Multi-Agent tourism plan generation method and device based on large language model

Abstract

The invention provides a multi-Agent tourism plan generation method and device based on a large language model, and relates to the technical field of computers, and the method comprises the steps: responding to a query request of a client, analyzing the query request, and extracting request features from the query request; the request features comprise a travel destination, travel time and limiting conditions; inquiring scenic spot information related to the request feature from a vector database, or collecting the scenic spot information related to the request feature on the Internet by using an external tool; a large amount of scenic spot data existing in a vector form is stored in the vector database; and based on the scenic spot information and the limiting conditions, calling an external query tool to obtain and aggregate various related information, and generating target information. Through interaction between the Agent and the outside, the user can directly input the instruction to obtain a complete travel plan, the efficiency of travel strategy generation is optimized, and the travel experience of the user is improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a multi-agent tourism planning generation method and device based on a large language model. Background Art

[0002] The rapid development of tourism has brought a series of problems to the city's tourism system, especially the tourism transportation system. Among them, the more prominent problem is that the excessive number of tourists during the peak tourist season leads to congestion in scenic spots, which is extremely disadvantageous to the management of scenic spots and has a great negative impact on tourists' travel experience. When traveling, tourists will choose tourist destinations and determine itineraries based on personal preferences, travel time budget, etc. These travel behavior characteristics of tourists ultimately present a complex spatial distribution. It is difficult to analyze the laws of these distributions using traditional mathematical statistical methods. Accurately grasping the laws of tourists' travel behavior is of great significance for formulating tourism transportation plans, improving tourism quality, and developing the tourism industry. In the existing technology, artificial intelligence, especially large language models, has made significant progress in understanding natural language and generating text content, providing new technical means to solve the above problems. However, how to effectively integrate various data sources and automatically generate travel guides through AI Agents is still a technical challenge. Summary of the invention

[0003] The purpose of the present invention is to solve the current problem of difficulty in effectively integrating various data sources and automatically generating travel guides. The present invention provides a multi-agent travel planning generation method and device based on a large language model.

[0004] The technical solution of the embodiment of the present application is implemented as follows: The first aspect of the embodiment of the present application provides a multi-agent tourism planning generation method based on a large language model, which is applied to an agent system, including: In response to a query request from a client, analyzing the query request and extracting request features from the query request; the request features include a travel destination, travel time, and restriction conditions; Querying scenic spot information related to the requested feature from a vector database, or using an external tool to collect scenic spot information related to the requested feature on the Internet; the vector database stores a large amount of tourist attraction data in the form of vectors; Based on the scenic spot information and the restriction conditions, an external query tool is called to obtain and aggregate various related information to generate target information; The target information is transmitted to the client.

[0005] Optionally, the responding to the query request from the client, analyzing the query request, and extracting request features from the query request includes: The query request is analyzed using thought chain technology to extract request features from the query request.

[0006] Optionally, the scenic spot information related to the request feature includes weather and route; the querying of the scenic spot information related to the request feature from the vector database, or collecting the scenic spot information related to the request feature on the Internet using an external tool, includes: Querying the scenic spot information related to the requested feature from the vector database; If the scenic spot information related to the requested feature is not found in the vector database, the scenic spot information related to the requested feature is collected on the Internet using an external tool.

[0007] Optionally, based on the scenic spot information and the restriction condition, calling an external query tool to obtain and aggregate a variety of related information to generate target information includes: The multiple related information are displayed in a preset order to generate at least one target information.

[0008] The second aspect of the embodiment of the present application provides a multi-agent tourism planning generation device based on a large language model, including: a feature extraction module, a query collection module, a call generation module and an information transmission module, wherein: A feature extraction module is configured to respond to a query request from a client, analyze the query request, and extract request features from the query request; the request features include a travel destination, travel time, and restriction conditions; A query collection module, configured to query the scenic spot information related to the request feature from a vector database, or to collect the scenic spot information related to the request feature on the Internet using an external tool; the vector database stores a large amount of tourist attraction data in the form of vectors; A calling generation module is configured to call an external query tool to obtain and aggregate a variety of related information based on the scenic spot information and the restriction conditions to generate target information; The information transmission module is configured to transmit the target information to the client.

[0009] Optionally, the feature extraction module is specifically configured as follows: The query request is analyzed using thought chain technology to extract request features from the query request.

[0010] Optionally, the scenic spot information related to the request feature includes weather and route; the query collection module is specifically configured as follows: Querying the scenic spot information related to the requested feature from the vector database; If the scenic spot information related to the requested feature is not found in the vector database, the scenic spot information related to the requested feature is collected on the Internet using an external tool.

[0011] Optionally, the call generation module is specifically configured as follows: The multiple related information are displayed in a preset order to generate at least one target information.

[0012] A third aspect of an embodiment of the present application provides an electronic device, including a processor and a memory; the memory stores a computer program, wherein the computer program, when executed by the processor, implements the multi-agent tourism planning generation method based on a large language model described in the first aspect.

[0013] A fourth aspect of an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in the first aspect.

[0014] Compared with the prior art, the technical solution provided by this application has the following beneficial effects: The present invention provides a multi-agent travel planning generation method and device based on a large language model, which analyzes the query request in response to the query request of the client, extracts the request feature from the query request, queries the scenic spot information related to the request feature from the vector database, or uses an external tool to collect the scenic spot information related to the request feature on the Internet, and based on the scenic spot information and restriction conditions, calls the external query tool to obtain and aggregate various related information to generate target information. Through the interaction between the agent and the outside, the user can directly input the command to obtain a complete travel plan, optimizes the efficiency of generating travel strategies, and improves the user's travel experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A flowchart of a multi-agent travel planning generation method based on a large language model provided in an embodiment of the present application; Figure 2 A schematic diagram of the feature extraction principle provided in the embodiment of the present application; Figure 3 A schematic diagram of the Agent information interaction principle provided in the embodiment of the present application; Figure 4 A schematic diagram of the structure of a multi-agent travel planning generation device based on a large language model provided in an embodiment of the present application; Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0016] Below, embodiments of the present application will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present application. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present application. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present application. The terms used herein are only for describing specific embodiments and are not intended to limit the present application. The terms "include", "comprising", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components. All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0017] Some block diagrams and / or flow charts are shown in the accompanying drawings. It should be understood that some blocks or combinations thereof in the block diagrams and / or flow charts may be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that these instructions, when executed by the processor, may create a device for implementing the functions / operations described in these block diagrams and / or flow charts.

[0018] In some embodiments, see Figure 1 , Figure 1 A flowchart of a multi-agent tourism planning generation method based on a large language model provided in an embodiment of the present application; the multi-agent tourism planning generation method based on a large language model provided in an embodiment of the present application is applied to an Agent system, including: S110, in response to the query request of the client, analyzing the query request and extracting request features from the query request; the request features include travel destination, travel time and restriction conditions.

[0019] Here, the server needs to receive the travel query request submitted by the client. This request can be transmitted in text form through protocols such as HTTP and HTTPS, or it may be transmitted in JSON or other structured data formats through an API interface. The query request is analyzed by extracting keywords from the query request or associating and inferring based on the contextual semantics of the keywords. The travel destination is the place where the user wants to go, which can be a specific city, a country, or a specific tourist area. For example, "Paris", "Japan" or "Hawaii". The travel time includes the dates when the user plans to start and end the travel, as well as the possible duration. For example, the user may plan to travel between January 1, 2025 and January 7, 2025. Restrictions may include a variety of factors, such as budget, transportation mode, accommodation type, attraction preferences, etc. These restrictions will help further narrow the search scope and provide travel information that better meets the needs of users. For example, budget: The user may have a specific budget range to pay for the entire travel itinerary. Restrictions may include budget, transportation mode, accommodation type, attraction preferences. Budget, that is, the user may have a specific budget range to pay for the entire travel itinerary. Transportation mode, i.e., users may want to use a specific mode of transportation, such as airplane, train, car, etc. Accommodation type, i.e., users may prefer a certain type of accommodation, such as hotel, B&B, youth hostel, etc. Attraction preference, i.e., users may have a special interest in certain types of attractions, such as historical sites, natural scenery, theme parks, etc.

[0020] In some embodiments, S110, in response to a query request from a client, analyzing the query request and extracting request features from the query request includes: The query request is analyzed using the thought chain technology, and the request features are extracted from the query request.

[0021] Thinking chain refers to breaking down a logically complex problem into a complete thinking process through a series of logically related thoughts. By building a logical reasoning framework, complex problems are broken down into multiple sub-problems, and these sub-problems are solved one by one to finally get a complete answer. Here, using thinking chain technology to analyze query requests can improve the accuracy of extracting request features.

[0022] In an example, see Figure 2 , Figure 2Schematic diagram of the feature extraction principle provided for the embodiment of the present application; wherein the question corresponds to the query request of the client, the standard prompt is one of the functional requirements of the Agent system, and the thought chain prompt corresponds to the analysis and processing logic of the query request. A variety of request features can be extracted from the query request, such as the travel destination "Wuhan", the travel time "November 6 to November 10", and the restriction conditions "the travel budget is 6,000 yuan, take the high-speed rail, the living environment has convenient transportation, and it is not too far from the scenic spot".

[0023] S120, querying the scenic spot information related to the requested feature from the vector database, or using external tools to collect the scenic spot information related to the requested feature on the Internet; the vector database stores a large amount of tourist attraction data in the form of vectors.

[0024] In the stage of collecting scenic spot information, the Agent system can collect popular tourist text information on social platforms through external tools such as crawlers, filter and analyze these text information, and store the tourist-related information contained therein in the form of vectors in the vector database. This is a process that can be used as a partial increment. The vector database uses a vectorized query execution engine that can process multiple data at a time, which can significantly reduce the complexity of calculations and improve processing speed. This efficient data processing method enables the vector database to respond quickly when processing big data, meeting the needs of the big data era. In addition, the vector database also supports efficient parallel processing, which can make full use of the parallel computing capabilities of multi-core CPUs or GPUs to further improve the speed of data processing.

[0025] In some embodiments, the scenic spot information related to the requested feature includes weather and route; S120, querying the scenic spot information related to the requested feature from the vector database, or using an external tool to collect the scenic spot information related to the requested feature on the Internet, including: Querying scenic spot information related to the requested feature from the vector database; If the scenic spot information related to the requested feature is not found in the vector database, the scenic spot information related to the requested feature is collected on the Internet using an external tool.

[0026] In the scenic spot query stage, the vector database is first searched for scenic spot information related to the requested features. If a scenic spot related to the user's travel destination is found in the vector database, the relevant information of the scenic spot is input into the relevant module of the Agent system; otherwise, external tools are used to collect relevant information on the Internet for analysis and storage.

[0027] S130, based on the scenic spot information and restriction conditions, call an external query tool to obtain and aggregate various related information to generate target information.

[0028] In this embodiment, in addition to considering the travel destination, travel time, budget, mode of transportation, accommodation type, attraction preference, etc. mentioned in the above embodiments, weather information, user reviews, etc. can also be comprehensively considered, and based on this, target information that better meets user expectations is generated.

[0029] In some embodiments, S130, based on the scenic spot information and the restriction conditions, an external query tool is called to obtain and aggregate various related information to generate target information, including: Display various related information in a preset order to generate at least one target information.

[0030] Here, the queried target information is organized in a natural, fluent, logical, and organized natural language to generate at least one target information that conforms to the user's reading habits. When there are multiple target information, each request feature can be weighted and scored, and the target information can be displayed in descending order.

[0031] S140, transmitting the target information to the client.

[0032] In this embodiment, if the target information meets the user's needs, the target information is successfully output; if the target information does not meet the user's needs, the process returns to S110 and the target information is re-determined according to the needs.

[0033] In an alternative embodiment, see Figure 3 , Figure 3Schematic diagram of the agent information interaction principle provided for the embodiment of the present application; the agent is based on a large language model. When the large language model receives a user instruction, it issues an instruction to the agent. The agent starts to act on the task contained in the instruction, and then the agent begins to interact with the external environment. While interacting, the agent will also get the results obtained when taking action. At this time, these results will be fed back to the agent. The agent generates a description of these feedbacks for the large language model, and the large language model then performs the next instruction based on the description. This is a cyclic process until the user is satisfied with the result. The large model agent can be composed of multiple key parts, which are responsible for task decomposition and strategy evaluation, information storage and recall, environmental perception and decision assistance, and converting thinking into actual action. Its specific functions may include: 1. Using the CoT reasoning model to realize the decomposition of complex tasks into executable subtasks and evaluate the execution strategy. 2. Storage and recall, including short-term memory and long-term memory. Short-term memory is used to store the context of the conversation and support multiple rounds of dialogue; long-term memory stores user features, business data, etc., and the storage method usually selects a vector database. 3. Perceive the environment, execute decisions, and expand the capabilities of the Agent through external tools such as API calls and plug-in extensions. For example, ChatPDF parses documents, etc. 4. Transform output, such as intelligent customer service answers, query weather forecasts, etc.

[0034] The embodiment of the present invention analyzes the query request by responding to the query request of the client, extracts the request feature from the query request, queries the scenic spot information related to the request feature from the vector database, or uses external tools to collect the scenic spot information related to the request feature on the Internet, and based on the scenic spot information and restriction conditions, calls the external query tool to obtain and aggregate various related information to generate target information. Through the interaction between the Agent and the outside, the user can directly input the command to get a complete travel plan, optimizes the efficiency of generating travel guides, and improves the user's travel experience.

[0035] In some embodiments, see Figure 4 , Figure 4 A schematic diagram of the structure of a multi-agent tourism planning generation device based on a large language model provided in an embodiment of the present application; an embodiment of the present application provides a multi-agent tourism planning generation device 400 based on a large language model, including: a feature extraction module 410, a query collection module 420, a call generation module 430 and an information transmission module 440, wherein: The feature extraction module 410 is configured to respond to the query request of the client, analyze the query request, and extract request features from the query request; the request features include travel destinations, travel time and restrictions; The query collection module 420 is configured to query the scenic spot information related to the requested feature from the vector database, or collect the scenic spot information related to the requested feature on the Internet using an external tool; the vector database stores a large amount of tourist attraction data in the form of vectors; A calling generation module 430 is configured to call an external query tool to obtain and aggregate various related information based on the scenic spot information and restriction conditions to generate target information; The information transmission module 440 is configured to transmit the target information to the client.

[0036] In some embodiments, the feature extraction module 410 is specifically configured as follows: The query request is analyzed using the thought chain technology, and the request features are extracted from the query request.

[0037] In some embodiments, the scenic spot information related to the request feature includes weather and route; the query collection module 420 is specifically configured as follows: Querying scenic spot information related to the requested feature from the vector database; If the scenic spot information related to the requested feature is not found in the vector database, the scenic spot information related to the requested feature is collected on the Internet using an external tool.

[0038] In some embodiments, the call generation module 430 is specifically configured as follows: Display various related information in a preset order to generate at least one target information.

[0039] The multi-agent travel planning generation device based on a large language model provided in the embodiment of the present application can implement each process in the embodiment corresponding to the multi-agent travel planning generation method based on a large language model, and will not be described again here to avoid repetition.

[0040] It should be noted that the multi-agent tourism planning generation device based on a large language model provided in the embodiment of the present application and the multi-agent tourism planning generation method based on a large language model provided in the embodiment of the present application are based on the same application concept. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned multi-agent tourism planning generation method based on a large language model, and the repeated parts will not be repeated.

[0041] In some embodiments, see Figure 5 , Figure 5 The electronic device 500 provided in the embodiment of the present application includes a processor 510 and a memory 520; the memory 520 stores a computer program, wherein the computer program implements the above-mentioned multi-agent travel planning generation method based on a large language model when executed by the processor.

[0042] Specifically, the processor 510 may include, for example, a general-purpose microprocessor, an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 510 may also include an onboard memory for cache purposes. The processor 510 may be a single processing unit or multiple processing units for executing different actions of the method flow according to the embodiment of the present application.

[0043] The memory 520 may be any medium capable of containing, storing, conveying, propagating or transmitting instructions. For example, the memory 520 may include, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared or semiconductor system, device, component or propagation medium. Specific examples of the memory 520 include: a magnetic storage device, such as a magnetic tape or a hard disk (HDD); an optical storage device, such as a compact disk (CD-ROM); a random access memory (RAM) or a flash memory; and / or a wired / wireless communication link. The present application also provides a computer-readable medium on which a computer program is stored, and when the program is executed by a processor, the above-mentioned multi-agent tourism planning generation method based on a large language model is implemented. The computer-readable medium may be included in the device / apparatus / system described in the above-mentioned embodiment; or it may exist independently without being assembled into the device / apparatus / system. The above-mentioned computer-readable medium carries one or more programs, and when the above-mentioned one or more programs are executed, the method according to the embodiment of the present application is implemented.

[0044] According to an embodiment of the present application, a computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, wherein a computer-readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, optical cable, radio frequency signal, etc., or any suitable combination of the above.

[0045] Those skilled in the art will appreciate that the features described in the various embodiments and / or claims of the present application may be combined and / or combined in a variety of ways, even if such combinations or combinations are not explicitly described in the present application. In particular, without departing from the spirit and teachings of the present application, the features described in the various embodiments and / or claims of the present application may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present application. Therefore, the scope of the present application should not be limited to the above-described embodiments, but should be determined not only by the appended claims, but also by the equivalents of the appended claims.

Claims

1. A multi-agent travel planning generation method based on a large language model, applied to an agent system, characterized in that: include: In response to a query request from a client, analyzing the query request and extracting request features from the query request; The request features include travel destination, travel time and restrictions; Querying scenic spot information related to the requested feature from a vector database, or using an external tool to collect scenic spot information related to the requested feature on the Internet; the vector database stores a large amount of tourist attraction data in the form of vectors; Based on the scenic spot information and the restriction conditions, an external query tool is called to obtain and aggregate various related information to generate target information; The target information is transmitted to the client.

2. The multi-agent travel planning generation method based on a large language model according to claim 1 is characterized in that: The step of analyzing the query request in response to the query request of the client and extracting request features from the query request includes: The query request is analyzed using thought chain technology to extract request features from the query request.

3. The multi-agent travel planning generation method based on a large language model according to claim 1 is characterized in that: The scenic spot information related to the request feature includes weather and route; the querying of the scenic spot information related to the request feature from the vector database, or the collection of the scenic spot information related to the request feature on the Internet using an external tool, includes: Querying the scenic spot information related to the requested feature from the vector database; If the scenic spot information related to the requested feature is not found in the vector database, the scenic spot information related to the requested feature is collected on the Internet using an external tool.

4. The multi-agent travel planning generation method based on a large language model according to claim 1 is characterized in that: Based on the scenic spot information and the restriction conditions, calling an external query tool to obtain and aggregate various related information to generate target information includes: The multiple related information are displayed in a preset order to generate at least one target information.

5. A multi-agent travel planning generation device based on a large language model, characterized in that: include: Feature extraction module, query collection module, call generation module and information transmission module, among which, A feature extraction module is configured to respond to a query request from a client, analyze the query request, and extract request features from the query request; the request features include a travel destination, travel time, and restriction conditions; A query collection module, configured to query the scenic spot information related to the request feature from a vector database, or to collect the scenic spot information related to the request feature on the Internet using an external tool; the vector database stores a large amount of tourist attraction data in the form of vectors; A calling generation module is configured to call an external query tool to obtain and aggregate a variety of related information based on the scenic spot information and the restriction conditions to generate target information; The information transmission module is configured to transmit the target information to the client.

6. The multi-agent travel planning generation device based on a large language model according to claim 5 is characterized in that: The feature extraction module is specifically configured as follows: The query request is analyzed using thought chain technology to extract request features from the query request.

7. The multi-agent travel planning generation device based on a large language model according to claim 5 is characterized in that: The scenic spot information related to the request feature includes weather and route; the query collection module is specifically configured as follows: Querying the scenic spot information related to the requested feature from the vector database; If the scenic spot information related to the requested feature is not found in the vector database, the scenic spot information related to the requested feature is collected on the Internet using an external tool.

8. The multi-agent travel planning generation device based on a large language model according to claim 5 is characterized in that: The call generation module is specifically configured as follows: The multiple related information are displayed in a preset order to generate at least one target information.

9. An electronic device comprising a processor and a memory; the memory stores a computer program, wherein: When the computer program is executed by the processor, the computer program implements the multi-agent travel planning generation method based on a large language model according to any one of claims 1 to 4.

10. A computer-readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images