water transport characteristic parameters;

[0214] Extracting multi-time scale inventory characteristic data from inventory cycle fluctuation characteristics, the multi-time scale inventory characteristic data includes daily fluctuation characteristic parameters, weekly fluctuation characteristic parameters and seasonal fluctuation characteristic parameters;

[0215] Extracting typical resonance mode feature data from a resonance feature coding library, wherein the typical resonance mode feature data includes bullwhip effect feature parameters, seasonal resonance feature parameters and sudden interruption conduction feature parameters;

[0216] Extracting transportation-inventory linkage strategy data from the mapping relationship network, the transportation-inventory linkage strategy data including transportation mode switching rule parameters and inventory adjustment strategy parameters;

[0217] Extracting multi-scenario conversion parameter data from the resonance control parameter conversion relationship, the multi-scenario conversion parameter data including reference scenario feature parameters and target scenario optimization parameters;

[0218] The multimodal transport feature data, multi-time scale inventory feature data, typical resonance mode feature data, transport-inventory linkage strategy data and multi-scenario conversion parameter data are fused to generate a multi-dimensional resonance feature matrix.

[0219] The knowledge base storage structure is constructed based on the multi-dimensional resonance feature matrix. The knowledge base storage structure includes:

[0220] Transport characteristic storage module, used to store and manage multimodal transport characteristic data;

[0221] Inventory feature storage module, used to store and manage multi-time scale inventory feature data;

[0222] A resonance characteristic storage module, used for storing and managing characteristic data of typical resonance modes;

[0223] Policy rule storage module, used to store and manage transportation-inventory linkage policy data;

[0224] A scene conversion storage module is used to store and manage multi-scene conversion parameter data;

[0225] Optimize the index of the knowledge base storage structure, establish a fast retrieval mechanism for feature parameters and a dynamic call interface for policy rules;

[0226] The multidimensional resonance feature matrix is ​​integrated with the knowledge base storage structure to form a multidimensional resonance management knowledge base including transportation feature dimension, inventory feature dimension, resonance feature dimension and control strategy dimension.

[0227] In this example, road transport characteristic parameters include real-time changes in road conditions and vehicle dispatch efficiency indicators. Railway transport characteristic parameters reflect train punctuality and track capacity. Waterway transport characteristic parameters characterize port loading and unloading efficiency and waterway navigation conditions. Daily fluctuation characteristic parameters capture daily replenishment and consumption patterns, weekly fluctuation characteristic parameters reflect the different patterns between weekdays and holidays, and seasonal fluctuation characteristic parameters record cyclical demand trends.

[0228] The bullwhip effect characteristic parameter quantifies the progressive amplification of demand signals within the supply chain, the seasonal resonance characteristic parameter describes the cumulative effect of cyclical fluctuations, and the sudden disruption transmission characteristic parameter characterizes the propagation path of sudden events within the transportation-inventory system. Transportation-inventory linkage strategy data is generated through the aforementioned mapping network and includes transportation priority adjustment plans and inventory buffering strategies tailored to different resonance levels. Multi-scenario conversion parameter data is derived from the resonance control parameter conversion relationship, enabling the intelligent transfer of baseline scenario experience to target scenarios.

[0229] The knowledge base storage structure utilizes a distributed database architecture. The transportation feature storage module enables real-time updates of multi-source transportation data. The inventory feature storage module supports adaptive recognition of fluctuation patterns. The resonance feature storage module provides matching and retrieval of typical patterns. The policy rule storage module implements version management of control policies. The scenario conversion storage module maintains cross-scenario parameter mapping relationships. The index optimization process utilizes inverted indexing technology. The feature parameter fast retrieval mechanism achieves millisecond-level response through hash mapping. The policy rule dynamic call interface supports real-time policy push.

[0230] This embodiment, through the construction of a multidimensional resonance management knowledge base, achieves digital representation and intelligent management of all elements of the transportation-inventory system, providing a unified knowledge support platform for resonance risk prevention and control in complex supply chain environments. The multidimensional resonance management knowledge base's dimensional storage and intelligent retrieval mechanism enables the system to quickly identify risk patterns and invoke optimal control strategies, significantly improving supply chain resilience and responsiveness.

[0231] In a second aspect, this embodiment further provides a computer-readable storage medium storing computer program instructions, which implement the method described in the first aspect when executed by a processor.

[0232] The computer program involved in this embodiment can be stored in a computer device readable storage medium, which includes but is not limited to a disk, a magnetic tape, a magnetic card, a floppy disk, a flash memory, an optical disc, an optical card, a read-only memory (ROM), a random access memory (RAM), an erasable programmable ROM (EPROM) and an electrically erasable programmable ROM (EEPROM), etc., and also includes other biological, physical or chemical structures that can achieve similar or equivalent functions to the storage media listed above, such as DNA, RNA, proteins and other units with information storage capabilities. In a specific embodiment, the storage medium involved can be one of the above-mentioned media types or a combination of the above-mentioned media types. In different embodiments, the computer program involved in the embodiment can be stored in a single medium in a centralized manner or in a distributed manner in multiple media. The memory containing the computer device readable storage medium can be a non-volatile memory or a random access memory. These computer device readable storage media can be built into the device or can be connected to the device involved in the embodiment as an external device or part of an external device. In some embodiments, a memory having a computer-readable storage medium is deployed locally; in other embodiments, a solution of deploying the memory away from the processor may also be adopted, such as a network-attached memory accessed via an RF circuit or an external port and a communication network, wherein the communication network may be the Internet, one or more intranets, a local area network (LAN), a wide area network (WLAN), a storage area network (SAN), etc., or a suitable combination thereof, as long as the computer device can access the memory. In addition, the computer program involved in the embodiment can be stored in plaintext / ciphertext form, or can be designed as training data, which can be integrated and reorganized through model training and implicitly stored in the parameter state of a deep neural network or other machine learning model.

[0233] See also Figure 5 In a third aspect, this embodiment further provides an electronic device 1, comprising a memory 11 and a processor 12, wherein the memory 11 is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor 12 to implement the method described in the first aspect.

[0234] The processor described in this embodiment can be implemented by hardware, firmware, software or a combination thereof, and can use a circuit, a single or multiple application-specific integrated circuits (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a central processing unit (CPU), a controller, a microcontroller, or at least one of a microprocessor. It also includes other physical, biological or chemical structures that can achieve similar or equivalent functions to the processors listed above, such as biological neurons, quantum computing units, DNA computing units, etc., so that the processor can execute some or all of the steps in the computer program or method involved in each embodiment of the present application, or any combination of the steps mentioned therein.

[0235] By adopting the above technical solutions, the present invention is different from the existing technology and has the following beneficial effects: by standardizing the processing of multimodal transport raw data to construct a transport time prediction model and an inventory spectrum analysis model, extracting dynamic transport characteristics and inventory cycle fluctuation characteristics, and revealing the resonance coupling characteristics of transport and inventory based on time-frequency correlation analysis; by constructing a mapping relationship network between transport mode switching rules and inventory adjustment strategies, as well as a resonance control parameter conversion relationship between the baseline transport scenario and the target transport scenario, a multidimensional resonance management knowledge base containing transport characteristics, inventory characteristics, resonance characteristics and control strategies is finally formed. The present invention realizes the coordinated prediction of transport delay risks and inventory fluctuations, establishes an intelligent linkage mechanism between multimodal transport schemes and inventory strategies, effectively suppresses the bullwhip effect amplification and seasonal fluctuation imbalance, significantly improves the adaptability of the supply chain system to changes in the transport environment and demand fluctuations, and provides a systematic solution for inventory resonance management in a complex multimodal transport environment.

[0236] Finally, it should be noted that although the above embodiments have been described in the specification and drawings of this application, this does not limit the scope of patent protection of this application. All technical solutions generated by replacing or modifying equivalent structures or equivalent processes based on the essential concepts of this application using the contents recorded in the specification and drawings of this application, as well as directly or indirectly implementing the technical solutions of the above embodiments in other related technical fields, are included in the scope of patent protection of this application.

Claims

1. A multimodal inventory resonance management method based on dynamic rules, characterized in that: include: Collecting raw data according to the transportation mode classification, the raw data including a first transportation status parameter, a first inventory level parameter, and a first demand fluctuation parameter; Preprocessing the original data to obtain standardized intermodal transport data, the standardized intermodal transport data including a second transport state parameter, a second inventory level parameter, and a second demand fluctuation parameter after time series alignment; A transportation timeliness prediction model is constructed based on the second transportation status parameter to extract dynamic transportation characteristics, including the delay probability distribution and switching response time of each transportation mode. In addition, a spectral analysis is performed on the second inventory level parameter to obtain inventory cycle fluctuation characteristics. Performing a time-frequency correlation analysis on the second transportation state parameter and the second inventory level parameter to extract resonance coupling features, including a transportation delay propagation coefficient and a demand amplification factor. Establishing a resonance feature encoding library containing typical resonance modes, wherein the typical resonance modes include at least one of a bullwhip effect, seasonal fluctuations, and sudden interruptions. The transportation delay propagation coefficient is used to quantify the transmission intensity of transportation delays on inventory fluctuations, and the demand amplification factor is used to calculate the degree of distortion of demand signals in the supply chain. Based on dynamic transportation characteristics, inventory cycle fluctuation characteristics, and resonance coupling characteristics, a mapping relationship network between transportation mode switching rules and inventory adjustment strategies is constructed. This mapping relationship network is constructed using a decision tree algorithm, which associates the delay probability in dynamic transportation characteristics with the inventory safety threshold. The transportation mode priority rules are generated based on the switching response time, and the replenishment timing strategy is formulated based on the inventory cycle fluctuation characteristics. Based on the dynamic transport characteristics and resonance coupling characteristics, a conversion relationship between the resonance control parameters of the baseline transport scenario and the target transport scenario is established. The resonance control parameters include the critical value of the transport delay propagation coefficient and the tolerance of the demand amplification factor. Dynamic transportation characteristics, inventory cycle fluctuation characteristics, resonance feature coding library, mapping relationship network and resonance control parameter conversion relationship are dynamically integrated to construct a multidimensional resonance management knowledge base including transportation characteristics, inventory characteristics, resonance characteristics and control strategies.

2. The multimodal transport inventory resonance management method based on dynamic rules according to claim 1 is characterized in that: The raw data is preprocessed to obtain standardized intermodal transport data. The standardized intermodal transport data includes the second transport status parameter, the second inventory level parameter, and the second demand fluctuation parameter after time series alignment, including: Selecting, from the first transport status parameters, transport node data that meets a preset integrity threshold as a second transport status parameter, wherein the preset integrity threshold includes a continuity index of GPS positioning and a sampling rate index of a sensor; Extracting, from the first inventory level parameter, inventory records that match a collection period of the second transport status parameter as second inventory level parameters, wherein a time deviation of the collection period during matching does not exceed a preset inventory synchronization window; Selecting order data corresponding to the second transport status parameter collection period from the first demand fluctuation parameter as a second demand fluctuation parameter; establishing a transport-inventory-demand association relationship between the second transport state parameter, the second inventory level parameter, and the second demand fluctuation parameter, and generating a unified transport batch identifier; Extracting multimodal transport features from the second transport status parameter to eliminate heterogeneity of data collected from different transport modes; performing outlier correction processing on the second inventory level parameter to retain inventory change characteristics within a valid fluctuation range; Generate standardized intermodal data.

3. The multimodal inventory resonance management method based on dynamic rules according to claim 1 is characterized in that: A transportation time efficiency prediction model is constructed based on the second transportation state parameter to extract dynamic transportation characteristics. The dynamic transportation characteristics include the delay probability distribution and switching response time of each transportation mode, including: Constructing a transportation time efficiency prediction model, selecting time efficiency data of multiple transportation modes from the second transportation status parameter as training input of the transportation time efficiency prediction model, and generating a time efficiency feature vector for each transportation mode, wherein the time efficiency feature vector includes a historical delay rate and a real-time operation status; Building a highway transport delay prediction sub-model based on the time efficiency feature vector and extracting a highway transport delay probability distribution, wherein the highway transport delay probability distribution includes the average delay duration and delay occurrence frequency of different road sections; Building a railway transport delay prediction sub-model based on the timeliness feature vector and extracting a railway transport delay probability distribution, wherein the railway transport delay probability distribution includes the fluctuation range of the on-time rate of each train and the probability of scheduling anomalies; Constructing a water transport delay prediction sub-model based on the time efficiency feature vector and extracting a water transport delay probability distribution, wherein the water transport delay probability distribution includes a port operation delay time and a waterway congestion coefficient; Calculate the switching response time between road transport, rail transport and water transport, including the transfer connection time from road to rail, the loading and unloading waiting time from rail to water transport, and the start delay time of emergency transport; The road transport delay probability distribution, the railway transport delay probability distribution, the water transport delay probability distribution and the switching response time are integrated to generate dynamic transport features.

4. The multimodal inventory resonance management method based on dynamic rules according to claim 1 is characterized in that: Perform spectral analysis on the second inventory level parameter to obtain inventory cycle fluctuation characteristics, including: Extracting a multi-time-scale inventory change sequence from the second inventory level parameter to obtain original inventory fluctuation data, wherein the multi-time-scale includes a daily fluctuation cycle, a weekly fluctuation cycle, and a seasonal fluctuation cycle; Performing wavelet transform processing on the original inventory fluctuation data to decompose and obtain inventory fluctuation components of different frequency bands, wherein the inventory fluctuation components include a high-frequency random fluctuation component, a medium-frequency business cycle component, and a low-frequency trend component; Perform extreme value statistical analysis on the high-frequency random fluctuation component to identify sudden fluctuation events, calculate the probability density function and autocorrelation characteristics of the sudden fluctuation events, and generate sudden fluctuation characteristic parameters; Based on the sudden fluctuation characteristic parameters, an inventory abnormal fluctuation warning model is constructed, wherein the abnormal fluctuation warning model includes a fluctuation threshold and a warning response rule; Identifying typical inventory fluctuation patterns from the medium-frequency business cycle components, including purchase cycle fluctuations, sales cycle fluctuations, and replenishment cycle fluctuations, and embedding the sudden fluctuation characteristic parameters as correction factors into the typical inventory fluctuation patterns; Extracting trend characteristic parameters based on the low-frequency trend component, wherein the trend characteristic parameters include a seasonal index and a long-term growth rate corrected for sudden fluctuations; The inventory fluctuation components, typical inventory fluctuation patterns, sudden fluctuation characteristic parameters and trend characteristic parameters are subjected to feature fusion to generate a multi-dimensional inventory cycle fluctuation characteristic matrix including normal fluctuations and abnormal fluctuations; The multi-dimensional inventory cycle fluctuation characteristic matrix is ​​standardized to obtain standardized inventory cycle fluctuation characteristics.

5. The multimodal inventory resonance management method based on dynamic rules according to claim 1 is characterized in that: A time-frequency correlation analysis is performed on the second transportation state parameter and the second inventory level parameter to extract resonance coupling features. The resonance coupling features include the transportation delay propagation coefficient and the demand amplification factor. A resonance feature coding library containing typical resonance modes is established. The typical resonance modes include at least one of the bullwhip effect, seasonal fluctuations, and sudden interruptions, including: Time-aligning a transportation delay event sequence in the second transportation status parameter with an inventory fluctuation sequence in the second inventory level parameter to establish a delay-fluctuation correlation time axis; Performing a cross-spectral analysis on the delay-fluctuation correlation time axis, calculating the coherent spectral density of transportation delay and inventory fluctuation, and extracting a transportation delay propagation coefficient, wherein the transportation delay propagation coefficient represents the amplitude of inventory fluctuation caused by a unit of transportation delay; Performing Granger causality analysis on the second demand fluctuation parameter and the second inventory level parameter to determine the transmission delay and amplification factor from demand fluctuation to inventory fluctuation, and generating a demand amplification factor; Identify typical resonance patterns from the delay-fluctuation correlation timeline, including bullwhip effect, seasonal fluctuations, and sudden interruptions; Performing parameterized modeling on the typical resonance mode to generate a resonance feature coding vector, wherein the coding vector includes a delay combination feature of a transport delay propagation coefficient and a demand amplification factor; storing the resonance feature encoding vector in association with the transport batch identifier to form a structured resonance feature; and, constructing a resonance intensity prediction model based on the resonance feature encoding vector to generate a resonance risk level as a supplementary resonance feature; The supplementary resonance features and the structured resonance features are organized into the resonance feature coding library.

6. The multimodal inventory resonance management method based on dynamic rules according to claim 1 is characterized in that: Based on dynamic transportation characteristics, inventory cycle fluctuation characteristics, and resonance coupling characteristics, a mapping relationship network between transportation mode switching rules and inventory adjustment strategies is constructed, including: Extracting time-sensitive risk indicators for each mode of transportation from the dynamic transportation characteristics, the time-sensitive risk indicators including a delay probability distribution curve and a switching response time threshold; Analyzing key fluctuation parameters in the inventory cycle fluctuation characteristics, wherein the key fluctuation parameters include a safety stock threshold and replenishment cycle sensitivity; determining a resonance control parameter based on the resonance coupling characteristic; Conduct multi-dimensional correlation analysis on the timeliness risk index, key fluctuation parameters and resonance control parameters to establish a three-dimensional decision space of transportation-inventory-resonance; Constructing a transportation mode switching rule set in the transportation-inventory-resonance three-dimensional decision space, wherein the transportation mode switching rule set includes transportation mode priority sorting and switching triggering conditions under different risk levels; generating an inventory adjustment strategy matrix in the transportation-inventory-resonance three-dimensional decision space based on the resonance control parameters, wherein the inventory adjustment strategy matrix includes a dynamic safety stock calculation formula and an emergency replenishment trigger mechanism; Co-optimizing the transportation mode switching rule set and the inventory adjustment strategy matrix to generate a strategy feature coding library containing typical scenario response solutions; and, visually modeling the strategy feature coding library and constructing a strategy decision tree as a supplementary control basis; The strategy decision tree, the transportation mode switching rule set and the inventory adjustment strategy matrix are integrated to form the mapping relationship network of transportation-inventory linkage.

7. The multimodal transport inventory resonance management method based on dynamic rules according to claim 1 is characterized in that: Based on the dynamic transport characteristics and resonance coupling characteristics, the resonance control parameter conversion relationship between the baseline transport scenario and the target transport scenario is established, including: Extracting, from the dynamic transport characteristics, baseline transport time efficiency characteristic parameters of a baseline transport scenario and target transport time efficiency characteristic parameters of a target transport scenario, wherein the baseline transport time efficiency characteristic parameters include a baseline delay probability distribution and a baseline switching response time, and the target transport time efficiency characteristic parameters include a target delay probability distribution and a target switching response time; Extracting a baseline resonance characteristic parameter of a baseline transport scenario and a target resonance characteristic parameter of a target transport scenario from the resonance coupling characteristic, wherein the baseline resonance characteristic parameter includes a baseline delay propagation coefficient and a baseline demand amplification factor, and the target resonance characteristic parameter includes a target delay propagation coefficient and a target demand amplification factor; Compare and analyze the baseline transport time characteristic parameters with the target transport time characteristic parameters, and calculate the equivalent conversion coefficient of the transport delay probability and the scenario adjustment factor of the switching response time; Matching and calibrating the reference resonance characteristic parameters with the target resonance characteristic parameters to determine the scenario scaling ratio of the delay propagation coefficient and the scenario correction parameters of the demand amplification factor; A resonance control parameter conversion model is constructed based on the equivalent conversion coefficient, the scene adjustment factor, the scene scaling ratio, and the scene correction parameter, wherein the resonance control parameter conversion model includes a scene conversion algorithm for transport timeliness characteristics and a scene adaptation rule for resonance characteristics; Inputting the characteristic parameters of the reference transport scenario into the resonance control parameter conversion model for parameter conversion, and outputting the optimized control parameters of the target transport scenario; Verifying the scenario applicability of the optimized control parameters and generating a conversion relationship correction coefficient including a scenario difference compensation term; A resonance control parameter conversion relationship is generated according to the resonance control parameter conversion model and the conversion relationship correction coefficient.

8. The multimodal transport inventory resonance management method based on dynamic rules according to claim 1 is characterized in that: Dynamic transportation characteristics, inventory cycle fluctuation characteristics, resonance feature coding library, mapping relationship network and resonance control parameter conversion relationship are dynamically integrated to build a multi-dimensional resonance management knowledge base containing transportation characteristics, inventory characteristics, resonance characteristics and control strategies, including: Extracting multimodal transport characteristic data from the dynamic transport characteristics, wherein the multimodal transport characteristic data includes road transport characteristic parameters, railway transport characteristic parameters, and water transport characteristic parameters; Extracting multi-time-scale inventory feature data from the inventory cycle fluctuation characteristics, wherein the multi-time-scale inventory feature data includes daily fluctuation feature parameters, weekly fluctuation feature parameters, and seasonal fluctuation feature parameters; Extracting typical resonance mode characteristic data from the resonance characteristic coding library, wherein the typical resonance mode characteristic data includes a bullwhip effect characteristic parameter, a seasonal resonance characteristic parameter, and a sudden interruption conduction characteristic parameter; Extracting transportation-inventory linkage strategy data from the mapping relationship network, wherein the transportation-inventory linkage strategy data includes transportation mode switching rule parameters and inventory adjustment strategy parameters; Extracting multi-scenario conversion parameter data from the resonance control parameter conversion relationship, wherein the multi-scenario conversion parameter data includes reference scene feature parameters and target scene optimization parameters; Performing feature fusion on the multimodal transport feature data, multi-time scale inventory feature data, typical resonance mode feature data, transport-inventory linkage strategy data, and multi-scenario conversion parameter data to generate a multi-dimensional resonance feature matrix; A knowledge base storage structure is constructed based on the multidimensional resonance feature matrix, and the knowledge base storage structure includes: A transport characteristic storage module, used for storing and managing the multimodal transport characteristic data; an inventory feature storage module, configured to store and manage the multi-time-scale inventory feature data; A resonance characteristic storage module, used for storing and managing the typical resonance mode characteristic data; A strategy rule storage module, used to store and manage the transportation-inventory linkage strategy data; A scene conversion storage module, used for storing and managing the multi-scene conversion parameter data; Optimize the index of the knowledge base storage structure, establish a fast retrieval mechanism for feature parameters and a dynamic call interface for policy rules; The multidimensional resonance feature matrix is ​​integrated with the knowledge base storage structure to form a multidimensional resonance management knowledge base including a transportation feature dimension, an inventory feature dimension, a resonance feature dimension and a control strategy dimension.

9. A computer-readable storage medium storing computer program instructions, characterized in that: The computer program instructions implement the method according to any one of claims 1 to 8 when executed by a processor.

10. An electronic device comprising a memory and a processor, characterized in that: The memory is configured to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method according to any one of claims 1 to 8.

Images