Management and control scheme generation method based on large language model and electronic equipment

Abstract

The invention discloses a management and control scheme generation method based on a big language model and electronic equipment, and the method comprises the steps: firstly obtaining a demand description input by a user, determining a target intention and an activity region corresponding to the demand description based on the big language model, and determining a target scheme template corresponding to the target intention; determining at least one target associated data type corresponding to the target scheme template; and obtaining target associated data corresponding to at least one target associated data type in a control range corresponding to the activity area. Inputting at least one piece of target associated data and a corresponding management and control scheme generation prompt word into the large language model to generate management and control scheme content; and displaying the management and control scheme content in a display area corresponding to the target scheme template, thereby generating a final management and control scheme. The method does not need to depend on manpower, improves the efficiency of formulating the management and control scheme, avoids the problem that the accuracy of the formulated management and control scheme is poor due to the influence of manual management, and improves the accuracy of the generated management and control scheme.

Description

Technical Field

[0001] This application relates to the field of intelligent transportation technology, and in particular to a method and electronic device for generating control schemes based on a large language model. Background Technology

[0002] Large-scale events involve massive numbers of travelers and demand high-quality transportation services, placing immense pressure on local and surrounding road networks and posing significant challenges to traffic management. Therefore, developing effective control plans is crucial to addressing these challenges and ensuring the smooth operation of large-scale events.

[0003] In related technologies, when organizing large-scale events, management personnel typically rely on human experience to develop control plans. On the one hand, developing control plans manually is inefficient, and on the other hand, the accuracy of the control plans may be poor due to the influence of human supervisors. Summary of the Invention

[0004] This application provides a method and electronic device for generating control schemes based on a large language model, which solves the problems in related technologies where control schemes are formulated by managers based on human experience. On the one hand, the efficiency of manually formulating control schemes is low, and on the other hand, the accuracy of the formulated control schemes is poor due to the influence of human supervisors.

[0005] Firstly, this application provides a method for generating control schemes based on a large language model, the method comprising:

[0006] Obtain the user's input requirement description, determine the target intent and activity area corresponding to the requirement description based on the large language model; determine the target solution template corresponding to the target intent based on the correspondence between at least one pre-saved intent and solution template;

[0007] Based on at least one pre-saved scheme template corresponding to at least one associated data type, determine at least one target associated data type corresponding to the target scheme template; and obtain the target associated data corresponding to the at least one target associated data type within the control range corresponding to the activity area; wherein, the target associated data includes at least one of the following: road network data, traffic flow data of various types of vehicles, traffic capacity data, passenger flow data, congestion index data, parking lot data, guidance screen data, public transportation data, and emergency support data;

[0008] Input at least one target-related data and its corresponding control scheme prompt words into the large language model to generate control scheme content; then display the control scheme content in the display area corresponding to the target scheme template.

[0009] The above technical solution has the following advantages or beneficial effects:

[0010] This application addresses the inefficiency and poor accuracy of control scheme development caused by relying on human experience and management personnel in related technologies. It proposes a control scheme generation method based on a large language model, thus achieving an intelligent method for generating control schemes based on objective data. Specifically, firstly, the user's input requirement description is obtained. Based on the large language model, the target intent and activity area corresponding to the requirement description, as well as the target scheme template corresponding to the target intent, are determined. This generates a target scheme template that matches the user's needs, ensuring that the subsequent control scheme generated based on the target scheme template also meets the user's needs, improving the user experience. Next, at least one target-related data type corresponding to the target scheme template is determined; and target-related data corresponding to at least one target-related data type within the control range corresponding to the activity area are obtained. Finally, at least one target-related data and the corresponding control scheme generation prompt are input into the large language model to generate control scheme content; the control scheme content is displayed in the display area corresponding to the target scheme template, thus generating the final control scheme. This application utilizes a large language model, combined with real and objective target-related data corresponding to at least one target-related data type within the control range corresponding to the activity area, to generate control schemes. This eliminates the need for manual intervention, improving the efficiency of developing control plans. Furthermore, this application avoids the problem of poor accuracy in the developed control plans due to the influence of human supervisors, thus improving the accuracy of the generated control plans.

[0011] Secondly, this application provides a control scheme generation device based on a large language model, the device comprising:

[0012] The determination module is used to obtain the user's input requirement description, determine the target intent and activity area corresponding to the requirement description based on the large language model, and determine the target solution template corresponding to the target intent according to the correspondence between at least one pre-saved intent and solution template.

[0013] The acquisition module is used to determine at least one target associated data type corresponding to the target scheme template based on at least one associated data type corresponding to at least one pre-saved scheme template; and to acquire the target associated data corresponding to the at least one target associated data type within the control range corresponding to the activity area; wherein, the target associated data includes at least one of the following: road network data, traffic flow data of various types of vehicles, traffic capacity data, passenger flow data, congestion index data, parking lot data, guidance screen data, public transportation data, and emergency support data;

[0014] The scheme generation module is used to input at least one target-related data and the corresponding control scheme generation prompt words into the large language model to generate control scheme content; and to display the control scheme content in the display area corresponding to the target scheme template.

[0015] Thirdly, this application provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0016] Memory, used to store computer programs;

[0017] A processor, used to execute a program stored in memory, implements the method described.

[0018] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described herein.

[0019] Fifthly, this application provides a computer program product comprising an executable program that is executed by a processor to implement the method described. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 A schematic diagram illustrating the generation process of the control scheme based on a large language model provided for this application;

[0022] Figure 2 A schematic diagram showing the passenger flow analysis content provided in this application;

[0023] Figure 3 A schematic diagram showing the congestion index analysis content provided in this application;

[0024] Figure 4 The flowchart for extracting user parameters provided in this application;

[0025] Figure 5 The flowchart of regional road network analysis based on Dify provided in this application;

[0026] Figure 6 This application provides a schematic diagram showing the location of the activity area.

[0027] Figure 7 The road network system diagram provided for this application;

[0028] Figure 8 A flowchart of traffic feature analysis based on Dify provided for this application;

[0029] Figure 9 The overall control strategy analysis flowchart based on Dify provided in this application;

[0030] Figure 10 This is a schematic diagram illustrating the control zone map provided in this application;

[0031] Figure 11 The public transportation assurance analysis flowchart based on Dify is provided for this application;

[0032] Figure 12 A flowchart of emergency support information recommendation and analysis based on Dify provided for this application;

[0033] Figure 13 A schematic diagram of the control scheme generation device based on a large language model provided in this application;

[0034] Figure 14 A schematic diagram of the electronic device structure provided in this application. Detailed Implementation

[0035] To make the objectives and implementation methods of this application clearer, the exemplary implementation methods of this application will be clearly and completely described below with reference to the accompanying drawings of the exemplary embodiments of this application. Obviously, the exemplary embodiments described are only some embodiments of this application, and not all embodiments.

[0036] It should be noted that the brief descriptions of terms in this application are only for the convenience of understanding the embodiments described below, and are not intended to limit the embodiments of this application. Unless otherwise stated, these terms should be understood in their ordinary and common meaning.

[0037] The terms "first," "second," "third," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar or related objects or entities, and do not necessarily imply a specific order or sequence, unless otherwise specified. It should be understood that such terms are interchangeable where appropriate.

[0038] The terms “comprising” and “having”, and any variations thereof, are intended to cover but not exclude inclusion, for example, a product or device that includes a range of components is not necessarily limited to all of the components that are clearly listed, but may include other components that are not clearly listed or that are inherent to such product or device.

[0039] The term "module" refers to any known or subsequently developed hardware, software, firmware, artificial intelligence, fuzzy logic, or combination of hardware and / or software code that is capable of performing the functions associated with that element.

[0040] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0041] For ease of explanation, the above description has been provided in conjunction with specific embodiments. However, the above exemplary discussion is not intended to be exhaustive or to limit the embodiments to the specific forms disclosed above. Various modifications and variations can be obtained based on the above teachings. The selection and description of the above embodiments are for the purpose of better explaining the principles and practical applications, thereby enabling those skilled in the art to better utilize the described embodiments and various different variations of embodiments suitable for specific use considerations.

[0042] Figure 1 The schematic diagram of the generation process of the control scheme based on a large language model provided for this application includes the following steps:

[0043] S101: Obtain the user's input requirement description, determine the target intent and activity area corresponding to the requirement description based on the large language model; determine the target solution template corresponding to the target intent according to the correspondence between at least one pre-saved intent and the solution template respectively;

[0044] S102: Based on at least one associated data type corresponding to at least one pre-saved scheme template, determine at least one target associated data type corresponding to the target scheme template; and obtain the target associated data corresponding to the at least one target associated data type within the control range corresponding to the activity area; wherein, the target associated data includes at least one of the following: road network data, traffic flow data of various types of vehicles, traffic capacity data, passenger flow data, congestion index data, parking lot data, guidance screen data, public transportation data, and emergency support data;

[0045] S103: Input at least one target-related data and the corresponding control scheme prompt words into the large language model to generate control scheme content; display the control scheme content in the display area corresponding to the target scheme template.

[0046] The control scheme generation method based on a large language model provided in this application is applied to electronic devices, such as PCs, computers, smart terminals, servers, etc.

[0047] The electronic device is equipped with a control scheme generation system. After logging into the system, users can enter a requirement description input box on the system interface. The electronic device receives the user's input requirement description and determines the target intent and activity area corresponding to the requirement description based on a large language model. For example, if the requirement description is "On a certain day and time, a beer festival opening ceremony will be held at a certain beer city, please generate a control scheme for me," the target intent is determined to be "generate a control scheme," the activity area is "a certain beer city," and the activity time is determined to be "a certain day and time," thus generating a control scheme for that specific time.

[0048] The electronic device pre-stores the correspondence between at least one intent and a solution template, whereby the at least one intent includes the intent to "generate a control solution". After determining the target intent corresponding to the requirement description, the electronic device determines the target solution template corresponding to the target intent based on the pre-stored correspondence between the at least one intent and the solution template. It should be noted that the process of pre-stored correspondence between at least one intent and the solution template includes: for the at least one intent, pre-stored the correspondence between the intent and the default solution template in the large language model; or pre-receiving a custom solution template corresponding to the intent input by the user, and storing the correspondence between the intent and the custom solution template.

[0049] The electronic device pre-stores at least one associated data type corresponding to at least one scheme template. This at least one associated data type includes at least one of the following: road network data type, traffic flow data type for various vehicle flows, capacity data type, passenger flow data type, congestion index data type, parking lot data type, guidance screen data type, public transportation data type, and emergency support data type. After determining the target scheme template, the electronic device determines at least one target associated data type corresponding to the target scheme template based on the pre-stored associated data types for each scheme template. Then, it acquires the target associated data corresponding to each of the at least one target associated data type within the control area corresponding to the activity area. The target associated data includes at least one of the following: road network data, traffic flow data for various vehicle flows, capacity data, passenger flow data, congestion index data, parking lot data, guidance screen data, public transportation data, and emergency support data. The control area corresponding to the activity area refers to the control area encompassing the control area itself. For example, it could be a control area with a radius of 1 km or 2 km centered on the center of the control area, or a control area encompassing a specific district or township. This application does not limit the size of the control area. Optionally, the user can also define the control area corresponding to the activity area.

[0050] After the electronic device obtains target-related data corresponding to at least one target-related data type, it inputs the at least one target-related data and the corresponding control scheme prompt words into the large language model to generate control scheme content; and displays the control scheme content in the display area corresponding to the target scheme template.

[0051] This application addresses the inefficiency and poor accuracy of control scheme development caused by relying on human experience and management personnel in related technologies. It proposes a control scheme generation method based on a large language model, thus achieving an intelligent method for generating control schemes based on objective data. Specifically, firstly, the user's input requirement description is obtained. Based on the large language model, the target intent and activity area corresponding to the requirement description, as well as the target scheme template corresponding to the target intent, are determined. This generates a target scheme template that matches the user's needs, ensuring that the subsequent control scheme generated based on the target scheme template also meets the user's needs, improving the user experience. Next, at least one target-related data type corresponding to the target scheme template is determined; and target-related data corresponding to at least one target-related data type within the control range corresponding to the activity area are obtained. Finally, at least one target-related data and the corresponding control scheme generation prompt are input into the large language model to generate control scheme content; the control scheme content is displayed in the display area corresponding to the target scheme template, thus generating the final control scheme. This application utilizes a large language model, combined with real and objective target-related data corresponding to at least one target-related data type within the control range corresponding to the activity area, to generate control schemes. This eliminates the need for manual intervention, improving the efficiency of developing control plans. Furthermore, this application avoids the problem of poor accuracy in the developed control plans due to the influence of human supervisors, thus improving the accuracy of the generated control plans.

[0052] In one optional implementation, obtaining the target association data corresponding to at least one target association data type within the control range corresponding to the activity area includes:

[0053] Acquire road network data, traffic flow data, and capacity data of various types of traffic on each arterial road within the control area corresponding to the activity area during the historical time period; the various types of traffic include destination vehicles, originating vehicles, internal vehicles, and transit vehicles;

[0054] At least one target-related data point and its corresponding control scheme are input into the large language model to generate prompt words. The generated control scheme content includes:

[0055] Input the traffic flow data and capacity data of various types of traffic on each arterial road, as well as the table analysis prompts, into the large language model to generate road network table analysis content;

[0056] Input the traffic flow data of various types of traffic on each main road and the prompt words displayed in the first chart into the large language model to generate traffic flow chart analysis content;

[0057] The road network data, traffic flow data and capacity data of various types of traffic on each arterial road, and road network analysis prompts are input into the large language model to generate road network analysis content.

[0058] The historical time period can be a period preceding the event date specified in the requirement description. For example, if the event date is May 10th, the historical time period could be May 9th, or a period from May 5th to May 9th, etc. Preferably, the historical time period can be a period during which large-scale events have occurred in the activity area specified in the requirement description.

[0059] The road network data within the control area corresponding to the activity zone includes road names, intersection names, road grades, road types, distances between roads and the activity zone, and routes. It also includes traffic flow data and capacity data for various types of traffic on each arterial road. These types of traffic include destination vehicles, originating vehicles, internal vehicles, and through vehicles. Destination vehicles are those entering and stopping within the activity zone from outside; originating vehicles are those moving from within the activity zone to outside; internal vehicles are those traveling within the activity zone; and through vehicles are those entering and exiting the activity zone from outside.

[0060] The traffic flow and capacity data for various types of traffic on each arterial road, along with table analysis prompts, are input into the large language model to generate road network table analysis content. This content includes traffic flow and capacity data for various types of traffic displayed in tabular form. Additionally, the capacity (in vehicles) for each road can be displayed in the corresponding display area of ​​the target solution template.

[0061] For example, the traffic capacity involved in each road is shown in Table 1 below:

[0062] road name Road type name Design capacity (vehicles) Road 1 provincial highway 4000 Road 2 County road 1000 Road 3 County road 1000 Road 4 rural roads 1000 Road 5 rural roads 1000

[0063] Table 1

[0064] The traffic flow data for various traffic types on each main road, along with the prompts for the first chart, are input into the large language model to generate traffic flow chart analysis content. The first chart can be a bar chart, pie chart, etc. The traffic flow chart analysis content includes traffic flow data for various traffic types displayed in the first chart format.

[0065] The road network data, traffic flow data and capacity data of various types of vehicles on each arterial road, and road network analysis prompts are input into the large language model to generate road network analysis content. The road network analysis content includes the geographical location analysis of the event area. For example, the road network analysis content might be: "A certain beer city is located in a certain district of a certain city, specifically near a certain new district of a certain county. The surrounding area has densely populated areas such as residential areas, shopping malls, office buildings, and schools. The presence of these places may place higher demands on crowd control and traffic dispersal during the event. The surrounding transportation facilities are relatively complete; the nearest subway station to the beer city is a station on Line 1, about a 10-minute walk away. In addition, multiple bus lines cover the area, connecting the city center with the new district, facilitating travel for tourists and residents to the event site. The distribution of subway and bus stops will help with crowd dispersal during the event, but may also put some pressure on surrounding traffic due to their concentration."

[0066] In one optional implementation, obtaining the target association data corresponding to at least one target association data type within the control range corresponding to the activity area includes:

[0067] Acquire passenger flow data, congestion index data, parking lot data, and road classification of each road within the control area corresponding to the activity area during historical time periods; the parking lot data includes parking lot location data, number of parking spaces, charging standard data, distance data between the parking lot and the activity area, and route data from the parking lot to the activity area.

[0068] At least one target-related data point and its corresponding control scheme are input into the large language model to generate prompt words. The generated control scheme content includes:

[0069] Input the passenger flow data within the activity area and the prompts displayed on the second chart into the large language model to generate passenger flow chart analysis content;

[0070] Input the congestion index data and the prompts displayed on the third chart within the activity area into the large language model to generate congestion index chart analysis content;

[0071] Input the parking lot data and the road network data within the control area corresponding to the activity area, along with the level data of each intersection and the table analysis prompts, into the large language model to generate traffic table analysis content;

[0072] The parking lot data and the road network data within the control area corresponding to the activity area, along with the level data of each intersection and traffic feature analysis prompts, are input into the large language model to generate traffic feature analysis content.

[0073] Obtain passenger flow data within the activity area during a historical time period. This can be done by acquiring passenger flow data within each step, such as 5-minute or 10-minute intervals. Input the passenger flow data within the activity area and the prompts for the second chart into the large language model to generate passenger flow chart analysis content. The passenger flow chart analysis content includes passenger flow data within each step, displayed in the form of a second chart. The second chart can be a line chart, bar chart, pie chart, etc. Figure 2 A schematic diagram showing the passenger flow analysis content provided in this application.

[0074] Obtain congestion index data for the activity area within a historical time period. This can be done by acquiring congestion index data for each step, such as 5-minute or 10-minute intervals. Input the congestion index data for the activity area and the prompts for the third chart into the large language model to generate congestion index chart analysis content. The congestion index chart analysis content includes congestion index data for each step, displayed in the form of a third chart. The third chart can be a line chart, bar chart, pie chart, etc. Figure 3 This is a schematic diagram showing the congestion index analysis provided in this application.

[0075] The system inputs parking lot data, road network data within the control area corresponding to the activity zone, and the level data of each intersection, along with table analysis prompts, into the large language model to generate traffic table analysis content. This traffic table analysis content includes table displays of parking lot data and table displays of intersection service levels.

[0076] For example, the table showing the service level at an intersection is shown in Table 2:

[0077]

[0078]

[0079] Table 2

[0080] For example, the table display content of the data for each parking lot is shown in Table 3:

[0081] Parking lot name Number of parking spaces Fee Standard Parking Lot 1 380 4 yuan per hour Parking Lot 2 290 Free for the first hour, then 3 yuan per hour thereafter. Parking lot 3 150 3 yuan per hour Parking lot 4 200 The actual charging rules on site shall prevail. Parking lot 5 240 The actual charging rules on site shall prevail.

[0082] Table 3

[0083] The parking lot data and the road network data within the control area corresponding to the activity area, including the level data of each intersection and traffic feature analysis prompts, are input into the large language model to generate traffic feature analysis content. For example, the traffic feature analysis content might be: "Daily total passenger flow forecast: 300,000 people, based on the 2023 Beer Festival opening ceremony's daily total passenger flow of 265,000 people; Instantaneous maximum carrying capacity: 110,000 people / square kilometer, based on the 2023 instantaneous maximum carrying capacity of 97,000 people / square kilometer, combined with the event scale growth and venue area; Traffic connection pressure period: 7 PM to 9 PM, opening ceremony time: 8 PM, peak traffic connection period expected 1 hour before and 1 hour after the event; Peak demand for ride-hailing services: 8 PM to 10 PM, during and after the event..." The peak hours for ride-hailing demand are around 10 PM. The core peak periods are: Entry Peak: Traffic flow is expected to surge around the Beer City from 7:30 PM to 8:00 PM, and queues may form at parking lot entrances. Ceremony Peak: Dense crowds are expected within the venue from 8:00 PM to 8:30 PM, with potential congestion in the stage area and surrounding food areas, requiring enhanced security management. Exit Peak: Significant pressure is expected at parking lot exits and surrounding roads from 8:30 PM to 9:30 PM; advance planning is recommended. Public transportation connection plan. Weather variables: July is the rainy season. Rainfall may lead to earlier entry and exit times, and increase the risk of wet roads and low visibility. Special events: If other human activities or emergencies occur in the vicinity on the day of the event, they may have a cumulative impact on the traffic flow of the beer festival. Real-time monitoring and emergency response need to be strengthened. "Special traffic conflict points around the Beer City are mainly concentrated at intersections. Due to the concentration of people and vehicles during the event, intersections may become high-risk areas for traffic incidents. Risk incidents include conflicts between pedestrians and vehicles, traffic accidents caused by the convergence of traffic from multiple directions, etc. The handling dimensions need to start from traffic light optimization, traffic control, and volunteer guidance to ensure traffic order." "Based on the passenger flow forecast, under the premise of traffic control and improving the level of public transportation (controlling the proportion of private car trips to within 30%), the predicted parking demand for private cars on a typical peak day of the beer festival is 10,000 spaces; the parking demand on an extreme peak day is 13,000 spaces. There are 10 parking lots with 2,872 parking spaces in the eastern area of ​​the Beer City. Combined with the parking demand, it is identified that there is a serious shortage of parking resources."

[0084] In one optional implementation, obtaining the target association data corresponding to at least one target association data type within the control range corresponding to the activity area includes:

[0085] The system acquires road network data within the control area corresponding to the activity area, including data on sub-control areas at all levels, road and intersection data within each sub-control area, parking lot data within the activity area, vehicle source data from traffic flow data of various types of vehicles within the activity area over a historical time period, and guidance screen data; wherein, the guidance screen data includes the location data of the guidance screens.

[0086] At least one target-related data point and its corresponding control scheme are input into the large language model to generate prompt words. The generated control scheme content includes:

[0087] Input the road network data of each level of sub-control area, the road and intersection data of each level of sub-control area, and the control area division prompt words into the large language model to generate control area analysis content;

[0088] Input the road network data of each level of sub-control scope, the road and intersection data of each level of sub-control scope, and the prompt words of each level of control plan into the big language model to generate the analysis content of each level of control plan;

[0089] The parking lot data within the activity area, vehicle source data from the traffic flow data of various types of vehicles within the activity area during historical time periods, and diversion route prompts are input into the large language model to generate inbound and transit diversion analysis content.

[0090] Input the parking lot data and the road network data within the control area corresponding to the activity area, along with the level data of each intersection and the table analysis prompts, into the large language model to generate traffic table analysis content;

[0091] The inducement screen data and inducement release strategy prompts are input into the large language model to generate inducement analysis content.

[0092] Data for each level of sub-control area includes, for example, Level 1, Level 2, and Level 3 sub-control area data. Road and intersection data for each level of sub-control area includes the name, location, and classification of roads and intersections. Analysis of controlled areas may include suggestions such as: "During peak activity periods, it is recommended to install temporary traffic lights or traffic control personnel at major intersections; open backup parking lots as needed to alleviate pressure on main parking lots; release real-time traffic information in advance through broadcasts and navigation platforms to guide traffic flow; when some parking lots are nearing saturation, promptly close entrances and guide traffic to other parking lots; set up temporary parking spots for short-term vehicle stops; and implement one-way traffic measures when necessary to improve vehicle traffic efficiency."

[0093] The analysis content of the control plans at each level is as follows: "Level 1 Control Zone: Level 1 Plan: 30% flow restriction in the inbound direction, with corresponding increases in the outbound direction; Level 2 Plan: 60% flow restriction in the inbound direction, with corresponding increases in the outbound direction; Level 3 Plan: Inbound roads are converted to one-way streets, and private vehicles are prohibited during severe congestion." "Level 2 Control Zone: Level 1 Plan: 30% flow restriction in the inbound direction, with corresponding increases in the outbound direction; Level 2 Plan: 50% flow restriction in the inbound direction, with corresponding increases in the outbound direction; Level 3 Plan: 50% flow restriction in the inbound direction, with corresponding increases in the outbound direction." "Level 3 Control Zone: Level 1 Plan: 10% flow restriction in the inbound direction, with corresponding increases in the outbound direction; Level 2 Plan: 20% flow restriction in the inbound direction, with corresponding increases in the outbound direction; Level 3 Plan: 40% flow restriction in the inbound direction, with corresponding increases in the outbound direction."

[0094] The analysis of entry and transit traffic diversion includes, for example, "To ensure smooth traffic during the event, the following transit traffic diversion routes are recommended: vehicles entering and leaving Qingdao city via Lijiang East Road and Lijiang West Road will be diverted along Changjiang Road and Zhujiang Road at the Jiaozhou Bay Tunnel Exit North Auxiliary Road, Jialingjiang Road Connector, Lijiang West Road Alishan Road Intersection, and Lijiang West Road Jiulianshan Road Intersection."

[0095] Traffic table analysis includes the level data of each intersection in the road network data, presented in tabular form. Guidance analysis includes, for example, "Entry guidance: Vehicles heading to the Beer City should turn right at the intersection of Lijiang West Road and Tianmushan Road, Qingyunshan Road, or Luofushan Road; Through traffic guidance: Binhai Avenue is currently congested; through traffic should detour via Changjiang Road and Zhujiang Road; Parking guidance: Parking lots east of the Beer City are nearing capacity; please park in parking lot P1 in the western area."

[0096] In one optional implementation, obtaining the target association data corresponding to at least one target association data type within the control range corresponding to the activity area includes:

[0097] Obtain public transportation data within the control area corresponding to the activity area, and vehicle source data from the traffic flow data of various types of vehicles within the activity area during historical time periods; wherein, the public transportation data includes public transportation stop location data and operating route data;

[0098] At least one target-related data point and its corresponding control scheme are input into the large language model to generate prompt words. The generated control scheme content includes:

[0099] The public transportation data within the control area corresponding to the activity area, the vehicle source data from the traffic flow data of various types of vehicles within the activity area in the historical time period, and the public transportation guarantee prompts are input into the big language model to generate public transportation guarantee analysis content.

[0100] The public transportation support analysis includes, for example, "Key routes: Tianmushan Road (metro station) Hengshan Road (metro station); Auxiliary routes: Anzi (metro station) Anzi East (metro station)." "Optimize bus route 809 at Xuejiadao Station (bus stop), increasing frequency during peak hours. It is recommended to shorten the departure interval from 10 minutes to 8 minutes, increasing capacity by approximately 20% to cope with high passenger flow during the event. Extend bus routes 807 and East 3 from Xuejiadao Primary School (bus stop) to the event area to reduce passenger transfer needs and improve travel efficiency. Utilize 18-meter articulated buses to increase single-unit capacity and meet the large passenger flow demand during the event." "After the event, it is recommended to open a midnight shuttle bus connecting the event area to surrounding major..." Residential areas and subway stations will be convenient for passengers traveling at night. Facial recognition and mobile payment fast-boarding lanes will be implemented on dedicated bus lines to improve passenger boarding and alighting efficiency and reduce queuing time. Temporary electric fences will be set up near the activity area to restrict random stops by taxis and ride-hailing vehicles, ensuring smooth traffic flow. The electric fence area will be divided into a core area and a buffer zone. Random parking is prohibited in the core area, while the buffer zone will be used for temporary turnover and dispatching. The electric fence area will be dynamically adjusted through real-time monitoring and an intelligent dispatching system to avoid traffic congestion. Shared bicycle management and deployment strategy: Electronic fence parking points for shared bicycles will be set up near major entrances and exits, subway stations, and bus stops near the activity area. One point will be set up every 500 meters to ensure coverage of major pedestrian flow nodes. Two hours before and two hours after the start of each day's activities, designated bicycles will be added to the parking areas. 500 bicycles will be added before 7:00 AM and 300 bicycles before 2:00 PM daily to meet peak-hour demand. Restricted Areas: Shared bicycles are prohibited from entering the core activity area (e.g., within 500 meters of the main venue). They are also prohibited from riding on main roads and key intersections within the activity area to ensure pedestrian safety and traffic order. Pedestrian System Optimization: The main pedestrian walkways from the main entrance to the activity area will be optimized, bottleneck sections will be widened, and clear signage will be provided to improve traffic efficiency. Rest areas equipped with seating and sunshades will be set up every 300 meters within the activity area to improve walking comfort. The areas around subway and bus stations will also be optimized. Pedestrian walkways will be widened and nighttime lighting will be increased to ensure pedestrian safety; temporary pedestrian walkways will be set up around activity areas to avoid traffic congestion in sections where large-scale crowds may occur. "When the load factor of three consecutive trains exceeds 150%, skip-stop operation will be triggered; when the passenger flow density in a subway station hall exceeds 2 people / m², a level-three response will be initiated; when the passenger boarding and alighting time of three consecutive trains in one direction at a subway station exceeds 2 minutes, the train departure frequency will be increased." "Based on historical data prediction: when the vehicle speed on a certain road section is <15 km / h, additional shuttle buses will be automatically added; based on real-time passenger flow data, when the load factor of a bus line exceeds 120%, additional shuttle buses and direct buses will be triggered; when the number of people queuing at a bus stop exceeds 50, the vehicle departure frequency will be increased."

[0101] In one optional implementation, obtaining the target association data corresponding to at least one target association data type within the control range corresponding to the activity area includes:

[0102] Obtain emergency support data within the control area corresponding to the activity area; wherein, the emergency support data includes hospital name data, level data, location data, distance data between the hospital and the activity area, and route data from the activity area to the hospital;

[0103] At least one target-related data point and its corresponding control scheme are input into the large language model to generate prompt words. The generated control scheme content includes:

[0104] The emergency support data and emergency support prompts within the control area corresponding to the activity area are input into the large language model to generate emergency support analysis content.

[0105] The emergency response analysis includes, for example, "Emergency medical treatment: The optimal route from the Beer City to the Affiliated Hospital's emergency room is as follows: Starting from the Beer City, drive south along Wutaishan Road for approximately 3 kilometers, passing through major roads such as Shuangzhu Road and Lizhu Road, and finally arrive at the Affiliated Hospital's emergency room. This route is approximately 5 kilometers long, and the estimated travel time is 10-15 minutes. It is recommended to coordinate with the traffic police department during the event to ensure that special vehicles can pass through green waves, and to arrange traffic police to direct traffic at key intersections to deal with emergencies."

[0106] In an optional implementation, the method further includes:

[0107] Input traffic information service prompts into the large language model to generate traffic control plan content; input police deployment prompts into the large language model to generate police deployment control plan content; display the traffic control plan content and the police deployment control plan content in the display area corresponding to the target plan template.

[0108] The police deployment and control plan includes, for example, "Recommended police deployment guidelines: Each squadron is responsible for area control, with fixed posts and patrol posts. The accident investigation department will deploy officers to handle rapid response in the surrounding area. The deployment of posts in each squadron's controlled area is as follows: Fixed posts: Fixed posts are set up at key points of the event venue (such as entrances and exits, stage areas, parking lots, etc.) to maintain order, verify credentials, and guide crowd flow. Patrol posts: Patrol posts are set up on the perimeter of the event venue and in the main internal passages to handle daily patrols, emergency response, and audience guidance. Rapid response team: Composed of officers from the accident investigation department, responsible for rapid response and emergency handling in the surrounding area to ensure timely handling of emergencies. Through the above deployment, the security capabilities of the event can be effectively improved, ensuring the smooth running of the event."

[0109] The traffic control plan includes, for example, "through in-depth cooperation with mainstream navigation platforms such as Baidu Maps and Gaode Maps, the transportation department inputs traffic control information and recommended detour routes into the system one week in advance. This cooperation model ensures that when citizens plan their trips, the navigation system can automatically avoid controlled road sections and provide multiple alternative route options, thereby effectively improving travel efficiency and optimizing the travel experience for citizens. Traffic information is released in advance through multiple channels such as WeChat official accounts, Weibo accounts, traffic radio stations, and television stations, covering traffic control notices, travel tips, controlled road sections and times, parking information, and travel suggestions. In addition, citizens are advised to choose green travel modes such as subways and buses, and detailed travel guides are provided to help citizens plan their trips in advance and ensure smooth travel."

[0110] In this application, the method further includes:

[0111] Upon receiving a map display instruction, the system displays map data containing the activity area in the map display area of ​​the target scheme template, and displays an activity area icon at the location corresponding to the activity area in the map data.

[0112] Obtain the target-related data and corresponding data filtering conditions carried in the map display instruction, determine the related data to be displayed based on the target-related data and corresponding data filtering conditions, and display the related data to be displayed in the map data.

[0113] Data filtering criteria include, for example, filtering for congested intersections and congested roads. Congested intersections and roads refer to those whose congestion index exceeds a threshold. Related data to be displayed includes: road network system, congested intersections, congested roads, control zones, inbound traffic diversion, transit traffic diversion, parking lots, guidance systems, traffic signal equipment, bus stops, and medical routes.

[0114] In this application, after obtaining the user-inputted requirement description, the method further includes:

[0115] If the target intent and activity area corresponding to the requirement description fail to be determined based on the large language model, a first prompt message is output to prompt the user to re-enter the requirement description; after receiving the user's re-entered requirement description, the target intent and activity area corresponding to the re-entered requirement description are determined based on the large language model;

[0116] After determining the target solution template corresponding to the target intent based on the pre-saved correspondence between at least one intent and a solution template, the method further includes:

[0117] Determine the template generation parameters corresponding to the target scheme template; the template generation parameters include activity theme parameters, activity time parameters, and activity area parameters.

[0118] If the requirement description is determined to contain the template generation parameters based on the large language model, the process proceeds to determine at least one target associated data type corresponding to the target solution template based on at least one associated data type corresponding to at least one pre-saved solution template; if the requirement description is determined not to contain the first template generation parameters based on the large language model, a second prompt message is output to prompt the user to re-enter the first template generation parameters; after receiving the first template generation parameters input by the user, the process proceeds to determine at least one target associated data type corresponding to the target solution template based on at least one associated data type corresponding to at least one pre-saved solution template.

[0119] The parameters for generating each template include the event time, event area, and event theme.

[0120] This application belongs to the field of efficient command and dispatch in intelligent transportation and police systems, involving key technologies such as domain knowledge embedding, multi-module prompt engineering design, and multi-source data access. Utilizing Dify prompt word engineering, it decomposes activity management domain knowledge to generate structured prompt words and integrates multimodal data; it quickly accesses mainstream large-scale models and integrates existing systems and databases; based on feedback, it optimizes prompt words and dynamically updates solutions, improving generation efficiency and accuracy. This application significantly enhances intelligence. The large language model integrating activity management knowledge effectively improves solution generation efficiency, and the control strategies recommended by combining road network and traffic data have high accuracy. It also supports map-based interaction and natural language interaction, improving user convenience. Compared to traditional solution generation, it significantly shortens the time, improving production efficiency and control accuracy. Secondly, it supports users to upload custom control solution templates, referring to the structural framework and core content of existing solutions to quickly generate new large-scale activity control solutions tailored to specific needs. This application, through Dify's prompt word engineering embedded in existing systems, achieves an intelligent transformation of "data-driven + knowledge-guided" in activity control solution generation. Compared to traditional solution generation methods, the efficiency and accuracy of solution generation have been greatly improved. It effectively solves the core pain points of traditional models in terms of adaptability to complex scenarios, regulatory compliance, and cross-domain collaboration efficiency, and has significant industry application value.

[0121] Figure 4 The user parameter extraction flowchart provided for this application includes: user triggering a solution generation request; obtaining an existing template list; user inputting a requirement description; performing prompt word engineering analysis to match the optimal template; if matching fails, prompting the user to re-enter the request; if matching succeeds, obtaining the necessary parameters for the template; performing prompt word engineering analysis to obtain the specific values ​​of the necessary parameters for the template from the user input; if parameters are missing, prompting the user to input the specific values ​​of the missing parameters; if all parameters are included, generating a control solution.

[0122] When a user enters their requirements in the dialog box, the backend receives the request and calls the large model to execute the prompt word project, matching the optimal template. If a match is successful, the backend queries the database for the required parameters of that template based on the matched template, and then calls the large model again to execute the parameter extraction prompt word project. If the user's requirements description covers all parameters, the backend prompts the user that a corresponding solution has been planned; clicking to view triggers a solution generation request. If any parameters are missing, the backend prompts the user to add the specific values ​​for those parameters. If a match fails, the backend prompts the user to re-enter their request.

[0123] The Chatflow generation process based on Dify's management solution is designed as follows:

[0124] 1. Regional road network analysis.

[0125] Figure 5 The regional road network analysis flowchart based on Dify provided in this application includes:

[0126] Input: Event time, event location, event theme; retrieve the latitude and longitude of the event location using the address-to-coordinates interface; query the main roads around the event location based on real road network data; query the actual traffic flow and designed capacity of the main roads based on the main roads; retrieve the vehicle profile interface to obtain the traffic flow composition during previous events in the event area; execute the large model prompt word engineering, generate a visual pie chart based on the vehicle composition data; structure prompt words using real data, execute prompt word engineering, and provide a regional road network analysis of the event area. Figure 6 A schematic diagram showing the location of the activity area provided in this application. Figure 7 A schematic diagram of the road network system provided for this application.

[0127] like Figure 7 As shown, while generating text, it will also link with the map. First, it will locate the activity area. Clicking the icon after the title "Road Network System Analysis" will display the specific location of the road network system on the map.

[0128] 2. Traffic characteristics analysis.

[0129] Figure 8 The Dify-based traffic feature analysis flowchart provided for this application includes:

[0130] Input: Event time, event location, event theme; Based on the event location, obtain the latitude and longitude of the event area; Call the interface to obtain passenger flow data during past events; Execute the large model prompt word engineering to generate a visual line chart based on passenger flow data; Call the interface to obtain the congestion index of the event area during past events; Execute the large model prompt word engineering to generate a visual line chart based on the regional congestion index data; Call the address-to-coordinate interface to obtain the latitude and longitude of the event location; Based on the real road network system data, call the interface to query the main intersections and service levels of the event location; Based on the event area, call the interface to query the parking data near the area, including the number of parking spaces, charging standards, distance, etc.; Utilize real data to structure prompt words, execute prompt word engineering, and provide a traffic characteristic analysis of the event area.

[0131] 3. Overall management and control strategy.

[0132] Figure 9 The overall control strategy analysis flowchart based on Dify provided for this application includes:

[0133] Input: Activity time, activity location, activity theme; Retrieve the latitude and longitude of the activity location using the address-to-coordinate interface; Retrieve the interface to obtain the first, second, and third-level control zones, including the zone range and road and intersection information within each zone; Utilize real data to structure prompt words, execute prompt word engineering, and provide the control zone division for the activity area; Combine expert experience and real data to structure prompt words, execute prompt word engineering, and provide control plans and activation suggestions for each level of control zone before, during, and after the activity; Based on the activity's pifolio, retrieve the interface to query parking data near the area, including the number of parking spaces, pricing standards, and distance; Retrieve the interface to obtain vehicle traceability information during past activities; Utilize real data to structure prompt words, execute prompt word engineering, and provide inbound and outbound traffic diversion routes for the activity area; Based on the activity area, retrieve the interface to query the locations of nearby traffic signal equipment and guidance screens; Utilize real data to structure prompt words, execute prompt word engineering, and provide guidance and information dissemination strategies for the activity area. Figure 10 This is a schematic diagram illustrating the control zone map provided in this application.

[0134] 4. Traffic information services.

[0135] The traffic information service executes a large-scale model prompt word project and outputs corresponding traffic information service analysis results.

[0136] 5. Public transportation guarantee.

[0137] Figure 11 The public transport assurance analysis flowchart based on Dify provided for this application includes:

[0138] Input: Event time, event location, event theme; retrieve the latitude and longitude of the event location using the address-to-coordinates API; based on the event area, retrieve information on nearby subway and bus stations, including subway lines and bus routes that stop there; retrieve vehicle traceability data from previous events using the API; use real data to structure prompts, perform prompt engineering, and provide enhancement plans for the subway and bus systems in the event area; perform prompt engineering and provide public transportation support analysis for the event area.

[0139] 6. Emergency support.

[0140] Figure 12 The emergency support information recommendation and analysis flowchart provided in this application includes:

[0141] Input: Event time, event location, event theme; retrieve the latitude and longitude of the event location using the address-to-coordinates API; based on the event area, retrieve nearby tertiary hospitals using the API, including hospital name, level, distance, etc.; retrieve the locations of nearby hospitals based on the event area location using the API, and obtain medical routes; use real data to structure prompt words, perform prompt word engineering, and provide emergency support information recommendations for the event area.

[0142] 7. Police deployment.

[0143] The traffic information service executes a large-scale model prompting word project and outputs corresponding police force deployment analysis results.

[0144] The following explains how to embed Dify's open API into existing systems:

[0145] This system achieves deep integration of the chatflow functional module with existing systems through the standardized RESTful API interface provided by the Dify platform, specifically including the following steps:

[0146] 1. Interface Authentication and Configuration: In the backend management module of the existing system, set up the DIFY API key management unit. The administrator obtains the unique authentication key through the Dify platform and stores it in the security configuration file of the existing system in an encrypted manner.

[0147] 2. Data Request and Transmission: When a user triggers a request to generate a management and control scheme within the existing system, the system sends a POST request to Dify via a pre-defined API endpoint. The request content includes core parameters such as the activity time, location, and theme.

[0148] 3. Response Parsing and Content Presentation: Upon receiving a request, Dify uses its internal large language model to process it, generating a response data package containing management plan text, visualization charts, and map annotation data. Existing systems use a pre-defined parsing module to convert the JSON-formatted response data into visual interface elements: rendering the plan text to the document editing area, converting chart instructions into Echarts-formatted configuration items, and loading the geographic coordinates and POI data returned by Dify into the map component, achieving integrated display of multimodal content.

[0149] Figure 13 A schematic diagram of the control scheme generation device based on a large language model provided in this application includes:

[0150] The determination module 11 is used to obtain the user's input requirement description, determine the target intent and activity area corresponding to the requirement description based on the large language model, and determine the target solution template corresponding to the target intent according to the correspondence between at least one pre-saved intent and the solution template.

[0151] The acquisition module 12 is used to determine at least one target associated data type corresponding to the target scheme template based on at least one associated data type corresponding to at least one pre-saved scheme template; and to acquire the target associated data corresponding to the at least one target associated data type within the control range corresponding to the activity area; wherein, the target associated data includes at least one of the following: road network data, traffic flow data of various types of vehicles, traffic capacity data, passenger flow data, congestion index data, parking lot data, guidance screen data, public transportation data, and emergency support data;

[0152] The scheme generation module 13 is used to input at least one target-related data and the corresponding control scheme generation prompt words into the large language model to generate control scheme content; and to display the control scheme content in the display area corresponding to the target scheme template.

[0153] The acquisition module 12 is specifically used to acquire road network data, traffic flow data and capacity data of various types of traffic flows on each arterial road within the control range corresponding to the activity area during the historical time period; the various types of traffic flows include destination vehicles, originating vehicles, internal vehicles and transit vehicles;

[0154] The scheme generation module 13 is specifically used to input the traffic flow data and capacity data of various types of traffic on each arterial road, as well as the table analysis prompts, into the large language model to generate road network table analysis content;

[0155] Input the traffic flow data of various types of traffic on each main road and the prompt words displayed in the first chart into the large language model to generate traffic flow chart analysis content;

[0156] The road network data, traffic flow data and capacity data of various types of traffic on each arterial road, and road network analysis prompts are input into the large language model to generate road network analysis content.

[0157] The acquisition module 12 is specifically used to acquire passenger flow data, congestion index data, parking lot data, and road level of each road in the road network data within the control range corresponding to the activity area during the historical time period; the parking lot data includes parking lot location data, parking lot number of parking spaces data, charging standard data, distance data between the parking lot and the activity area, and route data from the parking lot to the activity area.

[0158] The scheme generation module 13 is specifically used to input passenger flow data and second chart display prompts within the activity area into the large language model to generate passenger flow chart analysis content; input congestion index data and third chart display prompts within the activity area into the large language model to generate congestion index chart analysis content; input parking lot data and the level data of each intersection in the road network data within the control range corresponding to the activity area, as well as table analysis prompts, into the large language model to generate traffic table analysis content; and input parking lot data and the level data of each intersection in the road network data within the control range corresponding to the activity area, as well as traffic feature analysis prompts, into the large language model to generate traffic feature analysis content.

[0159] The acquisition module 12 is specifically used to acquire road network data at all levels of sub-control ranges within the control range corresponding to the activity area, road and intersection data at all levels of sub-control ranges, parking lot data within the activity area, vehicle source data and guidance screen data from traffic flow data of various types of vehicles within the activity area during historical time periods; wherein, the guidance screen data includes the location data of the guidance screen;

[0160] The scheme generation module 13 is specifically used to input the sub-control area data at all levels, the road and intersection data at all levels, and control area division prompts from the road network data within the control area corresponding to the activity area into the large language model to generate control area analysis content; input the sub-control area data at all levels, the road and intersection data at all levels, and control plan prompts from the road network data within the control area corresponding to the activity area into the large language model to generate control plan analysis content at all levels; input the parking lot data within the activity area, the vehicle source data from the traffic flow data of various types of vehicles within the activity area in historical time periods, and diversion route prompts into the large language model to generate entry and transit diversion analysis content; input the parking lot data and the level data of each intersection in the road network data within the control area corresponding to the activity area, as well as table analysis prompts into the large language model to generate traffic table analysis content; and input the guidance screen data and guidance release strategy prompts into the large language model to generate guidance analysis content.

[0161] The acquisition module 12 is specifically used to acquire public transportation data within the control area corresponding to the activity area and vehicle source data from the traffic flow data of various types of vehicles within the activity area during historical time periods; wherein, the public transportation data includes public transportation stop location data and operating route data;

[0162] The scheme generation module 13 is specifically used to input public transportation data within the control range corresponding to the activity area, vehicle source data from various types of traffic flow data within the activity area in historical time periods, and public transportation guarantee prompts into the large language model to generate public transportation guarantee analysis content.

[0163] The acquisition module 12 is specifically used to acquire emergency support data within the control range corresponding to the activity area; wherein, the emergency support data includes hospital name data, level data, location data, distance data between the hospital and the activity area, and route data from the activity area to the hospital;

[0164] The solution generation module 13 is specifically used to input emergency support data and emergency support prompts within the control range corresponding to the activity area into the large language model to generate emergency support analysis content.

[0165] The scheme generation module 13 is also used to input traffic information service prompts into the big language model to generate traffic control scheme content; input police deployment prompts into the big language model to generate police deployment control scheme content; and display the traffic control scheme content and the police deployment control scheme content in the display area corresponding to the target scheme template.

[0166] The device further includes:

[0167] Display module 14 is configured to receive a map display instruction, display map data containing the activity area in the map display area of ​​the target scheme template, and display an activity area icon at the location corresponding to the activity area in the map data; obtain target association data and corresponding data filtering conditions carried in the map display instruction, determine the association data to be displayed based on the target association data and corresponding data filtering conditions; and display the association data to be displayed in the map data.

[0168] The determination module 11 is further configured to, if it fails to determine the target intent and activity area corresponding to the requirement description based on the large language model, output a first prompt message to prompt the user to re-enter the requirement description; after receiving the user's re-entered requirement description, determine the target intent and activity area corresponding to the re-entered requirement description based on the large language model;

[0169] The determining module 11 is further configured to determine each template generation parameter corresponding to the target solution template; the template generation parameters include activity theme parameters, activity time parameters, and activity area parameters; if the requirement description is determined to contain each template generation parameter based on the large language model, a subsequent process is performed to determine at least one target associated data type corresponding to the target solution template based on at least one associated data type corresponding to at least one pre-saved solution template; if the requirement description is determined not to contain the first template generation parameter based on the large language model, a second prompt message is output to prompt the user to re-enter the first template generation parameter; after receiving the first template generation parameter input by the user, a subsequent process is performed to determine at least one target associated data type corresponding to the target solution template based on at least one associated data type corresponding to at least one pre-saved solution template.

[0170] This application also provides an electronic device, such as Figure 14 As shown, it includes: processor 21, communication interface 22, memory 23 and communication bus 24, wherein processor 21, communication interface 22 and memory 23 communicate with each other through communication bus 24;

[0171] The memory 23 stores a computer program, which, when executed by the processor 21, causes the processor 21 to perform any of the above method steps.

[0172] The communication bus mentioned in the above electronic devices can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0173] Communication interface 22 is used for communication between the above-mentioned electronic device and other devices.

[0174] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0175] The processors mentioned above can be general-purpose processors, including central processing units, network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits, field-programmable gate arrays or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0176] This application also provides a computer-readable storage medium storing a computer program executable by an electronic device, which, when run on the electronic device, causes the electronic device to perform any of the above method steps.

[0177] This application provides a computer program product, which includes an executable program that, when executed by a processor, implements the method described herein.

[0178] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0179] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for generating control schemes based on a large language model, characterized in that, The method includes: Obtain the user's input requirement description, determine the target intent and activity area corresponding to the requirement description based on the large language model; determine the target solution template corresponding to the target intent based on the correspondence between at least one pre-saved intent and solution template; Based on at least one pre-saved scheme template corresponding to at least one associated data type, determine at least one target associated data type corresponding to the target scheme template; and obtain the target associated data corresponding to the at least one target associated data type within the control range corresponding to the activity area; wherein, the target associated data includes at least one of the following: road network data, traffic flow data of various types of vehicles, traffic capacity data, passenger flow data, congestion index data, parking lot data, guidance screen data, public transportation data, and emergency support data; Input at least one target-related data and its corresponding control scheme prompt words into the large language model to generate control scheme content; then display the control scheme content in the display area corresponding to the target scheme template.

2. The method as described in claim 1, characterized in that, Obtaining the target association data corresponding to at least one target association data type within the control range corresponding to the activity area includes: Acquire road network data, traffic flow data, and capacity data of various types of traffic on each arterial road within the control area corresponding to the activity area during the historical time period; the various types of traffic include destination vehicles, originating vehicles, internal vehicles, and transit vehicles; At least one target-related data point and its corresponding control scheme are input into the large language model to generate prompt words. The generated control scheme content includes: Input the traffic flow data and capacity data of various types of traffic on each arterial road, as well as the table analysis prompts, into the large language model to generate road network table analysis content; Input the traffic flow data of various types of traffic on each main road and the prompt words displayed in the first chart into the large language model to generate traffic flow chart analysis content; The road network data, traffic flow data and capacity data of various types of traffic on each arterial road, and road network analysis prompts are input into the large language model to generate road network analysis content.

3. The method as described in claim 1, characterized in that, Obtaining the target association data corresponding to at least one target association data type within the control range corresponding to the activity area includes: Acquire passenger flow data, congestion index data, parking lot data, and road classification of each road within the control area corresponding to the activity area during historical time periods; the parking lot data includes parking lot location data, number of parking spaces, charging standard data, distance data between the parking lot and the activity area, and route data from the parking lot to the activity area. At least one target-related data point and its corresponding control scheme are input into the large language model to generate prompt words. The generated control scheme content includes: Input the passenger flow data within the activity area and the prompts displayed on the second chart into the large language model to generate passenger flow chart analysis content; Input the congestion index data and the prompts displayed on the third chart within the activity area into the large language model to generate congestion index chart analysis content; Input the parking lot data and the road network data within the control area corresponding to the activity area, along with the level data of each intersection and the table analysis prompts, into the large language model to generate traffic table analysis content; The parking lot data and the road network data within the control area corresponding to the activity area, along with the level data of each intersection and traffic feature analysis prompts, are input into the large language model to generate traffic feature analysis content.

4. The method as described in claim 1, characterized in that, Obtaining the target association data corresponding to at least one target association data type within the control range corresponding to the activity area includes: The system acquires road network data within the control area corresponding to the activity area, including data on sub-control areas at all levels, road and intersection data within each sub-control area, parking lot data within the activity area, vehicle source data from traffic flow data of various types of vehicles within the activity area over a historical time period, and guidance screen data; wherein, the guidance screen data includes the location data of the guidance screens. At least one target-related data point and its corresponding control scheme are input into the large language model to generate prompt words. The generated control scheme content includes: Input the road network data of each level of sub-control area, the road and intersection data of each level of sub-control area, and the control area division prompt words into the large language model to generate control area analysis content; Input the road network data of each level of sub-control scope, the road and intersection data of each level of sub-control scope, and the prompt words of each level of control plan into the big language model to generate the analysis content of each level of control plan; The parking lot data within the activity area, vehicle source data from the traffic flow data of various types of vehicles within the activity area during historical time periods, and diversion route prompts are input into the large language model to generate inbound and transit diversion analysis content. Input the parking lot data and the road network data within the control area corresponding to the activity area, along with the level data of each intersection and the table analysis prompts, into the large language model to generate traffic table analysis content; The inducement screen data and inducement release strategy prompts are input into the large language model to generate inducement analysis content.

5. The method as described in claim 1, characterized in that, Obtaining the target association data corresponding to at least one target association data type within the control range corresponding to the activity area includes: Obtain public transportation data within the control area corresponding to the activity area, and vehicle source data from the traffic flow data of various types of vehicles within the activity area during historical time periods; wherein, the public transportation data includes public transportation stop location data and operating route data; At least one target-related data point and its corresponding control scheme are input into the large language model to generate prompt words. The generated control scheme content includes: The public transportation data within the control area corresponding to the activity area, the vehicle source data from the traffic flow data of various types of vehicles within the activity area in the historical time period, and the public transportation guarantee prompts are input into the big language model to generate public transportation guarantee analysis content.

6. The method as described in claim 1, characterized in that, Obtaining the target association data corresponding to at least one target association data type within the control range corresponding to the activity area includes: Obtain emergency support data within the control area corresponding to the activity area; wherein, the emergency support data includes hospital name data, level data, location data, distance data between the hospital and the activity area, and route data from the activity area to the hospital; At least one target-related data point and its corresponding control scheme are input into the large language model to generate prompt words. The generated control scheme content includes: The emergency support data and emergency support prompts within the control area corresponding to the activity area are input into the large language model to generate emergency support analysis content.

7. The method as described in claim 1, characterized in that, The method further includes: Input traffic information service prompts into the large language model to generate traffic control plan content; input police deployment prompts into the large language model to generate police deployment control plan content; display the traffic control plan content and the police deployment control plan content in the display area corresponding to the target plan template.

8. The method as described in claim 1, characterized in that, The method further includes: Upon receiving a map display instruction, the system displays map data containing the activity area in the map display area of ​​the target scheme template, and displays an activity area icon at the location corresponding to the activity area in the map data. Obtain the target-related data and corresponding data filtering conditions carried in the map display instruction, determine the related data to be displayed based on the target-related data and corresponding data filtering conditions, and display the related data to be displayed in the map data.

9. The method as described in claim 1, characterized in that, After obtaining the user input requirement description, the method further includes: If the target intent and activity area corresponding to the requirement description fail to be determined based on the large language model, a first prompt message is output to prompt the user to re-enter the requirement description; after receiving the user's re-entered requirement description, the target intent and activity area corresponding to the re-entered requirement description are determined based on the large language model; After determining the target solution template corresponding to the target intent based on the pre-saved correspondence between at least one intent and a solution template, the method further includes: Determine the template generation parameters corresponding to the target scheme template; the template generation parameters include activity theme parameters, activity time parameters, and activity area parameters. If the requirement description is determined to contain the template generation parameters based on the large language model, the process proceeds to determine at least one target associated data type corresponding to the target solution template based on at least one associated data type corresponding to at least one pre-saved solution template; if the requirement description is determined not to contain the first template generation parameters based on the large language model, a second prompt message is output to prompt the user to re-enter the first template generation parameters; after receiving the first template generation parameters input by the user, the process proceeds to determine at least one target associated data type corresponding to the target solution template based on at least one associated data type corresponding to at least one pre-saved solution template.

10. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, when executing a program stored in memory, implements the method described in any one of claims 1-9.

Images