MaaS platform-oriented big language model-driven multimodal transport path planning method and system, medium and electronic equipment

Abstract

The invention provides a MaaS platform-oriented big language model-driven multimodal transport path planning method and system, a medium and electronic equipment. The method comprises the following steps: acquiring basic data based on a MaaS platform to construct a description text; analyzing and converting the description text by utilizing a preset large language model to obtain a multi-dimensional decision parameter; performing logic verification on the multi-dimensional decision parameter based on a preset topological structure to obtain a target parameter, wherein the target parameter comprises a standardized calculation parameter required by a travel path; and performing global optimization on the target parameter based on a preset target collaborative constraint model to obtain an optimal solution, and converting data content corresponding to the optimal solution into natural language description by using a large language model and outputting the natural language description. According to the method, the precision of processing natural language requirements in a multimodal transport path optimization scene is improved, intelligent processing of path planning is realized, and excellent robustness and accuracy are achieved when complex node information and dynamic requirements are processed.

Description

Technical Field

[0001] This application belongs to the field of data processing technology, and in particular relates to a method, system, medium and electronic equipment for multimodal transport route planning driven by a large language model for MaaS platforms. Background Technology

[0002] Multimodal transport is a freight transport service that relies on multiple modes of transport such as rail, road, water and air to achieve integrated freight transport throughout the entire process. Unlike traditional logistics methods that rely on a single mode of transport, multimodal transport can significantly improve transport efficiency, reduce the overall logistics costs of enterprises, and optimize the cross-regional logistics network structure through the seamless connection of different transport tools. It is a strategic direction for the development of the global logistics industry.

[0003] With the rise of MaaS (Model as a Service) platforms for logistics, which integrate transportation information data from water, rail, road, and air through a one-stop portal and provide users with combined solutions such as rail-water intermodal transport and road-water intermodal transport, it has become an important infrastructure for building a modern logistics system.

[0004] Currently, Large Language Models (LLMs) have been initially applied in the field of multimodal transport route planning. However, existing related patents mostly focus on multimodal data and maritime scenarios, and generally adopt multi-objective optimization algorithms, without fully considering the impact of variable constraints. Among them, traditional multimodal transport planning methods (such as operations research optimization methods based on hard constraint rules, simulated annealing algorithms, etc.) face the following prominent technical challenges in practical implementation compared to Large Language Models:

[0005] 1. The interaction method is outdated and there is a barrier to human-computer interaction. Traditional methods rely heavily on manual input of parameters into pre-set structured forms, which is a cumbersome process.

[0006] 2. The planning efficiency is low and the response agility is insufficient. When facing complex and ever-changing cross-regional intermodal transport scenarios, the entire process from demand proposal to solution generation is slow and the conversion efficiency is low, which fails to meet the core requirements of modern logistics for agility.

[0007] 3. Limited technical framework and lack of semantic understanding capabilities: Traditional algorithm frameworks themselves do not have semantic parsing capabilities. Their optimization logic requires a lot of manual "translation" and preprocessing work to meet the user's natural language needs, resulting in low system intelligence and poor scalability. Summary of the Invention

[0008] In view of the shortcomings of the prior art described above, the purpose of this application is to provide a method, system, medium and electronic device for multimodal transport route planning driven by a large language model for MaaS platforms, so as to solve the problems of outdated interaction methods, low planning efficiency and limited technical framework in the multimodal transport route planning of the prior art.

[0009] Firstly, this application provides a multimodal transport route planning method driven by a large language model for MaaS platforms, the method comprising:

[0010] Based on the basic data obtained from the MaaS platform, a descriptive text is constructed. The content of the descriptive text includes at least the unit transportation cost of each segment, the transportation time of each segment, the transportation demand quantity of each type of cargo box, the optimization objective, the constraints, the origin / end point of the journey, and the transfer node information.

[0011] The descriptive text is parsed and transformed using a pre-defined large language model to obtain multidimensional decision parameters, which include structured data corresponding to the descriptive text.

[0012] The multidimensional decision parameters are logically verified based on a preset topology to obtain target parameters, which include standardized calculation parameters required for the travel path.

[0013] The optimal solution is obtained by globally optimizing the target parameters based on the preset target collaborative constraint model. The data content corresponding to the optimal solution is then converted into a natural language description and output using a large language model.

[0014] In some embodiments of the first aspect of this application, the step of acquiring basic data based on the MaaS platform to construct the descriptive text specifically includes:

[0015] Basic data is acquired based on the MaaS platform, including transportation information data and user requirements stored in tabular form.

[0016] The description text is constructed by converting information data, user needs, transportation costs, transportation time, demand quantity, optimization objectives, constraints, and the origin / destination and transfer nodes of the travel route into natural language instructions containing parameters to be parsed using a fixed text template. The formula of the description text is as follows:

[0017] ;

[0018] in, This refers to the descriptive text. Indicates specifications Cargo container on the journey path Transportation between nodes Real-time per-box shipping cost Indicates the route Transportation between nodes The time of delivery This represents the optimization objective, which includes a time objective and / or a cost objective. This represents the constraints, which include time constraints and / or cost constraints. Indicates the starting point of the journey. Indicates the end point of the journey. This indicates the total demand for containers. This represents a set of container specifications. A set of nodes representing the travel path.

[0019] In some embodiments of the first aspect of this application, the step of parsing and transforming the descriptive text using a preset large language model to obtain multidimensional decision parameters specifically includes:

[0020] The large language model is used to convert the descriptive text into standardized JSON structured data using a custom prompt word template;

[0021] The multidimensional decision parameters are obtained by parsing the corresponding JSON structured data and extracting feature parameters, as shown in the following formula:

[0022] ;

[0023] in, Represents multidimensional decision parameters. Indicates the starting point of the journey. Indicates the end point of the journey. This indicates the total demand for containers. This represents the set of nodes along the travel path. Represents a set of modes of transportation. Indicates specifications Cargo container on the journey path Transportation between nodes Real-time per-box shipping cost Indicates the optimization objective. Indicates time constraints. This indicates a cost constraint.

[0024] In some embodiments of the first aspect of this application, the step of performing logical verification on the multidimensional decision parameters based on a preset topology to obtain the target parameters specifically includes:

[0025] Path node information is verified to determine whether each path is connected, and path-mode decision variables are constructed based on the directed graph topology of the multimodal transport network, as shown in the following formula:

[0026] ;

[0027] in, Indicates the route Transportation between nodes The path-transportation mode decision vector, Indicates the first Various path planning schemes This represents the set of nodes along the travel path. Represents a set of modes of transportation;

[0028] The total cost of a single arc segment is defined as the sum of the freight costs for all specifications of goods, as shown in the following formula:

[0029] ;

[0030] in, This represents the sum of freight costs for goods of all specifications and routes. This indicates the number of containers of specification 's'. Indicates specifications Cargo container on the journey path Transportation between nodes Real-time per-box shipping cost Indicates the route Transportation between nodes The path-transportation mode decision vector, This indicates the total demand for containers. This represents a set of container specifications. Indicates the first Various path planning schemes;

[0031] The formula for calculating the route transportation time is as follows:

[0032] ;

[0033] in, Indicates the transit time along the route. Indicates the route Transportation between nodes The time of delivery Indicates the route Transportation between nodes The path-transportation mode decision vector, Indicates the first Various path planning schemes;

[0034] Based on the sum of freight costs for all specifications of goods and route transportation time Obtain the target parameters.

[0035] In some embodiments of the first aspect of this application, the step of globally optimizing the target parameters based on a preset target collaborative constraint model to obtain the optimal solution specifically includes:

[0036] The optimal solution is obtained by finding the optimal path that simultaneously satisfies the optimization objective and constraints using a preset solution tool.

[0037] The optimization objective includes minimizing the dynamic selection cost, calculated as follows:

[0038] ;

[0039] in, This represents the sum of freight costs for goods of all specifications and routes. Indicates the nodes of the travel path. Represents a set of nodes. Indicates the mode of transportation. A collection of transportation methods Indicates the first Path planning scheme, or

[0040] The optimization objective includes minimizing the dynamic selection time, calculated as follows:

[0041] ;

[0042] in, Indicates the transit time along the route. Indicates the nodes of the travel path. Represents a set of nodes. Indicates the mode of transportation. A collection of transportation methods Indicates the first Various path planning schemes;

[0043] Constraints include time constraints that are less than or equal to time constraints, calculated as follows:

[0044] ;

[0045] in, Indicates the transit time along the route. Indicates the nodes of the travel path. Represents a set of nodes. Indicates the mode of transportation. A collection of transportation methods Indicates time constraints. Indicates the first Path planning scheme, or

[0046] Constraints include those less than or equal to cost constraints, calculated as follows:

[0047] ;

[0048] in, This represents the sum of freight costs for goods of all specifications and routes. Indicates the nodes of the travel path. Represents a set of nodes. Indicates the mode of transportation. A collection of transportation methods Indicates cost constraints. Indicates the first A path planning scheme.

[0049] In some embodiments of the first aspect of this application, the step of using a large language model to convert the data content corresponding to the optimal solution into a natural language description and output it specifically includes:

[0050] Based on the optimal solution, the following data are extracted: route information, total transportation time, total transportation cost, unit transportation cost of each interval, and transportation time of each interval.

[0051] The data content corresponding to the optimal solution is converted into a natural language description using a large language model, and the converted natural language description is then output externally.

[0052] To achieve the above and other related objectives, a second aspect of this application provides a large language model-driven multimodal transport route planning system for MaaS platforms, the system comprising:

[0053] The construction module is used to acquire basic data based on the MaaS platform to construct the description text. The content of the description text includes at least the unit transportation cost of each segment, the transportation time of each segment, the transportation demand quantity of each type of cargo box, the optimization objective, the constraints, the origin / end point of the journey, and the transfer node information.

[0054] The parsing module is used to parse and transform the description text using a preset large language model to obtain multidimensional decision parameters, the multidimensional decision parameters including the structured data corresponding to the description text;

[0055] The verification module is used to perform logical verification on the decision parameters based on a preset topology to obtain target parameters, the target parameters including standardized calculation parameters required for the travel path;

[0056] The output module is used to perform global optimization of the target parameters based on a preset target collaborative constraint model to obtain the optimal solution, and to use a large language model to convert the data content corresponding to the optimal solution into a natural language description and output it.

[0057] To achieve the above and other related objectives, a third aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the large language model-driven multimodal transport path planning method for MaaS platforms as described in any of the preceding claims.

[0058] To achieve the above and other related objectives, a fourth aspect of this application provides a computer program product comprising computer program code, which, when executed on a computer, enables the computer to implement the large language model-driven multimodal transport path planning method for MaaS platforms as described above.

[0059] To achieve the above and other related objectives, a fifth aspect of this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory; the processor executes the computer program to implement the large language model-driven multimodal transport path planning method for MaaS platforms as described in any of the preceding claims.

[0060] As described above, the large language model-driven multimodal transport route planning method, system, medium, and electronic equipment for MaaS platforms described in this application improves the accuracy of handling natural language requirements in multimodal transport route optimization scenarios, realizes intelligent route planning, and exhibits excellent robustness and accuracy when handling complex node information and dynamic requirements. Specific effects are as follows:

[0061] 1. The interaction method has been upgraded, enabling automatic access to table data from the MaaS platform to generate descriptive text;

[0062] 2. Improved planning efficiency, enabling the rapid generation of optimal routes in complex and ever-changing cross-regional intermodal transport scenarios;

[0063] 3. It expands intelligent semantic understanding by using a large language model to perform multi-dimensional analysis of multi-dimensional data, thereby enhancing intelligence. Attached Figure Description

[0064] Figure 1 The diagram shows a step-by-step illustration of the large language model-driven multimodal transport route planning method for MaaS platforms in this application in one embodiment.

[0065] Figure 2 The diagram shows a flowchart of the method for obtaining descriptive text in one embodiment of the large language model-driven multimodal transport route planning method for MaaS platform according to this application.

[0066] Figure 3The diagram shows a flowchart of the method for obtaining multidimensional decision parameters in one embodiment of the large language model-driven multimodal transport route planning method for MaaS platform according to this application.

[0067] Figure 4 The diagram shows a global optimization process in one embodiment of the large language model-driven multimodal transport route planning method for MaaS platforms according to this application.

[0068] Figure 5 The diagram shows a step-by-step illustration of the large language model-driven multimodal transport route planning method for MaaS platforms in this application in one embodiment.

[0069] Figure 6 The diagram shown is a structural schematic of a MaaS platform-oriented multimodal transport route planning system driven by a large language model in one embodiment of this application.

[0070] Figure 7 The diagram shown is a structural schematic of the electronic device of this application in one embodiment.

[0071] Component designation explanation S102~S108 step S502~S504 step 60 A large language model-driven multimodal transport route planning system for MaaS platforms 61 Modules 62 Parsing module 63 Verification module 64 Output module 700 Electronic terminal 701 processor 702 memory 7021 operating system 7022 app 703 Network interface 704 bus system 705 User Interface Detailed Implementation

[0072] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0073] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. Therefore, the drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0074] Furthermore, the use of terms such as "first" and "second" in this application is for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of the various embodiments can be combined with each other, but only on the basis of being achievable by those skilled in the art. If the combination of technical solutions is contradictory or impossible to implement, such a combination of technical solutions should be considered non-existent and not within the scope of protection claimed in this application.

[0075] With the continuous growth of personalized trade demands and the increasing complexity of multimodal transport solutions, the ability to quickly plan optimal routes from massive amounts of data has become a key requirement for MaaS platform management and scheduling. Utilizing digital means to analyze transport conditions and using optimization algorithms to automatically generate routes is a currently effective technical approach. Among these, the LLM (Large Language Model) is a typical deep learning model used for semantic understanding and logical reasoning. Specifically, it treats the identification of complex logistics demands as a parameter parsing and intent extraction process: through direct processing of natural language instructions, the network can simultaneously identify key variables such as origin and destination points, timeliness preferences, and cost constraints, unlike traditional methods that rely on manually pre-set structured forms or tedious manual parameter entry. Thanks to this end-to-end semantic processing approach, the large model has advantages such as strong interactivity, high generalization, and the ability to handle unstructured information. It has been widely applied in scenarios such as intelligent customer service, decision support, and automated office work, and has evolved into multiple vertical domain versions. It achieves a good balance between logical reasoning and knowledge retrieval, making it very suitable for intelligent route planning tasks under the MaaS platform.

[0076] With the in-depth development of large language model technology in the transportation field, many key technologies for driving multimodal transport route optimization using large language models have emerged in the industry. However, the existing problems mentioned in the background still exist. Therefore, in view of the technical challenges of traditional methods and the limitations of LLM in the application of multimodal transport, this application proposes a method, system, medium, and electronic equipment for multimodal transport route planning driven by a large language model for MaaS platforms. It is applicable to multimodal transport route planning. By constructing a professional natural language instruction set and using an LLM large language model to reconstruct unstructured natural language descriptions into JSON structured data, a node information verification mechanism is introduced. After verification, the JSON structured data is transformed into standardized parameters that can be directly called by the route planning engine. Finally, relying on variable constraints, global optimization is achieved in the cost or time dimension, and the optimal transportation plan is generated by combining LLM. In this application, the technical challenges of traditional methods mentioned in the background are effectively alleviated. By integrating the advantages of large language models and route optimization solution tools, the processing accuracy of natural language descriptions for multimodal transport route optimization is effectively improved. The technical solutions in the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0077] like Figure 1 As shown, in one embodiment of the invention, the large language model-driven multimodal transport route planning method for MaaS platforms of this application includes the following steps:

[0078] Step S102: Obtain basic data based on the MaaS platform to construct the descriptive text;

[0079] Step S104: Use a preset large language model to parse and transform the description text to obtain multidimensional decision parameters;

[0080] Step S106: Logically verify the multidimensional decision parameters based on the preset topology to obtain the target parameters;

[0081] Step S108: Based on the preset target collaborative constraint model, the target parameters are globally optimized to obtain the optimal solution. The data content corresponding to the optimal solution is converted into a natural language description and output using a large language model.

[0082] It should be noted that, in this embodiment, basic data is first acquired based on the MaaS platform to construct a descriptive text. The content of the descriptive text includes at least the unit transportation cost of each segment, the transportation time of each segment, the transportation demand quantity of each type of cargo container, the optimization objective, constraints, the origin / end point and transfer node information of the travel path. Then, the descriptive text is parsed and transformed using a preset large language model to obtain multi-dimensional decision parameters. Correspondingly, the multi-dimensional decision parameters include the structured data corresponding to the descriptive text. After obtaining the multi-dimensional decision parameters of the structured data, the multi-dimensional decision parameters need to be logically verified based on a preset topology to obtain the target parameters. The target parameters include the standardized calculation parameters required for the travel path. Finally, the target parameters are globally optimized based on a preset target collaborative constraint model to obtain the optimal solution. The data content corresponding to the optimal solution is converted into a natural language description and output using the large language model.

[0083] Furthermore, in one embodiment of the invention, the step of acquiring basic data based on the MaaS platform to construct the descriptive text specifically includes:

[0084] Basic data is acquired based on the MaaS platform, including transportation information data and user requirements stored in tabular form.

[0085] The description text is constructed by converting information data, user needs, transportation costs, transportation time, demand quantity, optimization objectives, constraints, and the origin / destination and transfer nodes of the travel route into natural language instructions containing parameters to be parsed using a fixed text template. The formula of the description text is as follows:

[0086] ;

[0087] in, This refers to the descriptive text. Indicates specifications Cargo container on the journey path Transportation between nodes Real-time per-box shipping cost Indicates the route Transportation between nodes The time of delivery This represents the optimization objective, which includes a time objective and / or a cost objective. This represents the constraints, which include time constraints and / or cost constraints. Indicates the starting point of the journey. Indicates the end point of the journey. This indicates the total demand for containers. This represents a set of container specifications. A set of nodes representing the travel path.

[0088] It should be noted that, in this embodiment, as Figure 2 As shown, this step aims to retrieve corresponding basic data based on the MaaS platform to construct a standardized prompt framework with domain knowledge guidance capabilities as descriptive text. Specifically, it retrieves multimodal transport information data and user demands from relevant regions, sets time or cost constraints and optimization goals, and uses a fixed text template to transform the information data, user demands, transport costs, transport time, demand quantity, optimization goals, constraints, origin / destination and transfer nodes of the travel route into natural language instructions containing parameters to be parsed, thus constructing the descriptive text. The formula for the descriptive text is as follows:

[0089] ;

[0090] in, This refers to the descriptive text. Indicates specifications Cargo container on the journey path Transportation between nodes Real-time per-box shipping cost Indicates the route Transportation between nodes The time of delivery This represents the optimization objective, which includes a time objective and / or a cost objective. This represents the constraints, which include time constraints and / or cost constraints. Indicates the starting point of the journey. Indicates the end point of the journey. This indicates the total demand for containers. Represents a set of container specifications, for example , A set of nodes representing the travel path.

[0091] Accordingly, with reference to Tables 1 and 2, examples are provided to visually illustrate this transformation process.

[0092] Table 1. Example of Transportation Information Data

[0093] starting point end Route 1 ··· Transportation methods 20GP freight rate (RMB) 45HQ Freight Rate (RMB) Transportation time (hours) Ma'anshan Zhenjiang Cihu Station ··· railway 448.8 561.0 2.7 Ma'anshan Zhenjiang Nanjing Port ··· waterway 180.5 326.4 5.4 Zhenjiang Changzhou Danyang Station ··· railway 351.8 483.0 1.6 Zhenjiang Changzhou Yangzhou Port ··· waterway 154.2 278.8 6.5

[0094] Table 2. Example of User Requirements

[0095] starting point end 20GP quantity … 45HQ quantity Ma'anshan Changzhou 3 ··· 3

[0096] Furthermore, in this embodiment, the corresponding natural language description is, for example:

[0097] "There is a multimodal transport logistics demand from Ma'anshan via Zhenjiang to Changzhou, requiring the completion of a specified number of container transport tasks. Four types of containers need to be used simultaneously: three 20GP containers, two 40GP containers, four 40HQ containers, and three 45HQ containers. The maximum cost constraint for this transport is set at 9,000 yuan. The core requirement is to minimize the total transport time while strictly adhering to the requirement that the total transport cost does not exceed 9,000 yuan. This transport task consists of two segments. The first segment is from Ma'anshan to Zhenjiang, with two transport modes available: rail and waterway. The rail transport originates at Ma'anshan Station and terminates at Zhenjiang Station, passing through Cihu Station, Tongjing Station, Jianning Station, Cangbomen Station, Longtan Station, and Gaozi Station. The transport time is 2.7 hours, and the corresponding freight rates for the different containers are 448.8 yuan for 20GP, 534.3 yuan for 40GP, 534.3 yuan for 40HQ, and 45HQ." The first leg is from Zhenjiang to Changzhou. The transportation options are rail and waterway. Rail transport starts at Ma'anshan Port and ends at Zhenjiang Port, passing through Nanjing Port. The transport time is 5.4 hours, and the freight rates for different containers are 180.5 yuan for 20GP, 310.8 yuan for 40GP, 310.8 yuan for 40HQ, and 326.4 yuan for 45HQ. The second leg is from Zhenjiang to Changzhou. Rail transport starts at Zhenjiang Station and ends at Changzhou East Station, passing through Danyang Station. The transport time is 1.6 hours, and the freight rates for different containers are 351.8 yuan for 20GP, 460.0 yuan for 40GP, 460.0 yuan for 40HQ, and 483.0 yuan for 45HQ. Waterway transport starts at Zhenjiang Port and ends at Changzhou Port, passing through Yangzhou Port and Taizhou Port. The transport time is 6.5 hours, and the freight rates for different containers are as follows: 20GP 180.5 yuan, 40GP 310.8 yuan, 40HQ 310.8 yuan, and 45HQ 326.4 yuan. The costs are: 154.2 yuan for 40GP, 265.5 yuan for 40HQ, and 278.8 yuan for 45HQ. Based on the known transportation methods, transit points, transit times, and freight rates, and considering both cost constraints and the core requirement of minimizing transit time, the optimal transportation combination for completing this container transport task needs to be selected.

[0098] Furthermore, in one embodiment of the invention, the step of parsing and transforming the descriptive text using a preset large language model to obtain multidimensional decision parameters specifically includes:

[0099] The large language model is used to convert the descriptive text into standardized JSON structured data using a custom prompt word template;

[0100] The multidimensional decision parameters are obtained by parsing the corresponding JSON structured data and extracting feature parameters, as shown in the following formula:

[0101] ;

[0102] in, Represents multidimensional decision parameters. Indicates the starting point of the journey. Indicates the end point of the journey. This indicates the total demand for containers. This represents the set of nodes along the travel path. Represents a set of modes of transportation. Indicates specifications Cargo container on the journey path Transportation between nodes Real-time per-box shipping cost Indicates the optimization objective. Indicates time constraints. This indicates a cost constraint.

[0103] It should be noted that, in this embodiment, as Figure 3 As shown, the semantic parsing capabilities of the Large Language Model (LLM) are used to transform natural language instructions into standardized JSON structured data, and multi-dimensional decision parameters are extracted accordingly. Specifically, the LLM uses pre-defined prompt word templates to map the corresponding descriptive text into standardized JSON structured data. Examples of these prompt word templates include:

[0104] "You are a professional logistics data analysis expert, skilled at extracting structured information from complex descriptions of transportation demand."

[0105] # Task

[0106] Please read the provided shipping requirements text, extract the key information, and convert it into JSON format.

[0107] # Schema Field Details

[0108] - start_point (String): The name of the starting city for the transport.

[0109] - end_point (String): The name of the destination city for the transport.

[0110] - target_mode (String): Optimization target. If the text emphasizes "shortest duration" or "lowest cost under time constraints", please enter "time"; if it emphasizes "lowest cost", please enter "cost".

[0111] - time_limit (Float / Integer): The maximum requirement for the total transportation time (in hours). Enter 0 if there is no specific requirement.

[0112] - cost_limit (Float / Integer): The upper limit of the total transportation cost (in yuan). Enter 0 if there is no explicit requirement.

[0113] - 20GP_Num (Integer): The number of 20GP type cargo boxes.

[0114] - 40GP_Num (Integer): The number of 40GP type cargo boxes.

[0115] - 40HQ_Num (Integer): The number of 40HQ type cargo boxes.

[0116] - 45HQ_Num (Integer): The number of 45HQ type cargo boxes.

[0117] - transport_segments (List): A list containing each transport segment.

[0118] - segment_start (String): The starting point of the road segment.

[0119] - segment_end (String): The end point of the road segment.

[0120] - transport_options (List): The available transport options for this route.

[0121] - transport_mode (String): The name of the mode of transport (e.g., railway, waterway, highway).

[0122] - passing_stations (List of Strings): A list of all stations / ports along the route (excluding segment_start and segment_end).

[0123] - freight_rate_RMB (Object): Unit price for each container size. Includes subfields: 20GP, 40GP, 40HQ, 45HQ (all Float type).

[0124] - transport_hours (Float): The required transport time for this mode of transport on this route.

[0125] # Constraint

[0126] 1. Strictly adhere to the JSON format for output, and do not include any explanatory text.

[0127] 2. Do not include units in numeric fields.

[0128] 3. If a parameter is not mentioned in the description, please enter null or 0 according to its type.

[0129] Furthermore, in this embodiment, the multidimensional decision parameters are obtained by parsing the corresponding JSON structured data and extracting feature parameters, as shown in the following formula:

[0130] ;

[0131] in, Represents multidimensional decision parameters. Indicates the starting point of the journey. Indicates the end point of the journey. This indicates the total demand for containers. This represents the set of nodes along the travel path. Represents a set of modes of transportation. Indicates specifications Cargo container on the journey path Transportation between nodes Real-time per-box shipping cost Indicates the optimization objective. Indicates time constraints. This indicates a cost constraint.

[0132] Furthermore, in one embodiment of the invention, the step of performing logical verification on the multidimensional decision parameters based on a preset topology to obtain the target parameters specifically includes:

[0133] Path node information is verified to determine whether each path is connected, and path-mode decision variables are constructed based on the directed graph topology of the multimodal transport network, as shown in the following formula:

[0134] ;

[0135] in, Indicates the route Transportation between nodes The path-transportation mode decision vector, Indicates the first Various path planning schemes This represents the set of nodes along the travel path. Represents a set of modes of transportation;

[0136] The total cost of a single arc segment is defined as the sum of the freight costs for all specifications of goods, as shown in the following formula:

[0137] ;

[0138] in, This represents the sum of freight costs for goods of all specifications and routes. This indicates the number of containers of specification 's'. Indicates specifications Cargo container on the journey path Transportation between nodes Real-time per-box shipping cost Indicates the route Transportation between nodes The path-transportation mode decision vector, This indicates the total demand for containers. This represents a set of container specifications. Indicates the first Various path planning schemes;

[0139] The formula for calculating the route transportation time is as follows:

[0140] ;

[0141] in, Indicates the transit time along the route. Indicates the route Transportation between nodes The time of delivery Indicates the route Transportation between nodes The path-transportation mode decision vector, Indicates the first Various path planning schemes;

[0142] Based on the sum of freight costs for all specifications of goods and route transportation time Obtain the target parameters.

[0143] It should be noted that, in this embodiment, after the parameter parsing is completed, the extracted parameters need to be verified for path node information, mainly to determine whether each path is connected. Then, based on the directed graph topology of the multimodal transport network, the corresponding path-transportation mode decision variables are constructed. ,in, ,and, Indicates the route Transportation between nodes The path-transportation mode decision vector, Indicates the first Various path planning schemes This represents the set of nodes along the travel path. This represents a set of transportation methods.

[0144] Furthermore, in this embodiment, the target parameter includes the sum of freight costs for each type of cargo. and route transportation time The target parameters include standardized calculation parameters required by the route planning engine, specifically, the sum of freight costs for each type of cargo. ,in, This represents the sum of freight costs for goods of all specifications and routes. This indicates the number of containers of specification 's'. This indicates the number of containers of specification 's'. Indicates the route Transportation between nodes The path-transportation mode decision vector, This indicates the total demand for containers. This represents a set of container specifications. Indicates the first Various route planning schemes; and route transportation time ,in, Indicates the transit time along the route. Indicates the route Transportation between nodes The time of delivery Indicates the route Transportation between nodes The path-transportation mode decision vector, Indicates the first A path planning scheme.

[0145] Furthermore, in one embodiment of the invention, the step of globally optimizing the target parameters based on a preset target collaborative constraint model to obtain the optimal solution specifically includes:

[0146] The optimal solution is obtained by finding the optimal path that simultaneously satisfies the optimization objective and constraints using a preset solution tool.

[0147] The optimization objective includes minimizing the dynamic selection cost, calculated as follows:

[0148] ;

[0149] in, This represents the sum of freight costs for goods of all specifications and routes. Indicates the nodes of the travel path. Represents a set of nodes. Indicates the mode of transportation. A collection of transportation methods Indicates the first Path planning scheme, or

[0150] The optimization objective includes minimizing the dynamic selection time, calculated as follows:

[0151] ;

[0152] in, Indicates the transit time along the route. Indicates the nodes of the travel path. Represents a set of nodes. Indicates the mode of transportation. A collection of transportation methods Indicates the first Various path planning schemes;

[0153] Constraints include time constraints that are less than or equal to time constraints, calculated as follows:

[0154] ;

[0155] in, Indicates the transit time along the route. Indicates the nodes of the travel path. Represents a set of nodes. Indicates the mode of transportation. A collection of transportation methods Indicates time constraints. Indicates the first Path planning scheme, or

[0156] Constraints include those less than or equal to cost constraints, calculated as follows:

[0157] ;

[0158] in, This represents the sum of freight costs for goods of all specifications and routes. Indicates the nodes of the travel path. Represents a set of nodes. Indicates the mode of transportation. A collection of transportation methods Indicates cost constraints. Indicates the first A path planning scheme.

[0159] It should be noted that, in this embodiment, as Figure 4As shown, different conditions exist during global optimization. In this embodiment, the optimization objective includes dynamically selecting the minimum cost or dynamically selecting the minimum time, while the constraints include time constraints and cost constraints. Specifically, when the minimum cost is selected as the optimization objective, the corresponding constraint is a time constraint; when the minimum time is selected as the optimization objective, the corresponding constraint is a cost constraint. In application, OR-Tools is used as a preset solution tool to find the optimal path that simultaneously satisfies the constraints and minimizes the optimization objective within the feasible region, thereby obtaining the optimal solution.

[0160] Specifically, in this embodiment, based on the requirements of the customized natural language instructions, the option with the lowest cost or time is dynamically selected according to the optimization objective. or ,in, Indicates the transit time along the route. Indicates the transit time along the route. Indicates the nodes of the travel path. Represents a set of nodes. Indicates the mode of transportation. A collection of transportation methods Indicates the first For each path planning scheme, the minimization process is subject to constraints, including time constraints. Or cost constraints ,in, Indicates the transit time along the route. This represents the sum of freight costs for goods of all specifications and routes. Indicates the nodes of the travel path. Represents a set of nodes. Indicates the mode of transportation. A collection of transportation methods Indicates time constraints. Indicates cost constraints. Indicates the first A path planning scheme.

[0161] Furthermore, in one embodiment of the invention, as Figure 5 As shown, the process of converting the data content corresponding to the optimal solution into a natural language description and outputting it using a large language model specifically includes the following steps:

[0162] S502, based on the optimal solution, extract the data content corresponding to the route information, total transportation time, total transportation cost, unit transportation cost of each interval, and transportation time of each interval;

[0163] S504, use a large language model to convert the data content corresponding to the optimal solution into a natural language description, and output the converted natural language description to the outside world.

[0164] It should be noted that, in this embodiment, after OR-Tools solves for the minimum value, the text generation function of the Large Language Model (LLM) is used to extract the path information, total transportation time, total transportation cost, unit transportation cost of each interval, and transportation time of each interval based on the optimal solution. This data is then converted into a natural language description, and the converted natural language description is output externally. An example of the corresponding Prompt word is as follows:

[0165] "You are the logistics route optimization result interpreter. Please directly convert this determined optimal transportation route plan into a coherent, formal, and complete plain text description based on the input JSON data."

[0166] Meaning of JSON fields (all fields must be included)

[0167] 1) start_point: The starting point of the transportation;

[0168] 2) end_point: The final destination of the transportation;

[0169] 3) target_mode: Optimization objective, cost is the most cost-effective, time is the most time-effective;

[0170] 4) 20GP_Num, 40GP_Num, 40HQ_Num, 45HQ_Num: Corresponding container demand quantities;

[0171] 5) time_limit: Total time limit (hours), 0 indicates no limit;

[0172] 6) cost_limit: Total cost limit (yuan), 0 indicates no limit;

[0173] 7) transport_segments: An array of transport segments, which is already the optimal solution in this JSON;

[0174] -segment_start: The starting point of the road segment;

[0175] -segment_end: End point of the road segment;

[0176] The `-transport_options` section lists the available transportation options for this route. The current JSON is already the optimal result, so we can directly use the transportation information provided within it.

[0177] -transport_mode: Transportation mode;

[0178] -passing_stations: Stations along the route (excluding the start and end points of the route);

[0179] -freight_rate_RMB: Unit price for each type of container;

[0180] -transport_hours: Transport time for the route;

[0181] 8) time_all: Total time of the optimal solution multiplied by 1;

[0182] 9) cost_all: Total cost of the optimal solution.

[0183] Furthermore, to evaluate the technical advantages of this application, this embodiment introduces the evaluation metric GPA (Generation Plan Accuracy) for assessment. A comparative experiment is conducted between the method of directly generating the optimal solution based on a large language model and the method proposed in this application. The formula for the GPA metric is defined as follows:

[0184] ;

[0185] in, The number of optimal routes, transportation methods, costs, and time calculations that are consistent with the manually labeled reference solutions is determined. To determine the total number of test samples, we take the custom reference dataset used in this evaluation as an example. Its data source is constructed based on MaaS database information. The dataset contains 15 natural language description instructions to be solved. Each description involves 2-4 candidate transportation routes, and the node information is significantly different. This is intended to fully verify the robustness of the algorithm in various scenarios. The complete dataset is stored in JSON format. Each test case includes: the original natural language description, transportation time and cost information, a set of nodes and transportation methods, and the baseline optimal route information obtained through rigorous manual calculation.

[0186] Specifically, during the verification process, three mainstream large language models—GLM-32B, DeepSeek-R1-Distill-Qwen-32B, and Qwen2.5-32B-Instruct—were used for comparative testing. The experimental results showed that in the direct input instruction mode, the GPA performance fluctuated significantly due to the limited sensitivity of large models to complex numerical calculations and logical constraints. However, by combining the semantic understanding capability of LLM with the computational capability of OR-Tools, the invention achieved a significant improvement in GPA across all three models, verifying the superior robustness and accuracy of the invention in handling complex node information and dynamic requirements. The experimental results are shown in Table 3.

[0187] Table 3. Detailed list of comparative experiments

[0188] LLM model LLM Direct Generation Method (GPA) Four-Phase Approach (GPA) GLM-32B 46.67% 66.67% DeepSeek-R1-Distill-Qwen-32B 53.33% 73.33% Qwen2.5-32B-Instruct 20.00% 66.67%

[0189] The scope of protection of the large language model-driven multimodal transport route planning method for MaaS platforms described in this application is not limited to the execution order of the steps listed in this embodiment. Any solution implemented by adding, subtracting, or replacing steps in the prior art based on the principles of this application is included within the scope of protection of this application.

[0190] This application also provides a large language model-driven multimodal transport route planning system for MaaS platforms. This system can implement the large language model-driven multimodal transport route planning method for MaaS platforms described in this application. However, the implementation apparatus for the large language model-driven multimodal transport route planning method for MaaS platforms described in this application includes, but is not limited to, the structure of the large language model-driven multimodal transport route planning system for MaaS platforms listed in this embodiment. Any structural modifications and substitutions of existing technologies made based on the principles of this application are included within the protection scope of this application.

[0191] Please see Figure 6 In one embodiment, this embodiment provides a large language model-driven multimodal transport route planning system 60 for MaaS platforms, the system comprising:

[0192] Module 61 is used to acquire basic data based on the MaaS platform to construct a description text. The description text includes at least the unit transportation cost of each segment, the transportation time of each segment, the transportation demand quantity of each type of cargo box, the optimization objective, the constraints, the origin / end point of the journey, and the transfer node information.

[0193] Parsing module 62 is used to parse and transform the description text using a preset large language model to obtain multidimensional decision parameters, the multidimensional decision parameters including structured data corresponding to the description text;

[0194] The verification module 63 is used to perform logical verification on the decision parameters based on a preset topology to obtain target parameters, the target parameters including standardized calculation parameters required for the travel path;

[0195] The output module 64 is used to perform global optimization of the target parameters based on a preset target collaborative constraint model to obtain the optimal solution, and to use a large language model to convert the data content corresponding to the optimal solution into a natural language description and output it.

[0196] Since the specific implementation of this embodiment corresponds to the aforementioned method embodiment, the same details will not be repeated here, and those skilled in the art should also understand this. Figure 6 The division of the modules in the embodiments is only a logical functional division. In actual implementation, they can be fully or partially integrated into one or more physical entities. These modules can be fully implemented in software through processing element calls, fully implemented in hardware, or some modules can be implemented in software through processing element calls and some modules can be implemented in hardware.

[0197] It should be understood that the disclosed apparatus or method can be implemented in other ways, given the several embodiments provided in this application. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules or units may be electrical, mechanical, or other forms.

[0198] The modules / units described as separate components may or may not be physically separate. The components shown as modules / units may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules / units can be selected to achieve the objectives of the embodiments of this application, depending on actual needs. For example, the functional modules / units in the various embodiments of this application may be integrated into one processing module, or each module / unit may exist physically separately, or two or more modules / units may be integrated into one module / unit.

[0199] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0200] This application also provides an electronic device, such as... Figure 7 As shown, the electronic device includes a processor and a memory.

[0201] like Figure 7 As shown, the electronic device includes: at least one processor 701, a memory 702, at least one network interface 703, and a user interface 705. The various components in the device are coupled together via a bus system 704. It is understood that the bus system 704 is used to implement communication between these components. In addition to a data bus, the bus system 704 also includes a power bus, a control bus, and a status signal bus. However, for clarity, in... Figure 7 The general will label all buses as bus systems.

[0202] The user interface 705 may include a monitor, keyboard, mouse, trackball, clicker, button, touchpad, or touch screen.

[0203] It is understood that memory 702 can be volatile memory or non-volatile memory, or both. Non-volatile memory can be read-only memory (ROM) or programmable read-only memory (PROM), which serves as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM) and synchronous static random access memory (SSRAM). The memories described in the embodiments of this application are intended to include, but are not limited to, these and any other suitable categories of memory.

[0204] In this embodiment, the memory 702 is used to store various types of data to support the operation of the electronic terminal 700. Examples of this data include any executable program for operation on the electronic terminal 700, such as the operating system 7021 and application programs 7022. The operating system 7021 contains various system programs, such as the framework layer, core library layer, and driver layer, for implementing various basic services and handling hardware-based tasks. The application program 7022 may contain various applications, such as a media player and a browser, for implementing various application services. The methods provided in this embodiment can be included in the application program 7022.

[0205] The methods disclosed in the embodiments of this application can be applied to processor 701, or implemented by processor 701. Processor 701 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in processor 701 or by instructions in the form of software. The processor 701 may be a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Processor 701 can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. Processor 701 may be a microprocessor or any conventional processor, etc. The steps of the accessory optimization method provided in the embodiments of this application can be directly reflected as being executed by a hardware decoding processor, or being executed by a combination of hardware and software modules in the decoding processor. The software module may be located in a storage medium, which is located in memory. The processor reads the information in the memory and combines it with its hardware to complete the steps of the aforementioned method.

[0206] In an exemplary embodiment, the electronic terminal 700 may be used by one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), or complex programmable logic devices (CPLDs) to execute the aforementioned method.

[0207] According to the method provided in the embodiments of this application, this application also provides a computer program product, which includes: computer program code, which, when run on a computer, causes the computer to execute any of the methods in the embodiments of the large language model driven multimodal transport path planning method for MaaS platform described above.

[0208] According to the method provided in the embodiments of this application, this application also provides a computer-readable storage medium storing program code, which, when run on a computer, causes the computer to execute any of the methods in the embodiments of the large language model-driven multimodal transport path planning method for MaaS platforms described above.

[0209] The terms “component,” “module,” “system,” etc., used in this specification are used to refer to computer-related entities, hardware, firmware, combinations of hardware and software, software, or software in execution. For example, a component can be, but is not limited to, a process running on a processor, a processor, an object, an executable file, an execution thread, a program, and / or a computer. As illustrated, applications running on computing devices and computing devices can both be components. One or more components may reside in a process and / or an execution thread, and components may be located on a single computer and / or distributed among two or more computers. Furthermore, these components can be executed from various computer-readable media on which various data structures are stored. Components can communicate, for example, via local and / or remote processes based on signals having one or more data packets (e.g., data from two components interacting with another component between a local system, a distributed system, and / or a network, such as the Internet interacting with other systems via signals).

[0210] Those skilled in the art will recognize that the various illustrative logical blocks and steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0211] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0212] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0213] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0214] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0215] In the above embodiments, the functions of each functional unit can be implemented entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. A computer program product includes one or more computer instructions (programs). When the computer program instructions (programs) are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., high-density digital video discs (DVDs), or semiconductor media (e.g., solid-state disks (SSDs)).

[0216] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0217] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0218] In summary, this application provides a method, system, medium, and electronic device for multimodal transport route planning driven by a large language model for MaaS platforms. This improves the accuracy of handling natural language requirements in multimodal transport route optimization scenarios, realizes intelligent route planning, and exhibits excellent robustness and accuracy when handling complex node information and dynamic requirements.

[0219] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A multimodal transport route planning method driven by a large language model for MaaS platforms, characterized in that, include: Based on the basic data obtained from the MaaS platform, a descriptive text is constructed. The content of the descriptive text includes at least the unit transportation cost of each segment, the transportation time of each segment, the transportation demand quantity of each type of cargo box, the optimization objective, the constraints, the origin / end point of the journey, and the transfer node information. The descriptive text is parsed and transformed using a pre-defined large language model to obtain multidimensional decision parameters, which include structured data corresponding to the descriptive text. The multidimensional decision parameters are logically verified based on a preset topology to obtain target parameters, which include standardized calculation parameters required for the travel path. The optimal solution is obtained by globally optimizing the target parameters based on the preset target collaborative constraint model. The data content corresponding to the optimal solution is then converted into a natural language description and output using a large language model.

2. The method for multimodal transport path planning driven by a large language model for MaaS platforms according to claim 1, characterized in that, The process of acquiring basic data based on the MaaS platform to construct the descriptive text specifically includes: Basic data is acquired based on the MaaS platform, including transportation information data and user requirements stored in tabular form. The description text is constructed by converting information data, user needs, transportation costs, transportation time, demand quantity, optimization objectives, constraints, and the origin / destination and transfer nodes of the travel route into natural language instructions containing parameters to be parsed using a fixed text template. The formula of the description text is as follows: ； in, This refers to the descriptive text. Indicates specifications Cargo container on the journey path Transportation between nodes Real-time per-box shipping cost Indicates the route Transportation between nodes The time of delivery This represents the optimization objective, which includes a time objective and / or a cost objective. This represents the constraints, which include time constraints and / or cost constraints. Indicates the starting point of the journey. Indicates the end point of the journey. This indicates the total demand for containers. This represents a set of container specifications. A set of nodes representing the travel path.

3. The method for multimodal transport path planning driven by a large language model for MaaS platforms according to claim 1, characterized in that, The step of parsing and transforming the descriptive text using a pre-defined large language model to obtain multi-dimensional decision parameters specifically includes: The large language model is used to convert the descriptive text into standardized JSON structured data using a custom prompt word template; The multidimensional decision parameters are obtained by parsing the corresponding JSON structured data and extracting feature parameters, as shown in the following formula: ； in, Represents multidimensional decision parameters. Indicates the starting point of the journey. Indicates the end point of the journey. This indicates the total demand for containers. This represents the set of nodes along the travel path. Represents a set of modes of transportation. Indicates specifications Cargo container on the journey path Transportation between nodes Real-time per-box shipping cost Indicates the optimization objective. Indicates time constraints. This indicates a cost constraint.

4. The method for multimodal transport path planning driven by a large language model for MaaS platforms according to claim 1, characterized in that, The step of performing logical verification on the multidimensional decision parameters based on a preset topology to obtain the target parameters specifically includes: Path node information is verified to determine whether each path is connected, and path-mode decision variables are constructed based on the directed graph topology of the multimodal transport network, as shown in the following formula: ； in, Indicates the route Transportation between nodes The path-transportation mode decision vector, Indicates the first Various path planning schemes This represents the set of nodes along the travel path. Represents a set of modes of transportation; The total cost of a single arc segment is defined as the sum of the freight costs for all specifications of goods, as shown in the following formula: s ； in, This represents the sum of freight costs for goods of all specifications and routes. This indicates the number of containers of specification 's'. Indicates specifications Cargo container on the journey path Transportation between nodes Real-time per-box shipping cost Indicates the route Transportation between nodes The path-transportation mode decision vector, This indicates the total demand for containers. This represents a set of container specifications. Indicates the first Various path planning schemes; The formula for calculating the route transportation time is as follows: ； in, Indicates the transit time along the route. Indicates the route Transportation between nodes The time of delivery Indicates the route Transportation between nodes The path-transportation mode decision vector, Indicates the first Various path planning schemes; Based on the sum of freight costs for all specifications of goods and route transportation time Obtain the target parameters.

5. The method for multimodal transport path planning driven by a large language model for MaaS platforms according to claim 1, characterized in that, The process of obtaining the optimal solution by globally optimizing the target parameters based on a preset target collaborative constraint model specifically includes: The optimal solution is obtained by finding the optimal path that simultaneously satisfies the optimization objective and constraints using a preset solution tool. The optimization objective includes minimizing the dynamic selection cost, calculated as follows: ； in, This represents the sum of freight costs for goods of all specifications and routes. Indicates the nodes of the travel path. Represents a set of nodes. Indicates the mode of transportation. A collection of transportation methods Indicates the first Path planning scheme, or The optimization objective includes minimizing the dynamic selection time, calculated as follows: ； in, Indicates the transit time along the route. Indicates the nodes of the travel path. Represents a set of nodes. Indicates the mode of transportation. A collection of transportation methods Indicates the first Various path planning schemes; Constraints include time constraints that are less than or equal to time constraints, calculated as follows: ； in, Indicates the transit time along the route. Indicates the nodes of the travel path. Represents a set of nodes. Indicates the mode of transportation. A collection of transportation methods Indicates time constraints. Indicates the first Path planning scheme, or Constraints include those less than or equal to cost constraints, calculated as follows: ； in, This represents the sum of freight costs for goods of all specifications and routes. Indicates the nodes of the travel path. Represents a set of nodes. Indicates the mode of transportation. A collection of transportation methods Indicates cost constraints. Indicates the first A path planning scheme.

6. The method for multimodal transport path planning driven by a large language model for MaaS platforms according to claim 1, characterized in that, The step of using a large language model to convert the data content corresponding to the optimal solution into a natural language description and output it specifically includes: Based on the optimal solution, the following data are extracted: route information, total transportation time, total transportation cost, unit transportation cost of each interval, and transportation time of each interval. The data content corresponding to the optimal solution is converted into a natural language description using a large language model, and the converted natural language description is then output externally.

7. A multimodal transport route planning system driven by a large language model for MaaS platforms, characterized in that, include: The construction module is used to acquire basic data based on the MaaS platform to construct the description text. The content of the description text includes at least the unit transportation cost of each segment, the transportation time of each segment, the transportation demand quantity of each type of cargo box, the optimization objective, the constraints, the origin / end point of the journey, and the transfer node information. The parsing module is used to parse and transform the description text using a preset large language model to obtain multidimensional decision parameters, the multidimensional decision parameters including the structured data corresponding to the description text; The verification module is used to perform logical verification on the decision parameters based on a preset topology to obtain target parameters, the target parameters including standardized calculation parameters required for the travel path; The output module is used to perform global optimization of the target parameters based on a preset target collaborative constraint model to obtain the optimal solution, and to use a large language model to convert the data content corresponding to the optimal solution into a natural language description and output it.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the large language model-driven multimodal transport path planning method for MaaS platforms as described in any one of claims 1 to 6.

9. A computer program product, characterized in that, The computer program product includes computer program code, which, when run on a computer, enables the computer to implement the large language model-driven multimodal transport path planning method for MaaS platforms as described in any one of claims 1 to 6.

10. An electronic device, characterized in that, The electronic device includes a processor and a memory; wherein the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the electronic device executes the large language model-driven multimodal transport path planning method for MaaS platform as described in any one of claims 1 to 6.

Images