A multimodal digital logistics management method based on AI algorithms and big data analysis

Abstract

The present invention relates to the field of logistics transportation, and particularly to a digital logistics management method for multimodal transportation based on AI algorithms and big data analysis. The content includes: extracting features from the historical data of different transportation modes, constructing a feature dataset for transportation tasks, and calculating the similarity and distribution density between data points; based on the similarity and distribution density between data points, updating the positions of the data points of transportation tasks, comprehensively evaluating the paths, and selecting the optimal path. It solves the problems that the traditional transportation management method cannot comprehensively coordinate and optimize multiple transportation modes, resulting in unreasonable resource allocation and inaccurate path selection during transportation; there are decision-making delays and information lags, making it difficult to effectively respond to changes in urgent or sudden transportation tasks, lacking multi-dimensional comprehensive analysis means, unable to accurately evaluate the actual benefits of paths; lacking the path optimization ability based on big data and AI technologies, and unable to achieve the global optimum of path selection and resource scheduling.

Description

Technical Field

[0001] The present invention relates to the field of logistics and transportation, and in particular to a multimodal digital logistics management method based on AI algorithms and big data analysis. Background Art

[0002] With the rapid development of global trade and e-commerce, the logistics industry faces increasingly complex transportation tasks, especially with the growing demand for multimodal transport (such as road, rail, air, and water transport). Traditional logistics and transportation management systems typically focus on a single mode of transportation and fail to fully consider the coordination and optimization of multiple modes of transportation. This leads to low transportation efficiency, severe resource waste, and high costs. Furthermore, the rapid development of information technology has generated a large amount of transportation data, but traditional logistics and transportation management systems fail to effectively integrate this information. This is particularly true when dynamically adjusting transportation routes and scheduling, making it difficult to fully utilize real-time data for intelligent decision-making, which in turn affects the efficiency and accuracy of transportation tasks.

[0003] Furthermore, with increasingly diverse transportation needs, customers are placing higher demands on the timeliness, reliability, and transparency of logistics and transport. This makes traditional transportation scheduling methods difficult to meet the demands of the modern transportation industry. How to rationally allocate limited transportation resources, optimize transportation routes, and reduce transportation costs while ensuring transportation safety and improving transportation efficiency has become a critical issue that the logistics and transportation industry urgently needs to address.

[0004] The above-mentioned technologies have the following technical problems: lack of comprehensive coordination and optimization of various modes of transportation, resulting in irrational resource allocation and inaccurate route selection during the transportation process, which in turn increases transportation costs and time; there are problems of decision-making delays and information lags, making it difficult to effectively respond to urgent or sudden changes in transportation tasks, lack of multi-dimensional comprehensive analysis methods, and inability to accurately evaluate the actual benefits of the route; lack of route optimization capabilities based on big data and AI technology, and inability to achieve global optimization of route selection and resource scheduling, resulting in great limitations in the efficiency and accuracy of system operations. Summary of the Invention

[0005] The present invention provides a multimodal digital logistics management method based on AI algorithms and big data analysis to address the problems that traditional transportation management methods lack comprehensive coordination and optimization of multiple transportation modes, resulting in irrational resource allocation and inaccurate route selection during transportation, thereby increasing transportation costs and time; there are problems of decision-making delays and information lags, making it difficult to effectively respond to urgent or sudden changes in transportation tasks, lack multi-dimensional comprehensive analysis methods, and cannot accurately evaluate the actual benefits of routes; lack of route optimization capabilities based on big data and AI technology, and cannot achieve global optimization of route selection and resource scheduling, resulting in significant limitations in the efficiency and accuracy of system operations.

[0006] A multimodal digital logistics management method based on AI algorithms and big data analysis of the present invention specifically includes the following technical solutions:

[0007] A multimodal digital logistics management method based on AI algorithms and big data analysis includes the following steps:

[0008] S1. Extract features from the historical data of different transportation modes, construct a transportation task feature dataset, and calculate the similarity between data points; calculate the density based on the similarity between data points to obtain the distribution density of data points;

[0009] S2. Update the positions of the data points of the transportation task based on the similarity and distribution density between data points, and comprehensively evaluate the paths to select the optimal path.

[0010] Preferably, the S1 specifically includes:

[0011] Based on the transportation task feature dataset, calculate the straight-line distance and inner product between data points to obtain the similarity between data points.

[0012] Preferably, the S1 specifically includes:

[0013] Based on the similarity between data points, introduce the weighted factors of distance and inner product to obtain the distribution density of data points.

[0014] Preferably, the S2 specifically includes:

[0015] Update the data point positions based on the similarity between data points, external perturbations, and the current distribution density; the position update formula for data points is:

[0016]

[0017] where is the th data point at time , representing the position of the updated data point; is the th data point at time ; is the learning rate; represents the th data point and the th data point in the transportation task feature dataset; is the th data point at time Position; is the th data point of the external disturbance vector; is the adjustment coefficient of the external disturbance; is the adjustment coefficient; is the th data point of the distribution density; is the number of data points included in the transportation task feature dataset.

[0018] Preferably, the S2 specifically includes:

[0019] Introduce the weight calculation model of the multimodal transport path, and calculate the path weight based on the distribution density of the data points and the updated data point positions.

[0020] Preferably, the S2 specifically includes:

[0021] The weight calculation model of the multimodal transport path introduces the spatio-temporal inversion factor, and combines the distance, transport time, and historical reliability of the path to calculate the path weight; the calculation formula of the path weight is:

[0022]

[0023]

[0024] Among them, is the weight of the path at time ; is the distance of the path ; is the path at time of the transport time; is the path of the historical reliability; is the path at time of the spatio-temporal inversion factor; , , , , , are the weight coefficients; is the time decay factor; is the adjustment coefficient; is the space decay factor; is the length of the time window; is the time integral variable; is the path at time The weights on

[0025] Preferably, the S2 specifically includes:

[0026] Based on the path weights, combining the distance of the path, the transportation time, the historical reliability of the path, and the spatio-temporal inversion factor, construct an objective function, select the globally optimal path selection scheme, and obtain the optimal path.

[0027] Preferably, the S2 specifically includes:

[0028] Connect the optimal path with the logistics management system, convert the path plan and time arrangement into path instructions, and send them to each logistics operation system to achieve real-time data sharing and scheduling of each transportation mode.

[0029] The beneficial effects of the technical solution of the present invention are:

[0030] 1. By combining the AI algorithm and big data analysis, the present invention can achieve real-time path optimization and resource scheduling in a dynamically changing logistics environment; optimize the transportation path through a dynamic evolution mechanism, and introduce a weight calculation model for multimodal transportation paths, comprehensively considering multiple factors such as the path distance, transportation time, and historical reliability of different transportation modes, significantly reducing the additional costs caused by improper route selection or resource scheduling errors during transportation, and improving the overall transportation efficiency.

[0031] 2. Based on the density analysis and similarity evaluation of big data, the present invention can monitor and respond to complex changes in transportation tasks in real time, such as environmental factors and market demand fluctuations, dynamically adjust resources according to real-time data and historical data, ensure that each link in multimodal transportation can be accurately connected, and further enhance the flexibility and adaptability of the logistics management system.

[0032] 3. Through the accumulation and analysis of historical transportation data, the present invention can assign a reliability index to each transportation path, reflecting the performance of each path in past transportation tasks, helping to evaluate the stability and safety of the path, enabling the path selection to not only rely on real-time data, but also fully consider historical performance, and further improving the accuracy and safety of path selection.

[0033] 4. The present invention integrates multiple transportation modes such as highway, railway, waterway, and air, gives full play to the advantages of each transportation mode, and realizes real-time data sharing and scheduling between different transportation modes through a digital platform, not only optimizing the path selection, but also ensuring the smooth connection of each transportation link, and improving the overall logistics efficiency. Description of the Drawings

[0034] Figure 1Flowchart of a multimodal digital logistics management method based on AI algorithms and big data analysis according to the present invention. Detailed implementation manners

[0035] To further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0036] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0037] The following specifically describes the specific solution of a multimodal digital logistics management method provided by the present invention in conjunction with the accompanying drawings.

[0038] Specifically, as an embodiment, the specific solution of a multimodal digital logistics management method based on AI algorithms and big data analysis is as follows:

[0039] First, comprehensive data collection is carried out on all transportation tasks, including but not limited to the starting position, ending position, task load, required resource type, transportation time requirements, and historical transportation data of each transportation task, etc.; at the same time, feature extraction is performed on the historical data of different transportation modes through big data technology to construct a transportation task feature data set. , is the number of data points included in the transportation task feature data set; based on the transportation task feature data set, the similarity between data points is calculated, and the distribution density of each data point in space is evaluated.

[0040] Then, the position of each data point is dynamically evolved, and the position of the data point is accurately updated based on the similarity between data points, external perturbations, and the current distribution density. The position update formula of the data point is:

[0041]

[0042] Wherein, is the th data point at time , that is, the updated data point position; is the th data point at time The position, i.e., the position of the data point before update; is the learning rate, which is used to control the step size of the data point position update and is obtained through experiments; represents the th data point and the th data point between the similarities; is the th data point at time position; is the th data point external perturbation vector, indicating the external influencing factors acting on the th data point ; is the adjustment coefficient of the external perturbation, which is used to control the influence of the external influencing factors on the position update and is obtained through experiments; is the adjustment coefficient, which is used to control the influence of the distribution density on the position update and is obtained through experiments; is the th data point distribution density, reflecting the degree of aggregation of data points in space;

[0043] Furthermore, a weight calculation model for the multimodal transport path is designed, and the path weight is calculated according to the distribution density of the data points and the updated data point positions; the calculation formula for the path weight is:

[0044]

[0045]

[0046] Among them, is the weight of the path at time , indicating the quality of the path; is the distance of the path ; is the transport time of the path at time ; is the historical reliability of the path , based on the deviation between the actual time and the expected time experienced by the path at time calculated; is the spatio-temporal inversion factor of the path at time , which is used to reflect the spatio-temporal variability of the path; , , , , , are weight coefficients, used to adjust the contribution of each factor to the calculation of path weight, obtained through experiments; is the time decay factor, used to control the influence of the path changing over time, obtained through experiments; is the adjustment coefficient, used to control the smoothness of the spatio-temporal inversion factor, obtained through experiments; is the space decay factor, used to control the influence of the distance of the path on the spatio-temporal inversion factor, obtained through experiments; is the length of the time window, used to determine the calculation period of the spatio-temporal inversion factor; is the time integration variable; is the path at time weight;

[0047] Finally, based on the path weight, combined with the distance of the path, transportation time, historical reliability of the path, and spatio-temporal inversion factor, a target function is constructed to select the optimal path; the construction of the target function uses existing technologies and will not be elaborated here; the optimal path is docked with the logistics management system, and the specific path plan and time arrangement are converted into path instructions and sent to each logistics operation system to achieve real-time data sharing and scheduling of each transportation mode and ensure the smooth docking of each transportation link.

[0048] To better understand the above technical solution, the above technical solution will be described in detail below in combination with the accompanying drawings of the specification and specific implementation manners;

[0049] As an embodiment, referring to the appendix Figure 1 , which shows a flowchart of a multimodal digital logistics management method provided by an embodiment of the present invention based on AI algorithms and big data analysis. The method includes the following steps:

[0050] S1. Extract features from the historical data of different transportation modes, construct a transportation task feature data set, and calculate the similarity between data points; calculate the density based on the similarity between data points to obtain the distribution density of data points;

[0051] Comprehensively collect data for all transportation tasks, including but not limited to the starting location, ending location, task load, required resource types, transportation time requirements, and historical transportation data of each transportation task. Through sensor networks and GPS positioning technology, monitor the starting location, ending location, and transportation route of each transportation task in real time to ensure the accuracy of each data point; use Internet of Things technology (IoT) to monitor resource requirements in real time, such as monitoring the load of trucks, vehicle status, and specific requirements of goods, so as to obtain key data related to transportation tasks at any time and provide precise support for route optimization and resource scheduling.

[0052] In addition to collecting real-time data, it also relies on the integration of historical data, which contains various characteristics of past transportation tasks, such as the transportation timeliness, cost, delay situation, and problems during transportation of each route. The acquisition and processing of historical data will ensure its availability through steps such as data cleaning and normalization, and avoid deviations in subsequent optimization processes caused by incomplete or incorrect data.

[0053] Analyze the historical data of different transportation modes (such as land transportation, sea transportation, air transportation, etc.) through big data technology, extract the key characteristics of transportation tasks, and calculate the performance indicators of each transportation mode (such as transportation time, cost, risk, etc.) to construct a transportation task feature dataset , is the number of data points included in the transportation task feature dataset, which contains all transportation task information (such as: time, location, resource requirements, and load).

[0054] By evaluating the similarity between data points, the overall characteristics of transportation tasks can be better understood; through a weighted model based on Euclidean distance and inner product, calculate the similarity between data points and construct a similarity matrix to measure the degree of similarity between data points; the calculation of similarity is not only based on the straight-line distance, but also takes into account the inner product between data points, which can better express the similarity of data points in high-dimensional space, ensure that route selection and resource scheduling can be based on the comprehensive information of data points, and avoid inaccurate route selection caused by overly simple distance metrics. The construction formula of the similarity matrix is as follows:

[0055]

[0056] where, represents the similarity between the th data point and the th data point in the transportation task feature dataset, which is determined by calculating the distance between the two data points; and respectively represent the The th data point and the is the th data point and the th data point Euclidean distance between them; is the standard deviation of the similarity between the th data point and the th data point ; is the th data point and the th data point inner product between them; is a regularization term used to adjust the contribution of the inner product to the similarity, obtained through experiments. The Euclidean distance measures the direct spatial distance between data points, indicating the proximity of two data points in geometric space. At the same time, through the inner product term , the linear correlation of data points in different dimensions is further considered, helping to capture more complex similarity relationships.

[0057] In big data analysis, density analysis helps to identify the distribution characteristics of transportation tasks in space. Based on the calculated similarity matrix, density calculation is performed to evaluate the distribution density of each data point in space. The calculation of distribution density not only considers the similarity between data points, but also introduces weighted factors of distance and inner product to ensure that the density value can truly reflect the relative position and relationship of data points; by weighting the similarity and adjusting the influence range of density, it is ensured that the distribution density of each data point can accurately reflect the local aggregation characteristics of the transportation task feature dataset in space, which helps to further guide the evolution process of data points and the optimization of path selection and resource scheduling. The formula for calculating distribution density is:

[0058]

[0059] where, is the distribution density of the th data point , reflecting the degree of aggregation of data points in space; is the th data point and the th data point Euclidean distance between them, indicating the separation degree of these two data points in space; is the weight coefficient used to control the influence of the Euclidean distance on the distribution density, adjusted through experiments; is the weight coefficient, which is used to control the influence of the inner product on the distribution density and is adjusted through experiments. By calculating the distribution density of each data point, it reflects the degree of aggregation of the data points in the overall space.

[0060] S2. Based on the similarity and distribution density between data points, update the positions of the data points of the transportation tasks, and comprehensively evaluate the paths to select the optimal path.

[0061] In the process of path selection and resource scheduling, dynamic evolution is one of the core mechanisms. Each data point is updated through certain rules in each round of iteration, and the path selection is optimized through dynamic evolution. Specifically, in the process of evolution, each data point is adjusted not only by the similarity with other data points, but also by the influence of external perturbations and distribution density. Therefore, in each round of update process, it is necessary to consider how to accurately update its position according to the current state of the data point, combining the similarity matrix, external perturbations and the current distribution density. The position update formula of the data point is:

[0062]

[0063] where is the position of the th data point at time , that is, the position of the updated data point; is the position of the th data point at time , that is, the position of the data point before update; is the learning rate, which is used to control the step size of the data point position update and is obtained through experiments; is the position of the th data point at time ; is the position of the th data point ; is the external perturbation vector of the th data point , representing the external influencing factors acting on the th data point , and the influencing factors include physical forces, environmental changes, market changes, etc.; is the adjustment coefficient of the external perturbation, which is used to control the influence of the external influencing factors on the position update and is obtained through experiments; is the adjustment coefficient, which is used to control the influence of the distribution density on the position update and is obtained through experiments. The position update of each data point is not only affected by the similarity with other data points, but also affected by the external perturbation vector and the distribution density. Through dynamic evolution, the data points can gradually migrate to the optimal positions to play a better role in subsequent path selection and resource scheduling.

[0064] Considering that multimodal transportation involves different transportation modes, the path selection does not depend on a single transportation mode, but rather requires a comprehensive consideration of the advantages and disadvantages of various transportation modes. Therefore, a weight calculation model for multimodal transportation paths is designed; the weight of a path is a key indicator to measure the comprehensive benefits of the path over time and takes into account the influence of multiple factors, such as the distance of the path, transportation time, historical reliability of the path, and spatio-temporal inversion factor, etc. Specifically, the distance and transportation time of the path are the most basic measurement criteria, which respectively affect the physical distance and time cost during transportation; the historical reliability of the path reflects the performance of the path in past transportation tasks and can reflect the stability and safety of the path; the spatio-temporal inversion factor further introduces the influence of historical paths on the current path selection and considers the reference role of the running trajectory of historical paths in current decision-making; taking the distribution density and the position of updated data points as factors for dynamic adjustment can optimize the path selection in real time according to the actual changes in transportation tasks. Therefore, the calculation formula for path weight is:

[0065]

[0066]

[0067] where is the weight of path at time and represents the quality of the path; is the distance of path ; is the transportation time of path at time ; is the historical reliability of path , calculated based on the deviation between the actual time and the expected time experienced by path at time ; is the spatio-temporal inversion factor of path at time and is used to reflect the spatio-temporal variability of the path; , , , , , are weight coefficients, used to adjust the contribution of each factor to the path weight calculation, obtained through experiments; is the time decay factor, used to control the influence of the path over time, obtained through experiments; is the adjustment coefficient, used to control the smoothness of the spatio-temporal inversion factor, obtained through experiments; is the spatial attenuation factor, which is used to control the influence of the distance of the path on the spatio-temporal inversion factor and is obtained through experiments; is the length of the time window, which is used to determine the calculation period of the spatio-temporal inversion factor; is the time integration variable; is the path at time weight.

[0068] The goal of optimal path selection is to comprehensively evaluate the advantages and disadvantages of paths based on various factors and select the optimal path that can minimize the overall transportation cost and time. The factors include the distance of the path, transportation time, historical reliability of the path, and spatio-temporal inversion factor. Construct an optimal path objective function, calculate the costs of all paths through weighted calculation, and obtain a globally optimal path selection scheme to effectively meet the various requirements of the transportation task, reduce costs and improve efficiency. The optimal path objective function is as follows:

[0069]

[0070] where is the optimal path, representing the selected shortest path; is the set of all paths; and are adjustment coefficients, which are used to adjust the influence of the distance of the path and the spatio-temporal inversion factor on path selection.

[0071] Connect the optimal path with the logistics management system, convert the specific path plan and time arrangement into path instructions, and send them to each logistics operation system, including the control systems of various transportation tools such as transport vehicles, ships, and airplanes. The path instructions include information such as transportation time, starting point, ending point, and path nodes. The optimal path involves the coordinated operation of multiple transportation modes, such as road, railway, waterway, and air transportation. Through a digital platform, real-time data sharing and scheduling of each transportation mode are realized to ensure the smooth connection of each transportation link.

[0072] Through Internet of Things devices and sensors, real-time obtain data of each transportation link, including information such as transportation progress, cargo status, and environmental conditions. All information is summarized and analyzed through a digital platform and provided to operators and decision-makers to ensure that they can timely understand every detail in the logistics process and make necessary adjustments.

[0073] In summary, a digital logistics management method for multimodal transport based on AI algorithms and big data analysis is completed.

[0074] The order of the invention embodiments is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0075] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. The key point of each embodiment is to illustrate the differences from other embodiments.

[0076] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments or perform equivalent replacements for some of the technical features. These modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of each embodiment of the present invention and should all be included in the protection scope of the present invention.

Claims

1. A multimodal digital logistics management method based on AI algorithms and big data analysis, characterized in that, It includes the following steps: S1. Extract features from the historical data of different transportation modes, construct a transportation task feature dataset, and calculate the similarity between data points; calculate the distribution density of data points based on the similarity between data points; S2. Update the positions of the data points of the transportation task based on the similarity and distribution density between data points, comprehensively evaluate the paths, and select the optimal path.

2. The multimodal digital logistics management method based on AI algorithm and big data analysis according to claim 1, wherein, The S1 specifically includes: Based on the transportation task feature dataset, calculate the straight-line distance and inner product between data points to obtain the similarity between data points.

3. The multimodal digital logistics management method based on AI algorithm and big data analysis according to claim 2, characterized in that, The S1 specifically includes: Based on the similarity between data points, introduce the weighted factors of distance and inner product to obtain the distribution density of data points.

4. A multimodal digital logistics management method based on AI algorithms and big data analysis according to claim 1, characterized in that The S2 specifically includes: Update the positions of data points based on the similarity between data points, external perturbations, and the current distribution density; the position update formula of the data points is: Among them, is the th data point at the position of time , representing the position of the updated data point; is the th data point at the position of time ; is the learning rate; represents the similarity between the th data point and the th data point in the transportation task feature dataset; is the th data point at the position of time ; is the th data point 's external disturbance vector; is the adjustment coefficient of the external disturbance; is the adjustment coefficient; is the th data point 's distribution density; is the number of data points included in the transportation task feature dataset.

5. A multimodal digital logistics management method based on AI algorithms and big data analysis according to claim 4, characterized in that, The S2 specifically includes: Introduce a weight calculation model for the multimodal transportation path, and calculate the path weight based on the distribution density of data points and the updated positions of data points.

6. The multimodal digital logistics management method based on AI algorithm and big data analysis according to claim 5, characterized in that, The S2 specifically includes: The weight calculation model for the multimodal transportation path introduces a space-time inversion factor, and combines the distance, transportation time, and historical reliability of the path to calculate the path weight; the calculation formula of the path weight is: Among them, is the path at time weight; is the path distance; is the path at time transportation time; is the path historical reliability; is the path at time space-time inversion factor; , , , , , are weight coefficients; is the time decay factor; is the adjustment coefficient; is the space decay factor; is the length of the time window; is the time integration variable; is the path at time weight.

7. An intermodal digital logistics management method based on AI algorithms and big data analysis according to claim 6, characterized in that, The S2 specifically includes: Based on the path weight, combine the distance, transportation time, historical reliability of the path, and space-time inversion factor to construct an objective function, select the globally optimal path selection scheme, and obtain the optimal path.

8. A multimodal digital logistics management method based on AI algorithms and big data analysis according to claim 7, characterized in that, The S2 specifically includes: Connect the optimal path with the logistics management system, convert the path plan and time arrangement into path instructions, and send them to each logistics operation system to realize real-time data sharing and scheduling of each transportation mode.

Images