A method for optimizing multimodal transport service networks

Abstract

This invention discloses a method for optimizing a multimodal transport service network, relating to the field of freight transport technology. The method includes the following steps: constructing a unified nanosecond-level global synchronization clock reference; synchronizing the time of hub nodes under various transport modes; and collecting first and second frequency data in real time based on this clock reference. The first frequency data represents the information flow refresh frequency, and the second frequency data represents the physical displacement frequency of logistics. This invention achieves high-precision synchronization and deep calibration of information and logistics flows under multiple transport modes through a unified nanosecond-level synchronization clock and blockchain timestamps, improving the time consistency and data reliability of multimodal transport. By combining a speed benchmark matrix and wavelet phase difference analysis, synchronization imbalance is accurately quantified, and a real-time imbalance index drives the dynamic completion of the information flow using a digital twin, ensuring the continuity and accuracy of scheduling decisions and reducing operational errors and transport risks.

Description

Technical Field

[0001] This invention relates to the field of freight transportation technology, and more specifically to a method for optimizing a multimodal transport service network. Background Technology

[0002] Multimodal transport network optimization refers to the systematic design and dynamic collaborative management in the freight and transportation sector to achieve efficient connection and resource allocation between different modes of transport (such as rail, road, water, and air), thereby improving end-to-end transport efficiency, reducing logistics costs, shortening transport cycles, and enhancing the robustness and responsiveness of the transport network. This optimization process includes not only the joint planning of transport routes, hub nodes, vehicle scheduling, and vehicle switching, but also the integrated scheduling and intelligent decision-making of factors such as the timeliness of each transport mode, capacity matching, loading and unloading efficiency, and information flow synchronization. By applying operations research, intelligent algorithms, and big data analysis, optimizing the multimodal transport network can achieve optimal allocation of transport resources, dynamic relief of bottleneck nodes, and multi-scheme emergency scheduling, thereby constructing a low-carbon, efficient, flexible, and cost-sensitive integrated logistics service system and enhancing the overall stability and competitiveness of the supply chain.

[0003] The existing technology has the following shortcomings:

[0004] In multimodal transport hubs, existing technologies commonly suffer from a mismatch between the information flow and the physical flow of goods at hub nodes. This is particularly true in transshipment involving multiple modes of transport, where the data interaction and monitoring systems at hub nodes fail to provide comprehensive coverage of the entire logistics process, leading to synchronization discrepancies between information and cargo flows. This results in operational errors such as misdelivery, omissions, and duplicate loading and unloading, reducing the visibility and traceability of the entire multimodal transport process. This problem is especially pronounced in scenarios involving high-value or sensitive goods; missing or delayed information flow can lead to incorrect cargo destinations, abnormal delays, and even transport safety incidents, posing a serious threat to the stability and security of the logistics chain.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide a method for optimizing multimodal transport service networks. By unifying a nanosecond-level global synchronous clock reference and blockchain timestamps, it achieves high-precision time synchronization and deep calibration of information flow and logistics flow under various transportation modes, improving the time consistency and data reliability of the entire multimodal transport process. Through speed benchmarking matrix and multi-scale wavelet phase difference analysis, it accurately characterizes synchronization deviations and quantifies synchronization imbalances. Combining real-time imbalance index and incremental residual matching with digital twins, it dynamically fills information flow gaps, ensuring the continuity and accuracy of scheduling decisions. Overall, it improves the visualization and intelligent scheduling capabilities of multimodal transport, enhances the stability and security of the logistics network, and reduces operational errors and transportation risks, thereby solving the problems mentioned in the background technology.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for optimizing a multimodal transport service network, comprising the following steps:

[0008] S101. Construct a unified nanosecond-level global synchronization clock reference to synchronize the time of various hub nodes under multiple transportation modes. Based on the synchronization clock reference, collect the first frequency data and the second frequency data of each hub node in real time. The first frequency data is the refresh frequency of the information flow, and the second frequency data is the displacement frequency of the physical flow of logistics. Ensure that the collected first frequency data and the second frequency data have a unified nanosecond-level time calibration.

[0009] S102. Based on a unified nanosecond-level global synchronization clock reference and combined with a blockchain-based multi-party verifiable timestamp mechanism, the first frequency data and the second frequency data are subjected to cross-transportation mode time domain deep calibration. A timing consistency matrix is ​​dynamically generated based on the calibrated first frequency data and the second frequency data. The timing consistency matrix is ​​used to ensure the time synchronization of multiple transportation modes throughout the entire process.

[0010] S103. Based on the time-series consistency matrix, extract the transmission speed vector corresponding to the information flow and the displacement speed vector corresponding to the physical flow of logistics, and construct a speed benchmark matrix under a unified scale based on the transmission speed vector and the displacement speed vector. The speed benchmark matrix is ​​used to provide a data foundation and scale unification for subsequent frequency domain and phase analysis.

[0011] S104. Based on the velocity benchmark matrix, the multi-scale wavelet phase difference analysis method is applied to analyze the phase difference and mismatch amplitude of the transmission velocity vector and the displacement velocity vector in multiple time windows, obtain the phase offset spectrum and mismatch amplitude of the information flow and logistics physical flow velocity vector, and form a synchronization deviation feature set covering multiple time scales.

[0012] S105. Based on the synchronization deviation feature set, an adaptive entropy weighting mechanism is adopted to dynamically weight the synchronization deviation features at different time scales. The real-time imbalance index is calculated based on the weighting result. The real-time imbalance index is used to quantify the degree of synchronization imbalance between information flow and physical logistics flow.

[0013] S106. Based on the real-time imbalance index-driven incremental residual matching mechanism of digital twin, the digital twin continuously tracks the residual dynamics between the physical flow trajectory of logistics and information flow data. Based on the residual dynamics, high-frequency hypothetical state update frames are generated. In the case of information flow lag or breakpoints, the hypothetical state update frames fill the gaps in the system perception and maintain the continuity and accuracy of scheduling decisions in the multimodal transport service network.

[0014] Preferably, S101 includes:

[0015] In various hub nodes under multiple transportation modes, a unified nanosecond-level global synchronization clock reference is established. By configuring high-precision atomic clocks or global navigation satellite system timing devices and combining them with wide-area clock synchronization protocols, time signals are synchronized to all hub nodes, and time synchronization devices are deployed to maintain a nanosecond-level error range.

[0016] Real-time acquisition of first and second frequency data, and configuration of a high-precision timestamp marking mechanism based on a synchronous clock reference, to mark each acquired data item with a nanosecond-level timestamp and verify synchronously;

[0017] Consistency checks and data cleaning were performed on the collected first and second frequency data to remove outliers. The data sequence was then optimized through interpolation repair and time series smoothing to establish a time series index based on a unified time calibration.

[0018] Preferably, S102 includes:

[0019] Based on a unified nanosecond-level global synchronous clock reference, the first frequency data and the second frequency data are initially time aligned to correct the time error caused by local clock drift or signal delay under different transportation modes.

[0020] By combining the multi-party verifiable timestamp mechanism of blockchain, the initially aligned data timestamps are written into the blockchain ledger, and the timestamps are stored and verified through distributed consensus, ensuring the integrity and authenticity of the time stamps.

[0021] Based on the blockchain verification results, a time-domain deep calibration algorithm is used to dynamically calculate the time drift and delay compensation value, and dynamic time warping and time series interpolation techniques are used to achieve fine time alignment of the data.

[0022] Based on the deeply calibrated data, a time-series consistency matrix is ​​dynamically generated to record the time stamps and data status of information flow and physical logistics flow, which is used to reflect the time synchronization relationship and change trend of the entire process.

[0023] Preferably, S103 includes:

[0024] Based on the time-series consistency matrix, the transmission speed vector of information flow and the displacement speed vector of physical logistics flow are extracted. The information flow rate is calculated by frequency increment and time difference, and the logistics speed is calculated by displacement distance and time interval.

[0025] The extracted transmission velocity vector and displacement velocity vector are normalized by maximum-minimum or standard deviation to unify the numerical scale, and weight coefficients are set according to application requirements.

[0026] A uniform-scale speed benchmark matrix is ​​constructed based on the normalized speed vector. The matrix elements record the correspondence between the speed of information flow and the speed of logistics flow, and speed difference and trend indicators are introduced.

[0027] Smoothing and completion of the velocity calibration matrix by interpolation or moving average are performed to complete time series calibration and index optimization, ensuring the data continuity and stability of frequency domain and phase analysis.

[0028] Preferably, S104 includes:

[0029] Based on the speed benchmark matrix, multiple time windows are set to segment and sample the information flow transmission speed vector and the logistics physical flow displacement speed vector to form time series pairs;

[0030] Continuous wavelet transform is applied to the velocity vector sequence within each time window, and multi-scale decomposition is performed using wavelet basis functions to extract amplitude and phase information at different frequencies, thereby obtaining the phase spectrum of each vector.

[0031] Based on the wavelet transform results, the phase difference and mismatch amplitude of the two vectors at the same time scale and frequency are calculated to form the phase shift spectrum and mismatch amplitude spectrum.

[0032] Based on the phase offset spectrum and mismatch amplitude spectrum, a synchronization deviation feature set including indicators such as average phase difference, phase difference variance, and maximum mismatch amplitude is constructed, and scale unification and sequence encoding are achieved through feature normalization and time sequence identification.

[0033] Preferably, S105 includes:

[0034] Based on the synchronization deviation feature set, an initial entropy weight model for feature weight evaluation is established. The entropy value of each feature is calculated through information entropy to form the initial weight allocation.

[0035] Based on the entropy weight model, the volatility and trend of features under the current time series are dynamically evaluated by combining a sliding time window. An adaptive adjustment mechanism is adopted to dynamically correct the feature weights, and a time scale weight adjustment factor is introduced.

[0036] Based on dynamically weighted feature weights, the synchronization deviation feature set is weighted and summed to form a weighted comprehensive deviation sequence, which quantifies the global synchronization deviation.

[0037] Based on the weighted composite deviation sequence, combined with statistical normalization and dynamic baseline comparison, a real-time imbalance index is calculated, and the sensitivity and robustness of the real-time imbalance index are enhanced by exponential smoothing and trend correction factors.

[0038] Preferably, S106 includes:

[0039] Based on the dynamic changes of the real-time imbalance index, the activation conditions and frequency of incremental residual tracking of the digital twin are determined, and the residual increment between the physical flow trajectory of logistics and the information flow data is dynamically calculated.

[0040] Based on continuous tracking of incremental residuals, trend extraction, fluctuation analysis and mutation detection algorithms are applied to extract dynamic features of residuals and identify lag patterns, breakpoint patterns or abnormal drift patterns.

[0041] Based on the dynamic characteristics of the residuals, time series prediction and state space modeling are used to generate high-frequency hypothetical state update frames, predict the expected state of the information flow, and label the confidence level and error range.

[0042] Based on hypothetical state update frames, the scheduling decision logic is dynamically completed, logistics trajectory and information flow data are integrated to eliminate perception blind spots, and the accuracy and adaptability of the prediction model are optimized through continuous comparison and correction.

[0043] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0044] This invention constructs a unified nanosecond-level global synchronous clock benchmark and combines it with a blockchain-based multi-party verifiable timestamp mechanism to achieve high-precision time synchronization and deep cross-mode calibration of information flow and physical logistics flow under various transportation modes, comprehensively improving the time consistency and data reliability of the entire multimodal transport process. Through the combined application of a speed benchmark matrix and multi-scale wavelet phase difference analysis, it can accurately characterize the synchronization deviation features of information flow and physical logistics flow at different time scales, dynamically assessing and quantifying the degree of synchronization imbalance. Furthermore, relying on the real-time imbalance index and the incremental residual matching mechanism of digital twins, it continuously tracks and dynamically fills in the perception gaps in the information flow. Especially in scenarios where information flow is lagging or interrupted, it can maintain the continuity and accuracy of scheduling decisions through the generation of high-frequency hypothetical state update frames. Overall, it significantly improves the end-to-end visualization level and scheduling intelligence capabilities of the multimodal transport service network, enhances the stability, resilience, and security of the logistics network in complex and ever-changing environments, and effectively reduces the error rate of logistics operations and transportation safety risks. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0046] Figure 1 This is a flowchart of a multimodal transport service network optimization method according to the present invention. Detailed Implementation

[0047] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0048] This invention provides, for example Figure 1 The multimodal transport service network optimization method shown includes the following steps:

[0049] S101. Construct a unified nanosecond-level global synchronization clock reference to synchronize the time of various hub nodes under multiple transportation modes. Based on the synchronization clock reference, collect the first frequency data and the second frequency data of each hub node in real time. The first frequency data is the refresh frequency of the information flow, and the second frequency data is the displacement frequency of the physical flow of logistics. Ensure that the collected first frequency data and the second frequency data have a unified nanosecond-level time calibration.

[0050] The specific implementation method is as follows:

[0051] First, a unified nanosecond-level global synchronization clock reference is established. Specifically, this is achieved by configuring high-precision atomic clocks or global navigation satellite system timing devices, combined with a wide-area clock synchronization protocol, to synchronously transmit time signals to all hub nodes participating in multimodal transport. This ensures that various hub nodes operating under different transport modes collaborate within the same time reference frame. By deploying highly stable time synchronization devices at each hub node, the local clocks of all hub nodes are guaranteed to maintain an error range within the nanosecond range compared to the global synchronization clock reference. Furthermore, to ensure time synchronization accuracy over long-term operation, a self-checking and dynamic correction mechanism for time synchronization is implemented. This mechanism periodically compares the error between the local clock and the global synchronization clock reference and automatically corrects the timing accuracy of the local clock based on the comparison results.

[0052] Based on achieving nanosecond-level time synchronization across the entire domain, real-time acquisition of first and second frequency data is implemented for various hub nodes under multiple transportation modes. Specifically, during the data acquisition process at hub nodes, the frequency of operations such as data refresh, information interaction, and data update is detected and recorded in real time along the information flow processing link to form the first frequency data. Simultaneously, during the physical flow of logistics, the displacement frequency of goods at different nodes is measured in real time by continuously monitoring the physical movement path, displacement distance, and arrival and departure times of goods to obtain the second frequency data. Both types of frequency data acquisition require precise time stamping under a unified time base to avoid time deviations caused by different time sources.

[0053] To ensure a unified nanosecond-level time calibration for both the first and second frequency data, a high-precision timestamp marking mechanism must be configured during the acquisition process. This mechanism, based on a globally synchronized clock reference, assigns a precise timestamp to each information stream refresh or physical displacement capture event and performs nanosecond-level synchronization verification on all timestamps. Through timestamp marking and verification, it is ensured that the first and second frequency data have a one-to-one time reference on the same time scale, achieving time alignment and precision consistency for different data types.

[0054] To ensure the accuracy and integrity of the data calibration, consistency verification and data cleaning are also required for the collected first and second frequency data. Specifically, this includes removing data anomalies caused by sensor errors, signal delays, or abnormal fluctuations; correcting and optimizing the data sequence using methods such as interpolation repair and time series smoothing; and establishing a time series index for the datasets of information flow and physical logistics flow based on the unified time calibration results. This will enable accurate modeling and analysis of the synchronization relationship, frequency response, and phase differences between information flow and logistics flow in subsequent operations.

[0055] By implementing the above steps in a coordinated manner, it is possible to effectively achieve full-domain time synchronization and high-precision frequency data acquisition for multimodal transport hubs, providing a solid data foundation and time guarantee for subsequent time domain calibration and synchronization analysis.

[0056] S102. Based on a unified nanosecond-level global synchronization clock reference and combined with a blockchain-based multi-party verifiable timestamp mechanism, the first frequency data and the second frequency data are subjected to cross-transportation mode time domain deep calibration. A timing consistency matrix is ​​dynamically generated based on the calibrated first frequency data and the second frequency data. The timing consistency matrix is ​​used to ensure the time synchronization of multiple transportation modes throughout the entire process.

[0057] The specific implementation method is as follows:

[0058] First, based on an established unified nanosecond-level global synchronous clock reference, preliminary time alignment is performed on the first and second frequency data collected from various hub nodes under different transportation modes. This step involves verifying the timestamps and time series of all collected data, correcting data with time deviations based on the nanosecond-level time reference, and adjusting time errors caused by equipment local clock drift and signal transmission delays in different transportation modes to a unified global time frame. This preliminary alignment operation ensures basic temporal consistency between different data sources, laying the foundation for subsequent in-depth calibration.

[0059] By combining a blockchain-based multi-party verifiable timestamp mechanism, reliable time stamping and verification are performed on the initially aligned first-frequency and second-frequency data. Specifically, the timestamp information of the first-frequency and second-frequency data is written into the blockchain ledger using a hash method. Through the distributed consensus mechanism of the blockchain, multi-party verification and tamper-proof storage of the timestamps are achieved. By storing and verifying all key time points on the blockchain, the integrity and authenticity of the timestamps are ensured, avoiding time inconsistencies caused by data forgery or record tampering in scenarios with multiple transportation modes and multiple participants. This ensures the credibility and transparency of the subsequent calibration process.

[0060] Based on the blockchain verification results using the aforementioned timestamps, a time-domain deep calibration algorithm is employed to perform fine-grained time alignment between the first and second frequency data. This deep calibration analyzes the time-series characteristics of the first and second frequency data, combines the timestamp differences before and after correction, and dynamically calculates the time drift and delay compensation values ​​for cross-transportation mode data, thereby achieving accurate mapping of data from different modes onto a unified timeline. During this process, to address the non-linear changes in time drift across multiple transportation modes, a dynamic time warping algorithm and time-series interpolation techniques are used to correct time alignment errors in different time periods, ensuring that each data point forms a strictly consistent time stamp throughout the entire process.

[0061] Based on the first and second frequency data that have undergone deep time-domain calibration, a time-series consistency matrix is ​​dynamically generated. This matrix maps the time series of the first and second frequency data onto the row and column axes of a two-dimensional matrix. Matrix elements record the time stamps and data states of information flow and physical logistics at corresponding time points. This matrix can intuitively reflect the time synchronization relationship and changing trends of data under different transportation modes. The time-series consistency matrix not only provides a standardized time mapping basis for subsequent speed benchmarking and phase analysis of information flow and logistics, but also serves as an important supporting tool for multimodal transport end-to-end visualization, scheduling optimization, and anomaly detection.

[0062] Through the progressive and synergistic effects of the above steps, the time consistency and data reliability of information flow and physical logistics flow across transportation modes have been comprehensively improved, achieving high precision, full-process time synchronization and dynamic optimization of the multimodal transport service network.

[0063] S103. Based on the time-series consistency matrix, extract the transmission speed vector corresponding to the information flow and the displacement speed vector corresponding to the physical flow of logistics, and construct a speed benchmark matrix under a unified scale based on the transmission speed vector and the displacement speed vector. The speed benchmark matrix is ​​used to provide a data foundation and scale unification for subsequent frequency domain and phase analysis.

[0064] The specific implementation method is as follows:

[0065] First, based on the temporal consistency matrix, for each time node of the information flow and the physical flow of logistics, the transmission velocity vector corresponding to the information flow and the displacement velocity vector corresponding to the physical flow of logistics are extracted respectively. Specifically, for the information flow, the rate of change of the information flow at each time point is quantified by calculating the frequency increment and time difference of adjacent information state changes at each time node, forming the transmission velocity vector of the information flow. For the physical flow of logistics, the displacement velocity per unit time is calculated based on the displacement distance of each item in physical space and the corresponding time interval, forming the displacement velocity vector of the physical flow of logistics. This step ensures that the data of the information flow and the physical flow of logistics are expressed in the form of velocity vectors, establishing a direct quantitative basis for their dynamic changes.

[0066] The extracted transmission velocity vector and displacement velocity vector are subjected to scale normalization. Because the transmission velocity of information flow and the displacement velocity of physical logistics differ in units, orders of magnitude, and magnitude of change, direct comparison would result in inconsistent data dimensions, affecting the accuracy of subsequent analysis. Therefore, methods based on max-min normalization or standard deviation normalization are used to map both types of velocity vectors to the same numerical range or standard distribution, ensuring comparability and operability of the velocity vectors at the same numerical scale. Simultaneously, during the normalization process, appropriate weighting coefficients are set for the velocity vectors of information flow and physical logistics according to actual application needs, to enhance numerical sensitivity to specific circulation stages or key nodes.

[0067] Based on the normalized transmission velocity vector and displacement velocity vector, a velocity benchmarking matrix with a unified scale is constructed. The row and column axes of this matrix correspond to the time series of information flow and physical logistics flow, respectively. Each element in the matrix forms a velocity benchmarking relationship by matching the velocity values ​​of information flow and physical logistics flow at the corresponding time point. To enhance the representational capability of the velocity benchmarking matrix, the absolute value of velocity differences, relative ratios, and trend indicators of velocity changes are further introduced as supplementary dimensions to comprehensively reflect the dynamic synchronization and velocity synergy of information flow and logistics flow at different time points. This velocity benchmarking matrix not only records static velocity comparison data but also covers the dynamic trajectory of velocity changes, forming a velocity benchmarking mapping with complete data dimensions and rich representational capabilities.

[0068] To ensure the adaptability and data continuity of the velocity benchmark matrix in subsequent frequency domain and phase analysis, data smoothing and missing data completion were performed on the constructed velocity benchmark matrix. Mathematical methods such as interpolation and moving averages were used to correct gaps and abrupt changes caused by discontinuous data acquisition or abnormal fluctuations, improving the stability and continuity of the matrix data. Furthermore, the matrix underwent time-series calibration and index optimization to ensure flexible access and parsing across multiple time scales, providing a stable, accurate, and traceable data foundation for subsequent frequency domain-based wavelet phase difference analysis and the extraction of synchronization deviation features.

[0069] Through the systematic implementation of the above steps, precise alignment of information flow and physical logistics flow at the speed level has been achieved, and a speed alignment matrix with unified scale, complete structure and dynamic response capability has been established, which greatly enhances the data processing depth and intelligent analysis capability of this invention in the optimization of multimodal transport service networks.

[0070] S104. Based on the velocity benchmark matrix, the multi-scale wavelet phase difference analysis method is applied to analyze the phase difference and mismatch amplitude of the transmission velocity vector and the displacement velocity vector in multiple time windows, obtain the phase offset spectrum and mismatch amplitude of the information flow and logistics physical flow velocity vector, and form a synchronization deviation feature set covering multiple time scales.

[0071] The specific implementation method is as follows:

[0072] First, based on the information flow transmission velocity vector and the physical displacement velocity vector of logistics recorded in the velocity benchmark matrix, multiple time windows are set for segmented data sampling. The multiple time windows are set according to the characteristics of the actual logistics circulation cycle and information flow refresh cycle, covering multiple scales from seconds, minutes, hours to days, to ensure that short-cycle changes, periodic fluctuations, and synchronization deviation characteristics under long-term trends can be captured. Within each time window, the corresponding velocity vector sequence is extracted to form the time series pair to be analyzed, providing raw data support for subsequent wavelet transform and phase difference calculation.

[0073] A continuous wavelet transform method is applied to the velocity vector sequences within each time window, projecting the velocity vectors of information flow and physical logistics displacement onto the time-frequency domain. By selecting wavelet basis functions with good time-frequency localization characteristics, the velocity vector sequences are decomposed at multiple scales to obtain amplitude and phase information at different frequency components. This step not only distinguishes different frequency characteristics in velocity changes but also reveals the phase distribution characteristics at each time scale. Through wavelet transform processing, the phase spectra of information flow and physical logistics are obtained, laying the foundation for accurate calculation of subsequent phase differences.

[0074] After performing wavelet transform and obtaining the phase information of each vector, the phase difference between the information flow velocity vector and the logistics physical flow displacement velocity vector is calculated based on the correspondence between the same time scale and frequency components. Specifically, for each frequency level and time segment, the phase shift of the two vectors is calculated point by point to form a complete phase difference sequence. Simultaneously, based on the amplitude difference and phase difference dynamics of the two vectors, the mismatch amplitude is further calculated, i.e., the degree of amplitude deviation between the information flow and logistics flow in terms of velocity change trends under the same time scale and frequency conditions. Through the joint calculation of phase difference and mismatch amplitude, the phase shift spectrum and mismatch amplitude spectrum of information flow and logistics flow are constructed, comprehensively characterizing the synchronicity and coordination between the two at different time scales.

[0075] Based on phase offset spectra and mismatch amplitude spectra obtained under multiple time windows and frequency levels, a synchronization deviation feature set covering multiple time scales is constructed. This feature set includes multi-dimensional indicators such as average phase difference, phase difference variance, maximum mismatch amplitude, minimum mismatch amplitude, and amplitude volatility at different time scales, used to quantitatively describe the synchronization deviation and inconsistency characteristics of information flow and logistics flow at various time scales. To improve the practicality and generalization ability of the feature set, feature normalization and time-series labeling methods are further adopted to unify the scale and serialize the feature data, ensuring that it can be directly applied in subsequent synchronization imbalance index calculation, dynamic scheduling optimization, and anomaly detection analysis.

[0076] Through the coordinated processing of the above steps, not only is in-depth analysis of information flow and physical logistics flow achieved across multiple time scales and frequency domains, but also solid feature support and quantitative basis are provided for a comprehensive and accurate understanding of their synchronicity and dynamic imbalance.

[0077] S105. Based on the synchronization deviation feature set, an adaptive entropy weighting mechanism is adopted to dynamically weight the synchronization deviation features at different time scales. The real-time imbalance index is calculated based on the weighting result. The real-time imbalance index is used to quantify the degree of synchronization imbalance between information flow and physical logistics flow.

[0078] The specific implementation method is as follows:

[0079] For the obtained synchronization deviation feature set covering multiple time scales, an initial entropy weight model for feature weight evaluation is established. This model first performs statistical analysis on the distribution characteristics of each feature in the full sample sequence, extracting statistics such as standard deviation, mean, and range for each feature. Based on the principle of information entropy, the entropy value of each feature is calculated. The smaller the entropy value, the greater the variability of that feature in the full sample, indicating a higher contribution of that feature to distinguishing different synchronization deviation states, and vice versa. Through this statistical process, a preliminary set of entropy weights reflecting the information content and variability of features is formed, providing a basic weight allocation framework for subsequent dynamic weighting.

[0080] Based on the entropy weight model, an adaptive adjustment mechanism is introduced to dynamically adjust feature weights. Specifically, the sensitivity and contribution of each synchronization deviation feature are evaluated based on the real-time feature performance in the current time series. In this evaluation process, a sliding time window method is used to dynamically calculate the coefficient of variation and fluctuation amplitude of each feature in the latest time period. If a feature exhibits significant volatility and trend within the current window, its weight is increased; conversely, its weight is decreased. Furthermore, considering the impact of different time scales on global synchronization deviation, a time scale weight adjustment factor is introduced to redistribute the weights of short-cycle, medium-cycle, and long-cycle features. This adaptive weight adjustment mechanism ensures the real-time nature and sensitivity of weight allocation, adjusting the importance of features according to the dynamic changes in the logistics scenario.

[0081] Based on the dynamically weighted feature weights, all features in the synchronization deviation feature set are summed in a weighted manner to form a comprehensive index sequence, namely the weighted comprehensive deviation sequence. Specifically, the normalized value of each feature at each time scale is multiplied by its corresponding dynamic weight, and all products are then summed according to the time scale to which the feature belongs, forming a dynamically weighted index that reflects the global synchronization state. This step ensures that the contribution of different features to the global synchronization deviation is fully reflected, and the weight allocation fully reflects the dynamic force of each feature in the current time state. Through this comprehensive method of weighted summation, the multi-dimensional, multi-time-scale feature set is transformed into a single, quantifiable index expression, facilitating further numerical analysis and application.

[0082] After obtaining the weighted composite deviation sequence, a real-time imbalance index is calculated as a core indicator to quantify the degree of synchronous imbalance between information flow and physical logistics flow. The calculation of the real-time imbalance index combines statistical normalization and dynamic baseline comparison methods, comparing the latest value in the weighted composite deviation sequence with the historical baseline range to determine its position within the normal fluctuation range. When the real-time imbalance index deviates from the fluctuation range set by the historical mean and variance, the real-time imbalance index is adjusted upwards accordingly, reflecting an aggravation of synchronous imbalance; conversely, the real-time imbalance index decreases, indicating a recovery in synchronicity. To enhance the sensitivity and robustness of the real-time imbalance index, exponential smoothing and trend correction factors are further introduced into the calculation to avoid misjudgments caused by short-term extreme fluctuations. The final output of the real-time imbalance index not only has the ability to quantitatively describe the current synchronous deviation but can also trigger early warnings of synchronous imbalance by setting thresholds, guiding dynamic optimization and risk intervention in logistics scheduling.

[0083] Through the orderly execution of the above steps, a mechanism for assessing synchronization deviation weights and quantifying the index is formed, which fully supports the accurate monitoring and intelligent adjustment of the synchronization between information flow and physical logistics flow in the multimodal transport service network.

[0084] S106. Based on the real-time imbalance index-driven incremental residual matching mechanism of digital twin, the digital twin continuously tracks the residual dynamics between the physical flow trajectory of logistics and information flow data, and generates high-frequency hypothetical state update frames based on the residual dynamics. In the scenario where there is a lag or break in the information flow, the hypothetical state update frames fill the gaps in the system perception and maintain the continuity and accuracy of scheduling decisions in the multimodal transport service network.

[0085] The specific implementation method is as follows:

[0086] First, based on the dynamic changes in the real-time imbalance index, the initiation conditions and tracking frequency of incremental residual tracking for the digital twin are determined. When the real-time imbalance index reaches a set dynamic threshold, it indicates that the synchronization between information flow and physical logistics flow has shown an imbalance trend, and the digital twin immediately initiates the incremental residual tracking process. This process constructs a virtual simulation path corresponding to the physical logistics flow path in the digital twin, and dynamically calculates the residual increment at each time node based on the numerical residual between the latest physical logistics flow trajectory and historical information flow data. At this time, the physical logistics flow trajectory uses its multi-dimensional dynamic attributes such as displacement, velocity, and acceleration as the basic reference, while the information flow data uses attributes such as timestamp, update frequency, and data integrity as the comparison basis. The residual between the two is output in real time in the form of a continuous numerical sequence, providing basic data support for the subsequent generation of hypothetical states.

[0087] Based on continuous tracking of incremental residuals, a feature extraction and pattern recognition model based on residual dynamics is established. By applying algorithms such as trend extraction, fluctuation analysis, and abrupt change detection to the incremental residual sequence, key characteristic parameters in the residual dynamics are identified, including residual accumulation, residual volatility, and residual abrupt change points, thereby determining the lag pattern, breakpoint pattern, or abnormal drift pattern of the information flow. The results of this feature extraction and pattern recognition are used to guide the generation strategy of hypothetical state update frames, ensuring that the generated hypothetical state update frames are not merely simple extensions of existing data, but intelligently predict the possible states and trends of the information flow based on the dynamic reality of the physical flow of logistics and the characteristics of residuals.

[0088] Based on the pattern recognition results of residual dynamics, high-frequency hypothetical state update frames are generated in real time. These hypothetical state update frames dynamically fit and intelligently complete the information flow in scenarios of lag or breakpoints. Specifically, in a digital twin, based on the actual trajectory of the current physical flow of logistics and the dynamic characteristics of residuals, combined with the residual evolution patterns of similar historical scenarios, time series prediction and state-space modeling methods are used to predict the expected state of the information flow in the next time slice and several future time slices, generating virtual data frames with high temporal resolution. These high-frequency hypothetical state update frames not only cover the estimated timestamp, information flow refresh frequency, and data integrity attributes, but also annotate the corresponding confidence level and error range, ensuring usability and risk controllability in practical applications.

[0089] Based on the generated high-frequency hypothetical state update frames, the scheduling decision-making logic of the multimodal transport service network is dynamically supplemented and corrected. By fusing the hypothetical state update frames with real-time logistics physical flow trajectories and existing information flow data, a global perception view is constructed, eliminating perception blind spots caused by information flow lags or breakpoints. This ensures that the scheduling system can maintain accurate understanding and scientific decision-making regarding the multimodal transport process even when perception data is incomplete. Simultaneously, the digital twin continuously compares and corrects the generated hypothetical state update frames with real data based on subsequent supplementation and recovery of actual information flow, dynamically revising the prediction model and improving the accuracy and adaptability of subsequent hypothetical state generation.

[0090] Through the coordinated operation of the above steps, not only is the problem of perception gap and lag between multimodal transport information flow and physical logistics flow effectively solved, but the scheduling continuity, decision-making accuracy and dynamic recovery capability of the multimodal transport network are also enhanced, giving full play to the innovative and practical value of digital twins and incremental residual matching in the field of smart logistics.

[0091] This invention constructs a unified nanosecond-level global synchronous clock benchmark and combines it with a blockchain-based multi-party verifiable timestamp mechanism to achieve high-precision time synchronization and deep cross-mode calibration of information flow and physical logistics flow under various transportation modes, comprehensively improving the time consistency and data reliability of the entire multimodal transport process. Through the combined application of a speed benchmark matrix and multi-scale wavelet phase difference analysis, it can accurately characterize the synchronization deviation features of information flow and physical logistics flow at different time scales, dynamically assessing and quantifying the degree of synchronization imbalance. Furthermore, relying on the real-time imbalance index and the incremental residual matching mechanism of digital twins, it continuously tracks and dynamically fills in the perception gaps in the information flow, especially in scenarios of information flow lag or breakpoints, maintaining the continuity and accuracy of scheduling decisions through the generation of high-frequency hypothetical state update frames. Overall, this invention significantly improves the end-to-end visualization level and scheduling intelligence capabilities of multimodal transport service networks, enhances the stability, resilience, and security of logistics networks in complex and ever-changing environments, effectively reduces logistics operation error rates and transportation safety risks, and is particularly suitable for the high-reliability transportation needs of high-value or sensitive goods.

[0092] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A method for optimizing a multimodal transport service network, characterized in that, Includes the following steps: S101. Construct a unified nanosecond-level global synchronization clock reference to synchronize the time of hub nodes under various transportation modes. Based on this clock reference, collect first frequency data and second frequency data in real time. The first frequency data is the information flow refresh frequency, and the second frequency data is the logistics physical flow displacement frequency. S102. Based on the clock reference and combined with the blockchain's multi-party verifiable timestamp mechanism, perform time-domain deep calibration on the first frequency data and the second frequency data to dynamically generate a timing consistency matrix. S103. Based on the time-series consistency matrix, extract the transmission speed vector of information flow and the displacement speed vector of physical logistics flow, and construct a speed benchmark matrix with a unified scale. S104. Based on the velocity calibration matrix, multi-scale wavelet phase difference analysis is applied to analyze the phase difference and mismatch amplitude between the transmitted velocity vector and the displacement velocity vector, forming a multi-time-scale synchronization deviation feature set. S105. Based on the synchronization deviation feature set, an adaptive entropy weighting mechanism is used to dynamically weight the synchronization deviation features, calculate the real-time imbalance index, and quantify the degree of synchronization imbalance between information flow and physical logistics flow. S106. Based on the real-time imbalance index, drive the incremental residual matching of the digital twin, continuously track the residual dynamics of the physical flow trajectory of logistics and information flow data, generate high-frequency hypothetical state update frames, and fill the perception gap when the information flow is lagging or interrupted.

2. The multimodal transport service network optimization method according to claim 1, characterized in that, S101 includes: In various hub nodes under multiple transportation modes, a unified nanosecond-level global synchronization clock reference is established. By configuring high-precision atomic clocks or global navigation satellite system timing devices and combining them with wide-area clock synchronization protocols, time signals are synchronized to all hub nodes, and time synchronization devices are deployed to maintain a nanosecond-level error range. Real-time acquisition of first and second frequency data, and configuration of a high-precision timestamp marking mechanism based on a synchronous clock reference, to mark each acquired data item with a nanosecond-level timestamp and verify synchronously; Consistency checks and data cleaning were performed on the collected first and second frequency data to remove outliers. The data sequence was then optimized through interpolation repair and time series smoothing to establish a time series index based on a unified time calibration.

3. The multimodal transport service network optimization method according to claim 1, characterized in that, S102 includes: Based on a unified nanosecond-level global synchronous clock reference, the first frequency data and the second frequency data are initially time aligned to correct time errors under different transportation modes. By combining the multi-party verifiable timestamp mechanism of blockchain, the initially aligned data timestamps are written into the blockchain ledger, and the storage and verification of timestamps are realized through distributed consensus. Based on the blockchain verification results, a time-domain deep calibration algorithm is used to dynamically calculate the time drift and delay compensation value, and dynamic time warping and time series interpolation techniques are used to achieve data time alignment. Based on the deeply calibrated data, a time-series consistency matrix is ​​dynamically generated to record the time stamps and data status of information flow and physical logistics flow.

4. The multimodal transport service network optimization method according to claim 3, characterized in that, S103 includes: Based on the time-series consistency matrix, the transmission speed vector of information flow and the displacement speed vector of physical logistics flow are extracted. The information flow rate is calculated by frequency increment and time difference, and the logistics speed is calculated by displacement distance and time interval. The extracted transmission velocity vector and displacement velocity vector are normalized to unify the numerical scale, and weighting coefficients are set according to application requirements. A uniform-scale speed benchmark matrix is ​​constructed based on the normalized speed vector. The matrix elements record the correspondence between the speed of information flow and the speed of logistics flow, and speed difference and trend indicators are introduced. Smoothing and completion of the velocity calibration matrix by interpolation or moving average are performed to complete the time series calibration and index optimization.

5. The method for optimizing a multimodal transport service network according to claim 1, characterized in that, S104 includes: Based on the speed benchmark matrix, multiple time windows are set to segment and sample the information flow transmission speed vector and the logistics physical flow displacement speed vector to form time series pairs; Continuous wavelet transform is applied to the velocity vector sequence within each time window, and multi-scale decomposition is performed using wavelet basis functions to extract amplitude and phase information at different frequencies, thereby obtaining the phase spectrum of each vector. Based on the wavelet transform results, the phase difference and mismatch amplitude of the two vectors at the same time scale and frequency are calculated to form the phase shift spectrum and mismatch amplitude spectrum. Based on the phase shift spectrum and mismatch amplitude spectrum, a synchronization deviation feature set is constructed, and scale unification and sequence encoding are achieved through feature normalization and time sequence identification.

6. The method for optimizing a multimodal transport service network according to claim 1, characterized in that, S105 includes: Based on the synchronization deviation feature set, an initial entropy weight model for feature weight evaluation is established. The entropy value of each feature is calculated through information entropy to form the initial weight allocation. Based on the entropy weight model, the volatility and trend of features under the current time series are dynamically evaluated by combining a sliding time window. An adaptive adjustment mechanism is adopted to dynamically correct the feature weights, and a time scale weight adjustment factor is introduced. Based on dynamically weighted feature weights, the synchronization deviation feature set is weighted and summed to form a weighted comprehensive deviation sequence, which quantifies the global synchronization deviation. Based on the weighted composite deviation sequence, combined with statistical normalization and dynamic baseline comparison, a real-time imbalance index is calculated, and the sensitivity and robustness of the real-time imbalance index are enhanced by exponential smoothing and trend correction factors.

7. The multimodal transport service network optimization method according to claim 6, characterized in that, S106 includes: Based on the dynamic changes of the real-time imbalance index, the activation conditions and frequency of incremental residual tracking of the digital twin are determined, and the residual increment between the physical flow trajectory of logistics and the information flow data is dynamically calculated. Based on continuous tracking of incremental residuals, dynamic features of residuals are extracted to identify hysteresis patterns, breakpoint patterns, or abnormal drift patterns. Based on the dynamic characteristics of the residuals, time series prediction and state space modeling are used to generate high-frequency hypothetical state update frames, predict the expected state of the information flow, and label the confidence level and error range. Based on hypothetical state update frames, the scheduling decision logic is dynamically completed, logistics trajectory and information flow data are integrated to eliminate perception blind spots, and the accuracy and adaptability of the prediction model are optimized through continuous comparison and correction.

Images