An ai-assisted differential travel course creation system

Abstract

The application provides an AI-assisted travel itinerary creation system, belonging to the technical field of itinerary creation, which automatically completes demand recognition, itinerary generation and dynamic adjustment through an AI algorithm, without human intervention, and the planning efficiency is improved by more than 80% compared with the manual mode; at the same time, the visual itinerary display and one-key reservation function further shorten the cycle of itinerary confirmation and execution. Through implicit demand mining and multi-objective optimization, the matching degree of the itinerary scheme and the user's personalized demand and official activity demand is improved by more than 90%, and the frequency of itinerary modification caused by demand mismatch is reduced. Through precise regulation and control of the cost sensitivity weight and multi-objective optimization, the total travel cost is reduced by an average of 15%-25% compared with the prior art under the premise of meeting the travel demand; at the same time, the additional expenditure caused by unreasonable travel is reduced.

Description

Technical Field

[0001] This invention belongs to the field of itinerary creation technology, specifically relating to an AI-assisted business travel itinerary creation system. Background Technology

[0002] With the increasing frequency of global business activities, business travel has become an important part of corporate operations and individual work. The creation of business travel itineraries involves multiple dimensions such as transportation booking, accommodation arrangement, business coordination, time planning, and cost control. It also needs to take into account timeliness, convenience, economy, and precise matching with business needs. Traditional business travel itinerary creation methods are no longer able to meet the current needs for efficient and personalized business travel and have many shortcomings.

[0003] Current business travel itinerary creation mainly relies on manual planning or simple online booking platform integration. Manual planning is inefficient and prone to inefficient itineraries due to information asymmetry (such as real-time traffic delays, hotel full occupancy, sudden weather changes at the destination), and it is difficult to balance multi-dimensional needs (such as achieving the shortest commuting time and optimal business connections within a limited budget). While simple online platforms can provide basic booking services for transportation and accommodation, they lack in-depth understanding of the personalized needs of business travelers (such as user travel preferences, past travel habits, business priorities, etc.), and cannot achieve dynamic optimization and intelligent adaptation of itineraries.

[0004] Specifically, the existing technology has the following shortcomings: (1) Low demand matching: It can only respond to the basic needs explicitly input by users (such as departure point, destination, and time), and cannot actively identify implicit needs (such as the preparation time for official activities, the user's preference weight for transportation mode, and the buffer time for transit connections in cross-city business trips); (2) Poor dynamic adaptability: When unexpected situations occur during the execution of the trip (such as flight cancellation, change of meeting time, or sudden weather disaster at the destination), it cannot quickly generate the optimal adjustment plan, and users need to manually replan; (3) Imbalance in multi-objective optimization: It is difficult to simultaneously take into account multiple optimization objectives such as time, cost, and convenience, and usually can only achieve the optimization of a single objective (such as the lowest price), ignoring travel efficiency and user experience; (4) Incomplete information integration: It does not fully integrate multi-source heterogeneous data (such as real-time traffic data, real-time hotel room status data, official contact person schedule data, destination government / business service data, weather warning data, etc.), resulting in insufficient accuracy and feasibility of itinerary planning.

[0005] Therefore, there is an urgent need for an AI-assisted business trip creation system that can deeply mine user needs, integrate multi-source data, and achieve multi-objective dynamic optimization, in order to solve the problems of low trip planning efficiency, poor demand matching, and weak dynamic adaptability in existing technologies. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of existing travel itinerary creation technologies, such as low demand matching degree, poor dynamic adaptability, imbalance of multi-objective optimization and incomplete information integration, and to provide an AI-assisted travel itinerary creation system. This system achieves intelligent planning, accurate matching and real-time adjustment of travel itineraries by integrating multi-source heterogeneous data, constructing a multi-dimensional demand identification model and dynamic optimization algorithm, thereby improving travel efficiency, reducing travel costs and optimizing user experience.

[0007] The present invention employs the following technical solution.

[0008] An AI-assisted business trip creation system includes:

[0009] The system includes a data acquisition module, a demand identification module, a trip planning module, a dynamic optimization module, an output interaction module, and a data storage module.

[0010] Data acquisition module: used to collect multi-source heterogeneous data, including basic travel data, real-time dynamic data, user personalized data, and official business related data;

[0011] Demand identification module: Based on AI algorithms, it performs in-depth analysis of the collected data to identify users' explicit and implicit needs;

[0012] Trip planning module: Based on the demand vector output by the demand identification module and combined with the multi-source data provided by the data acquisition module, an initial travel itinerary plan is generated through a multi-objective optimization algorithm;

[0013] Dynamic optimization module: Real-time monitoring of dynamic data acquired by the data acquisition module. When a sudden situation affecting the execution of the trip is detected, the dynamic optimization algorithm is triggered to adjust the existing trip plan and generate the optimal adjustment plan.

[0014] Output interaction module: used to display itinerary plans to users, receive user feedback, and pass user feedback to the demand identification module or dynamic optimization module;

[0015] Data storage module: Used to store collected multi-source data, user demand data, itinerary plan data, optimization log data and user feedback data, providing data support for the iterative optimization of AI algorithms.

[0016] Preferably, in the data acquisition module, basic travel data includes departure point, destination, travel period, budget range, and other basic user input data.

[0017] Preferably, in the data acquisition module, real-time dynamic data includes real-time traffic data, hotel real-time data, and destination real-time data.

[0018] Preferably, in the data collection module, user-personalized data includes the user's past business travel records, transportation preferences, accommodation standard preferences, dining preferences, and travel time habits; official business-related data includes official activity information, schedule data of the unit being contacted, and business travel approval process data.

[0019] Preferably, the demand identification module specifically includes constructing a demand identification model. This model takes basic user input data as initial input, combines user personalized data and official business related data to perform feature extraction and demand inference, and outputs a demand vector containing dimensions such as official business priority, time sensitivity, cost sensitivity, comfort preference, and connection buffer needs.

[0020] Preferably, in the itinerary planning module, the initial itinerary plan includes transportation plan, accommodation plan, official activity arrangement, meal and rest time planning, and cost details, and meets the constraints of reasonable time connection of each link, cost not exceeding the budget range, and matching the priority of official activities.

[0021] Preferred options also include:

[0022] The demand weight calculation function is used to calculate the weights for time sensitivity, cost sensitivity, comfort preference, and priority, providing a basis for multi-objective optimization.

[0023] ω t +ω e +ω_c+ω_b=1(1)

[0024] in:

[0025] ω t This is the time sensitivity weight, with a value range of [0,1], and is calculated as ω. t =a1×f t +a2×e_b, where a1 and a2 are weighting coefficients, f t The frequency of user time optimization needs is used as the percentage, where e_b is the urgency coefficient of official business.

[0026] ω e This represents the cost sensitivity weight, with a value range of [0,1].

[0027] ω_c is the comfort preference weight, with a value range of [0,1]. It is calculated as ω_c = γ1 × f tr +γ2×f_h, where γ1 and γ2 are weighting coefficients, f tr f_h represents the percentage of frequency of choosing high-end transportation and the percentage of frequency of choosing high-end accommodation.

[0028] ω_b is the priority weight of official duties, with a value range of [0,1]. It is calculated as ω_b=δ1×i_b+δ2×l_b+δ3×r_b, where δ1, δ2, and δ3 are weight coefficients, i_b is the importance coefficient of official duties, l_b is the level coefficient of the participants, and r_b is the time rigidity coefficient.

[0029] Preferably, ω e The calculation method is ω e =β1×(B max -B avg ) / B max +β2×f e Where β1 and β2 are weighting coefficients, B max B is the upper limit of the travel budget. avg f is the user's average past travel expenses. e This represents the percentage of times expenses were exceeded.

[0030] Preferred options also include:

[0031] With the optimization objectives of minimizing total travel time, reducing total cost, maximizing comfort, and ensuring the best match between travel and official business, a weighted summation multi-objective optimization function is constructed:

[0032] Min F=ω1×T tot +ω e ×C tot -ω_c×S_c-ω_b×M_b (2);

[0033] in:

[0034] F represents the objective function value to be optimized.

[0035] T tot The total travel time includes the sum of time spent on transportation, business activities, transfers, rest, and other aspects of the trip.

[0036] C tot The total cost of the business trip includes transportation, accommodation, meals, and business-related expenses.

[0037] S_c is the comfort score, calculated by weighting the comfort of transportation, accommodation, and connection smoothness. S_c = λ1 × S tr +λ2×S_h+λ3×S_con, where λ1, λ2, and λ3 are weighting coefficients, and S tr S_h is the rating for transportation comfort, S_h is the rating for accommodation comfort, and S_con is the rating for seamless connection.

[0038] M_b represents the matching degree of official business coordination, which is used to evaluate the degree of matching between the itinerary and the needs of official activities. M_b = μ1×M_p + μ2×M_t + μ3×M_s, where μ1, μ2, and μ3 are weighting coefficients, M_p is the score for the sufficiency of preparation time, M_t is the score for the accuracy of meeting time coordination, and M_s is the score for the matching degree of the schedule of the coordinating personnel.

[0039] Constraints: Ctot≤Bmax; Tstart≤T_b e 9in; T arrive +T_p re ≤T_b e 9in; S_h≥S_hmin; T_w ait ≤T_wmax.

[0040] Preferred options also include:

[0041] The dynamic adjustment trigger threshold function determines whether to trigger travel adjustment when an unexpected situation occurs.

[0042] ΔF'=F'_new-F'_old≥θ(3);

[0043] in:

[0044] ΔF' represents the change in the objective function value; F'_old represents the normalized objective function value of the original travel plan; F'_new represents the normalized objective function value of the original travel plan after the occurrence of the unexpected situation; θ represents the adjustment trigger threshold.

[0045] The beneficial effects of the present invention are as follows, compared with the prior art:

[0046] 1. Improve itinerary planning efficiency: AI algorithms automatically identify needs, generate itineraries, and dynamically adjust them without human intervention, improving planning efficiency by more than 80% compared to manual methods; at the same time, the visual itinerary display and one-click booking function further shorten the itinerary confirmation and execution cycle.

[0047] 2. Improve the matching degree of needs: Through the mining of implicit needs and multi-objective optimization, the matching degree of itinerary plans with users' personalized needs and business activity needs has been improved by more than 90%, reducing the frequency of itinerary modifications due to mismatch of needs.

[0048] 3. Reduce travel costs: Through precise control of cost sensitivity weights and multi-objective optimization, the total travel cost is reduced by an average of 15%-25% compared to existing technologies, while meeting travel needs; at the same time, it reduces additional expenses caused by unreasonable itineraries (such as rescheduling fees and extra accommodation fees).

[0049] 4. Enhanced dynamic adaptability: In the event of an emergency, the dynamic optimization algorithm can generate the optimal adjustment plan within 30 seconds with an accuracy rate of over 95%, avoiding business delays or itinerary failures caused by emergencies and improving the smoothness of business trips.

[0050] 5. Optimize user experience: By optimizing comfort and business suitability, user satisfaction with business travel itineraries has increased by over 90%; at the same time, it has reduced the operational costs for users in itinerary planning and adjustment, and improved the convenience and comfort of business travel.

[0051] 6. Facilitates corporate travel management: The system can connect to the corporate travel approval system to generate standardized itinerary reports and expense details, enabling transparent management of travel expenses and reducing corporate travel management costs. Attached Figure Description

[0052] Figure 1 This is a flowchart of the AI-assisted travel itinerary creation system in this invention. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, other embodiments obtained by those skilled in the art without creative effort are all within the protection scope of this invention.

[0054] like Figure 1 As shown, this invention proposes an AI-assisted business trip creation system, comprising the following steps:

[0055] The system includes a data acquisition module, a demand identification module, an itinerary planning module, a dynamic optimization module, an output interaction module, and a data storage module. These modules work together to achieve intelligent creation and dynamic adjustment of the entire business trip itinerary.

[0056] Data acquisition module: used to collect multi-source heterogeneous data, including basic travel data, real-time dynamic data, user personalized data, and official business related data;

[0057] In a preferred but non-limiting embodiment of the present invention, the basic travel data in the data acquisition module includes user-input data such as departure point, destination, travel period, and budget range.

[0058] In a preferred but non-limiting embodiment of the present invention, the real-time dynamic data in the data acquisition module includes real-time traffic data (flight / high-speed rail / bus schedules, available tickets, delay information, and ticket prices), hotel real-time data (room availability, prices, geographical location, and user ratings), and destination real-time data (weather, traffic congestion index, and government service hours).

[0059] In a preferred but non-limiting embodiment of the present invention, the user personalized data in the data acquisition module includes the user's past business travel records, transportation mode preferences, accommodation standard preferences, dining preferences, and travel time habits; the official business related data includes official activity information (meeting time, location, participants, and required preparation time), schedule data of the cooperating unit, and business travel approval process data.

[0060] Demand identification module: Based on AI algorithms, it performs in-depth analysis of the collected data to identify users' explicit and implicit needs;

[0061] In a preferred but non-limiting embodiment of the present invention, the demand identification module specifically includes constructing a demand identification model. This model takes basic user input data as initial input, combines user personalized data and official business related data to perform feature extraction and demand reasoning, and outputs a demand vector containing dimensions such as official business priority, time sensitivity, cost sensitivity, comfort preference, and connection buffer needs.

[0062] Trip planning module: Based on the demand vector output by the demand identification module and combined with the multi-source data provided by the data acquisition module, an initial travel itinerary plan is generated through a multi-objective optimization algorithm;

[0063] In a preferred but non-limiting embodiment of the present invention, the initial itinerary plan in the itinerary planning module includes a transportation plan, an accommodation plan, arrangements for official activities, planning of meal and rest time, and a cost breakdown, and meets constraints such as the reasonableness of the time connection of each link, the cost not exceeding the budget range, and the matching of the priority of official activities.

[0064] Dynamic optimization module: Real-time monitoring of dynamic data acquired by the data acquisition module. When a sudden situation affecting the execution of the trip is detected (such as traffic delays, flight cancellations, changes in meeting times, or abnormal weather at the destination), the dynamic optimization algorithm is triggered to adjust the existing itinerary plan and generate the optimal adjustment plan. At the same time, the itinerary can be further optimized based on real-time user feedback (such as dissatisfaction with a certain aspect).

[0065] Output interaction module: used to display the itinerary plan to users (including a visual itinerary timeline, detailed information of each stage, cost details, and booking links), receive user feedback (such as itinerary modification requests and confirmation instructions), and pass user feedback to the requirement identification module or dynamic optimization module; at the same time, it can generate itinerary reports required for business travel approval and connect to the enterprise business travel approval system.

[0066] Data storage module: Used to store collected multi-source data, user demand data, itinerary plan data, optimization log data and user feedback data, providing data support for the iterative optimization of AI algorithms.

[0067] The core of this invention lies in the demand identification model and the multi-objective dynamic optimization algorithm. The multi-objective dynamic optimization algorithm achieves a multi-dimensional balance of time, cost, convenience, and comfort by constructing an optimization objective function, as detailed below:

[0068] In a preferred but non-limiting embodiment of the present invention, the AI-assisted business trip creation system further includes:

[0069] The demand weight calculation function is used to calculate the weights for time sensitivity, cost sensitivity, comfort preference, and priority, providing a basis for multi-objective optimization.

[0070] ω t +ω e +ω_c+ω_b=1(1)

[0071] in:

[0072] ω t This represents the time sensitivity weight, ranging from [0,1]. A larger value indicates a higher requirement for the time efficiency of the user's business trip. It is calculated by weighting the frequency of the user's need to shorten trip time in their past business travel records and the urgency of official activities (e.g., 1.0 for urgent meetings and 0.6 for regular official business). The calculation method is ω. t =a1×f t +a2×e_b, where a1 and a2 are weighting coefficients (a1+a2=1), f t The frequency of user time optimization needs is used as the percentage, where e_b is the urgency coefficient of official business.

[0073] ω e This is the cost sensitivity weight, with a value range of [0,1]. The larger the value, the higher the user's requirements for controlling travel expenses. It is calculated based on the user's travel budget range, past travel expense overruns, and the company's travel expense control standards.

[0074] In a preferred but non-limiting embodiment of the present invention, ω e The calculation method is ω e=β1×(B max -B avg ) / B max +β2×f e Where β1 and β2 are weighting coefficients (β1 + β2 = 1), / B max B is the upper limit of the travel budget. avg f is the user's average past travel expenses. e This represents the percentage of times expenses were exceeded.

[0075] ω_c represents the comfort preference weight, ranging from [0,1]. A larger value indicates a higher requirement for travel comfort. It is calculated from the user's preference records for transportation mode and accommodation standard, such as the frequency of the user's past choices of business class / first class and the frequency of choosing four-star or higher hotels. The calculation method is ω_c=γ1×f tr +γ2×f_h, where γ1 and γ2 are weighting coefficients (γ1+γ2=1), f tr f_h represents the percentage of frequency of choosing high-end transportation and the percentage of frequency of choosing high-end accommodation.

[0076] ω_b is the priority weight of official activities, with a value range of [0,1]. The larger the value, the higher the priority of the official activities. It is calculated based on the importance of the official activities, the level of the participants, and the time rigidity. The calculation method is ω_b=δ1×i_b+δ2×l_b+δ3×r_b, where δ1, δ2, and δ3 are weight coefficients (δ1+δ2+δ3=1), i_b is the coefficient of importance of the official activities, l_b is the coefficient of the level of the participants, and r_b is the coefficient of time rigidity.

[0077] In a preferred but non-limiting embodiment of the present invention, the AI-assisted business trip creation system further includes:

[0078] With the optimization objectives of minimizing total travel time, reducing total cost, maximizing comfort, and ensuring the best match between travel and official business, a weighted summation multi-objective optimization function is constructed:

[0079] Min F=ω1×T tot +ω e ×C tot -ω_c×S_c-ω_b×M_b (2);

[0080] in:

[0081] F represents the objective function value; the smaller the value, the better the travel plan.

[0082] T tot The total travel time includes the sum of time spent on transportation, business activities, transfers, rest, and other aspects of the trip.

[0083] C tot The total cost of the business trip includes transportation, accommodation, meals, and business-related expenses.

[0084] S_c is the comfort score, calculated by weighting the comfort of transportation (e.g., business class 90-100 points, economy class 60-80 points), accommodation comfort (e.g., four-star hotel 80-90 points, three-star hotel 60-70 points), and connection smoothness (100 points for no waiting time, 10 points deducted for each additional hour of waiting time, minimum 0 points). S_c = λ1 × S tr +λ2×S_h+λ3×S_con, where λ1, λ2, and λ3 are weighting coefficients (λ1+λ2+λ3=1), S tr S_h is the rating for transportation comfort, S_h is the rating for accommodation comfort, and S_con is the rating for seamless connection.

[0085] M_b represents the matching degree of official business coordination (unit: points, value range [0,100]), used to evaluate the degree of matching between the itinerary and the needs of official activities. It is calculated by weighting the sufficiency of preparation time before official activities, the accuracy of meeting time coordination, and the matching degree of the schedule of the coordinating personnel. M_b=μ1×M_p+μ2×M_t+μ3×M_s, where μ1, μ2, and μ3 are weight coefficients (μ1+μ2+μ3=1). M_p is the score for the sufficiency of preparation time (unit: points, 100 points for preparation time ≥ required time, 20 points deducted for each hour less, minimum 0 points). M_t is the score for the accuracy of meeting time coordination (unit: points, 100 points for arrival time 30-60 minutes earlier than the start time of the meeting, 10 points deducted for each 10 minutes of deviation, minimum 0 points). M_s is the score for the matching degree of the schedule of the coordinating personnel (unit: points, 100 points for a complete match between the travel time and the available time of the coordinating personnel, 50 points for a partial match, and 0 points for no match).

[0086] Constraints:

[0087] C tot ≤B max (Total expenses shall not exceed the budget limit);

[0088] T start ≤T_b e9in (Departure time T) start No later than the latest departure time T_b corresponding to the start time of the official activity e9in );

[0089] T arrive +T_p re ≤T_b e9in (Traffic arrival time T) arrive +Preparation time T_pre ≤Start time of official activities T_b e9in );

[0090] S_h≥S_hmin (Accommodation comfort rating S_h is not lower than the user's lowest acceptable rating S_hmin);

[0091] T_w wait ≤T_wmax (the transit connection waiting time does not exceed the maximum acceptable waiting time for the user).

[0092] In Formula 2, each sub-item needs to be normalized in actual calculations. The hours, yuan, and minutes are all converted into dimensionless normalized values ​​(range [0,1]). The specific normalization method is as follows:

[0093] T tot '=(T tot' -T tmin ) / (T tmax -T tmin ) (T tmin For the theoretically shortest time, T tmax (the longest acceptable time for the user).

[0094] C tot' '=(C tot' -C tmin ) / (C tmax -C tmin ) (C tmin For the theoretical minimum cost, C tmax (for the budget ceiling)

[0095] S_c'=S_c / 100; M_b'=M_b / 100;

[0096] The normalized objective function is: min F'=ω t ×T tot '+ω e ×C tot' '-ω_c×S_c'-ω_b×M_b' (Formula 2'), where each sub-term is dimensionless.

[0097] In a preferred but non-limiting embodiment of the present invention, the AI-assisted business trip creation system further includes:

[0098] The dynamic adjustment trigger threshold function determines whether to trigger travel adjustment when an unexpected situation occurs.

[0099] ΔF'=F'_new-F'_old≥θ(3);

[0100] in:

[0101] ΔF' represents the change in the objective function value; F'_old represents the normalized objective function value of the original travel plan; F'_new represents the normalized objective function value of the original travel plan after the occurrence of the emergency; θ represents the adjustment trigger threshold (dimensionless, value range [0.05, 0.2], which can be adaptively adjusted according to the user's sensitivity to travel changes; if the sensitivity is high, θ takes a small value, and if the sensitivity is low, θ takes a large value).

[0102] When ΔF'≥θ, it indicates that the unexpected situation has caused a significant decrease in the optimization level of the original travel plan, triggering the dynamic optimization algorithm; otherwise, no adjustment is needed.

[0103] The features of this invention are:

[0104] 1. A multi-dimensional demand identification model was constructed, which breaks through the limitation of existing technologies that can only identify explicit demands: by integrating user personalized data and official related data, implicit demands (such as preparation time and connection buffer demand) are actively mined, and the demand dimensions are quantified through the demand weight calculation function, providing a precise basis for multi-objective optimization; existing technologies do not involve the quantitative identification of implicit demands and cannot achieve deep matching between demands and itineraries.

[0105] 2. A multi-objective weighted optimization algorithm is proposed to solve the imbalance problem of single-objective optimization in existing technologies: by constructing a multi-objective function that includes time, cost, comfort, and business matching degree, and combining normalization processing to achieve dimensional uniformity, the optimization logic is ensured to be smooth; at the same time, multi-dimensional constraints are set to ensure the feasibility and rationality of the trip; existing technologies mostly adopt single-objective (such as lowest price) optimization, ignoring travel efficiency and user experience, and cannot meet comprehensive needs.

[0106] 3. An adaptive dynamic adjustment mechanism was designed to improve the dynamic adaptability of the trip: By dynamically adjusting the trigger threshold function, the impact of sudden situations on the trip can be accurately judged, and dynamic optimization can be achieved by triggering on demand. Existing technologies lack accurate adjustment trigger judgment, which either leads to excessive adjustment and waste of resources or untimely adjustment and trip failure.

[0107] 4. Achieved deep integration and real-time interaction of multi-source heterogeneous data: Integrates real-time traffic, hotel, destination, and business coordination data to ensure the accuracy and timeliness of itinerary planning; at the same time, the data storage module enables data accumulation and iteration, providing data support for the optimization of AI algorithms; existing technologies only simply integrate some basic data and do not achieve real-time interaction and deep integration of multi-source data, resulting in insufficient accuracy in itinerary planning.

[0108] A specific embodiment of the present invention is shown below:

[0109] 1.1 System Architecture Deployment

[0110] In this embodiment, the AI-assisted business trip creation system adopts a cloud-edge collaborative architecture: the cloud deploys a data acquisition module, a demand identification module, a trip planning module, a dynamic optimization module, and a data storage module to achieve centralized data processing and efficient algorithm operation; the terminal (including mobile APP, computer web page, and enterprise OA system client) deploys an output interaction module to realize user interaction with the system.

[0111] The data storage module uses a distributed database to store multi-source data: real-time dynamic data (traffic, hotels, weather, etc.) is cached using Redis to ensure real-time data retrieval; user-personalized data, official business-related data, and itinerary data are persistently stored using a MySQL database to ensure data security and traceability.

[0112] The data acquisition module connects with third-party platforms via API interfaces: it connects with airline and high-speed rail ticketing systems to obtain real-time traffic data; it connects with mainstream hotel booking platforms to obtain real-time hotel data; it connects with meteorological departments and traffic management departments to obtain real-time destination data; it connects with enterprise OA systems and schedule management systems to obtain official-related data; and it collects user input and feedback data through terminals.

[0113] 1.2 System Workflow

[0114] In this embodiment, the system's workflow is as follows:

[0115] Step 1: The user inputs basic travel information through the terminal, including departure point (e.g., Beijing), destination (e.g., Shanghai), travel period (March 10-12, 2026), budget range (within 5,000 yuan), and business activity information (a meeting at a company in Shanghai at 10:00 AM on March 11, attended by company executives). At the same time, the system retrieves the user's personalized data through the data storage module (past travel choices mostly included high-speed rail business class and four-star hotels, no overspending records, and high time sensitivity).

[0116] Step 2: The data acquisition module initiates multi-source data acquisition: It acquires real-time high-speed rail / flight data from Beijing to Shanghai on March 10th (e.g., G1 high-speed rail, departing at 8:00 AM and arriving at 11:30 AM, business class seats are plentiful, fare 630 yuan; MU5101 flight, departing at 7:30 AM and arriving at 9:50 AM, first class seats are plentiful, fare 1200 yuan); it acquires real-time data on four-star hotels near the Shanghai conference venue from March 10th to March 11th (e.g., a hotel in Shanghai, 1 km from the conference venue, room rate 800 yuan / night, ample room availability, user rating 4.8); it acquires weather data for Shanghai from March 10th to March 12th (sunny, no precipitation, low traffic congestion index); and it acquires the schedule data of a company's contact person in Shanghai (available from 9:30 AM to 12:00 PM on March 11th).

[0117] Step 3: The demand identification module runs the demand weight calculation function (Formula 1) to calculate the demand weight:

[0118] Known user time optimization request frequency percentage f t =0.8, the urgency coefficient of official business e_b=1.0 (the senior management meeting is an urgent official business), a1=0.4, a2=0.6, then =0.4×0.8+0.6×1.0=0.92;

[0119] User budget limit B max =5000 yuan, average past travel expenses B avg =4000 yuan, the percentage of overspending frequency f e =0, β1=0.7, β2=0.3, then ω e =0.7×(5000-4000) / 5000+0.3×0=0.14;

[0120] f tr =0.9, the proportion of high-end accommodation selection frequency f_h=1.0, γ1=0.5, γ2=0.5, then ω_c=0.5×0.9+0.5×1.0=0.95;

[0121] The coefficients for importance of official duties are i_b=1.0, the coefficients for the level of participants are l_b=1.0, the coefficients for time rigidity are r_b=1.0, and δ1=δ2=δ3=1 / 3. Therefore, ω_b=(1.0+1.0+1.0) / 3=1.0.

[0122] After normalization, ω t =0.22, ω e ω'=0.03, ω_c'=0.23, ω_b'=0.52 (to ensure the sum is 1).

[0123] Step 4: The trip planning module runs the multi-objective optimization objective function (Formula 2') to generate an initial trip plan:

[0124] 1. Transportation Option: Choose Business Class on the G1 high-speed train (8:00-11:30), travel time 1.5 hours, cost 630 yuan; travel comfort rating: S. tr =95 points;

[0125] 2. Accommodation Option: Select a hotel in Shanghai (1 km from the conference venue), check-in on March 10th, check-out on March 11th, cost 800 yuan; accommodation comfort rating S_h = 90 points;

[0126] 3. Business Arrangements: Arrive at the meeting venue at 9:30 AM on March 11th, with 30 minutes for preparation (meeting requirements); Meeting time: 10:00 AM - 12:00 PM; Business coordination matching score M_b = 100.

[0127] 4. Other arrangements: Rest and free time are arranged on the afternoon of March 10; a business lunch is arranged on the afternoon of March 11; and the G2 high-speed train (14:00-17:30) will be taken back to Beijing on March 12.

[0128] 5. Total time T tot =2 days (48 hours), total cost C tot =630 (outbound) + 800 (accommodation) + 300 (food) + 630 (return) = 2360 yuan (≤5000 yuan);

[0129] 6. Comfort score S_c = 0.4 × 95 + 0.4 × 90 + 0.2 × 100 = 94 points, S_c' = 0.94;

[0130] 7. Normalized objective function value F'=0.22×(48-40) / (56-40)+0.03×(2360-1800) / (5000-1800)-0.23×0.94-0.52×1 .0=0.22×0.5+0.03×0.175-0.2162-0.52=0.11+0.00525-0.2162-0.52=-0.62095.

[0131] Step 5: The output interaction module displays the initial itinerary plan to the user in a visual timeline format, including information such as time, location, cost, and booking link for each stage; once the user confirms that there are no modification requests, the system generates a travel approval report and connects to the company's OA system for approval.

[0132] Step 6: During the trip, the data acquisition module monitors dynamic data in real time; if it is detected on the morning of March 10 that the G1 high-speed train is delayed by 2 hours due to weather (departing at 10:00 and arriving at 13:30), then calculate ΔF':

[0133] Original itinerary T arrive =11:30, delayed T arrive =13:30, resulting in insufficient preparation time (13:30 + 30 minutes = 14:00 > 10:00), M_b drops to 0 points, M_b' = 0; T tot Increased to 49.5 hours, T tot '=(49.5-40) / (56-40)=0.59375;C tot Unchanged, C tot '=0.175; S_c'=(0.4×95+0.4×90+0.2×(100-2×10)) / 100=(38+36+16) / 100=0.9;

[0134] F'_new=0.22×0.59375+0.03×0.175-0.23×0.9-0.52×0=0.130625+0.00525-0.207=-0.071125;

[0135] ΔF'=-0.071125-(-0.62095)=0.5498≥θ=0.1 (user sensitivity is high, θ is taken as 0.1), triggering the dynamic optimization algorithm.

[0136] Step 7: Dynamic optimization module replans the itinerary: Select flight MU5101 (departs at 7:30, arrives at 9:50), first class fare 1200 yuan; arrival time 9:50, pre-arrangement time 10 minutes (less than the required 30 minutes), then the coordinating personnel will adjust the meeting time to 10:30 (coordinating personnel's schedules match), pre-arrangement time 40 minutes, M_b=100 minutes; recalculate F'_new=0.22×(48-40) / (56-4 0)+0.03×(1200+800+300+630-1800) / (5000-1800)-0.23×0.94-0.52×1.0=0.11+0.03×0.353-0.2162-0.52=0.11+0.0106-0.2162-0.52=-0.6156, which meets the optimization requirements; the system will push the adjustment plan to the user, and after the user confirms, the itinerary will be updated and the flight booking and meeting time coordination will be completed.

[0137] 1.3 Module Function Verification

[0138] 1. Data Acquisition Module: Verification results show that the module can collect multi-source data in real time, with a data update delay of ≤5 seconds. The accuracy of traffic data and hotel data is ≥99%, and the accuracy of weather data and schedule data is ≥98%, meeting the data support requirements.

[0139] 2. Demand Identification Module: Tested on 100 users, the demand identification accuracy rate is ≥92%, of which the implicit demand identification accuracy rate is ≥85%, which is more than 40% higher than the existing technology.

[0140] 1. Trip Planning Module: The time to generate an initial trip plan is ≤10 seconds, the plan meets the constraints 100% of the time, and the user requirement matching degree is ≥90%.

[0141] 4. Dynamic optimization module: Response time for emergencies ≤30 seconds, accuracy of adjustment plan ≥95%, and smooth execution rate of adjusted itinerary ≥98%.

[0142] 5. Output interaction module: User operation convenience score ≥ 4.8 points (out of 5), and 100% success rate of business travel approval integration.

[0143] This invention achieves intelligent planning, precise matching, and real-time adjustment of business travel itineraries through the collaborative work of a data acquisition module, a demand identification module, an itinerary planning module, a dynamic optimization module, an output interaction module, and a data storage module, combined with a multi-dimensional demand identification model and a multi-objective dynamic optimization algorithm. This system overcomes the shortcomings of existing technologies, such as low demand matching accuracy, poor dynamic adaptability, imbalance in multi-objective optimization, and incomplete information integration. It boasts advantages such as high itinerary planning efficiency, high demand matching accuracy, strong dynamic adaptability, and low travel costs, and can be widely applied in scenarios such as corporate travel management and personal business travel planning, demonstrating significant practical application value and promising prospects for wider adoption.

[0144] The beneficial effects of the present invention are as follows, compared with the prior art:

[0145] 1. Improve itinerary planning efficiency: AI algorithms automatically identify needs, generate itineraries, and dynamically adjust them without human intervention, improving planning efficiency by more than 80% compared to manual methods; at the same time, the visual itinerary display and one-click booking function further shorten the itinerary confirmation and execution cycle.

[0146] 2. Improve the matching degree of needs: Through the mining of implicit needs and multi-objective optimization, the matching degree of itinerary plans with users' personalized needs and business activity needs has been improved by more than 90%, reducing the frequency of itinerary modifications due to mismatch of needs.

[0147] 3. Reduce travel costs: Through precise control of cost sensitivity weights and multi-objective optimization, the total travel cost is reduced by an average of 15%-25% compared to existing technologies, while meeting travel needs; at the same time, it reduces additional expenses caused by unreasonable itineraries (such as rescheduling fees and extra accommodation fees).

[0148] 4. Enhanced dynamic adaptability: In the event of an emergency, the dynamic optimization algorithm can generate the optimal adjustment plan within 30 seconds with an accuracy rate of over 95%, avoiding business delays or itinerary failures caused by emergencies and improving the smoothness of business trips.

[0149] 5. Optimize user experience: By optimizing comfort and business suitability, user satisfaction with business travel itineraries has increased by over 90%; at the same time, it has reduced the operational costs for users in itinerary planning and adjustment, and improved the convenience and comfort of business travel.

[0150] 6. Facilitates corporate travel management: The system can connect to the corporate travel approval system to generate standardized itinerary reports and expense details, enabling transparent management of travel expenses and reducing corporate travel management costs.

[0151] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention without departing from the spirit and scope of the present invention. Any modifications or equivalent substitutions should be covered within the scope of protection of the claims of the present invention.

Claims

1. An AI-assisted business travel itinerary creation system, characterized in that, include: The system includes a data acquisition module, a demand identification module, a trip planning module, a dynamic optimization module, an output interaction module, and a data storage module. Data acquisition module: used to collect multi-source heterogeneous data, including basic travel data, real-time dynamic data, user personalized data, and official business related data; Demand identification module: Based on AI algorithms, it performs in-depth analysis of the collected data to identify users' explicit and implicit needs; Trip planning module: Based on the demand vector output by the demand identification module and combined with the multi-source data provided by the data acquisition module, an initial travel itinerary plan is generated through a multi-objective optimization algorithm; Dynamic optimization module: Real-time monitoring of dynamic data acquired by the data acquisition module. When a sudden situation affecting the execution of the trip is detected, the dynamic optimization algorithm is triggered to adjust the existing trip plan and generate the optimal adjustment plan. Output interaction module: used to display itinerary plans to users, receive user feedback, and pass user feedback to the demand identification module or dynamic optimization module; Data storage module: Used to store collected multi-source data, user demand data, itinerary plan data, optimization log data and user feedback data, providing data support for the iterative optimization of AI algorithms.

2. The AI-assisted business trip creation system according to claim 1, characterized in that, In the data acquisition module, basic travel data includes departure point, destination, travel period, budget range, and other basic user input data.

3. The AI-assisted business trip creation system according to claim 2, characterized in that, In the data acquisition module, real-time dynamic data includes real-time traffic data, hotel real-time data, and destination real-time data.

4. The AI-assisted business trip creation system according to claim 3, characterized in that, In the data collection module, user-personalized data includes users' past business travel records, transportation preferences, accommodation standards preferences, dining preferences, and travel time habits; official business-related data includes official activity information, schedule data of the units they are in contact with, and business travel approval process data.

5. The AI-assisted business trip creation system according to claim 4, characterized in that, The demand identification module specifically includes building a demand identification model. This model takes basic user input data as the initial input, combines user personalized data and official business related data to perform feature extraction and demand inference, and outputs a demand vector containing dimensions such as official business priority, time sensitivity, cost sensitivity, comfort preference, and connection buffer needs.

6. The AI-assisted business trip creation system according to claim 5, characterized in that, In the itinerary planning module, the initial itinerary plan includes transportation plan, accommodation plan, official activity arrangement, meal and rest time planning, and cost details. It also meets the constraints of reasonable time connection of each link, cost not exceeding the budget range, and matching the priority of official activities.

7. The AI-assisted business trip creation system according to claim 6, characterized in that, Also includes: The demand weight calculation function is used to calculate the weights for time sensitivity, cost sensitivity, comfort preference, and priority, providing a basis for multi-objective optimization. ωt+ω e +ω_c+ω_b=1(1); in: ωt is the time sensitivity weight, with a value range of [0,1], and is calculated as ω t =a1×ft+a2×e_b, where a1 and a2 are weighting coefficients, ft is the proportion of user time optimization needs frequency, and e_b is the urgency coefficient of official business; ω e This represents the cost sensitivity weight, with a value range of [0,1]. ω_c is the comfort preference weight, with a value range of [0,1]. It is calculated as ω_c=γ1×ftr+γ2×f_h, where γ1 and γ2 are weight coefficients, ftr is the proportion of the frequency of choosing high-end transportation mode, and f_h is the proportion of the frequency of choosing high-end accommodation. ω_b is the priority weight of official duties, with a value range of [0,1]. It is calculated as ω_b=δ1×i_b+δ2×l_b+δ3×r_b, where δ1, δ2, and δ3 are weight coefficients, i_b is the importance coefficient of official duties, l_b is the level coefficient of the participants, and r_b is the time rigidity coefficient.

8. The AI-assisted business trip creation system according to claim 7, characterized in that, ω e The calculation method is ω e =β1×(Bmax-Bavg) / Bmax+β2×f e Where β1 and β2 are weighting coefficients, Bmax is the upper limit of the travel budget, Bavg is the user's average past travel expenses, and f e This represents the percentage of times expenses were exceeded.

9. The AI-assisted business trip creation system according to claim 8, characterized in that, Also includes: With the optimization objectives of minimizing total travel time, reducing total cost, maximizing comfort, and ensuring the best match between travel and official business, a weighted summation multi-objective optimization function is constructed: Min F=ω1×Ttot+ω e ×Ctot-ω_c×S_c-ω_b×M_b (2); in: F represents the objective function value to be optimized. Ttot is the total travel time, which includes the sum of time spent on transportation, business activities, transfers, rest, and other aspects of the trip. Ctot is the total cost of the business trip, which includes transportation, accommodation, meals, and business-related expenses. S_c is the comfort score, which is calculated by weighting the comfort of transportation, accommodation, and connection smoothness. S_c=λ1×Str+λ2×S_h+λ3×S_con, where λ1, λ2, and λ3 are weighting coefficients, Str is the transportation comfort score, S_h is the accommodation comfort score, and S_con is the connection smoothness score. M_b represents the matching degree of official business coordination, which is used to evaluate the degree of matching between the itinerary and the needs of official activities. M_b = μ1×M_p + μ2×M_t + μ3×M_s, where μ1, μ2, and μ3 are weighting coefficients, M_p is the score for the sufficiency of preparation time, M_t is the score for the accuracy of meeting time coordination, and M_s is the score for the matching degree of the schedule of the coordinating personnel. Constraints: Ctot≤Bmax; Tstart≤T_b e 9in; T arrive +T_p re ≤T_b e 9in; S_h≥S_hmin; T_w ait ≤T_wmax。 10. The AI-assisted business trip creation system according to claim 9, characterized in that, Also includes: The dynamic adjustment trigger threshold function determines whether to trigger travel adjustment when an unexpected situation occurs. ΔF'=F'_new-F'_old≥θ(3); in: ΔF' represents the change in the objective function value; F'_old represents the normalized objective function value of the original travel plan; F'_new represents the normalized objective function value of the original travel plan after the occurrence of the unexpected situation; θ represents the adjustment trigger threshold.

Images