Literature Review

1.Title: Design of An AI-Empowered Recommender System for Travelling Support

:Individual Traveler as an Instance

Authors: Kuan-Hua Lai, Neil Y. Yen, Mu-Yen Chen

Abstract:
There are many things, which are often time consuming, need to be prepared for individuals
who travel by themselves. There are lots of recommendation systems built to support such type
of travel. For instance, Skyscanner compares the fight ticket, recommends the cheapest ticket
and provides booking flight ticket service. This paper targets to support individual traveler
while arranging travel itinerary and travel related issues through a newly proposed intelligent
framework. This framework employs the techniques of artificial intelligence, collaborative
filtering, Term Frequency-Inverse Document Frequency, and optimization algorithms.
Systems, with support of our framework, may automatically suggest possible itineraries to
individual travellers, save time for making arrangement, and reach personalization.

Advantages:

 AI-based systems can tailor recommendations to individual preferences, past

behaviour, and contextual factors (e.g., location, weather, time).
 Travelers receive suggestions that align with their interests, reducing the time and effort
needed to plan trips.
 These systems leverage big data (e.g., user reviews, location data, social media) to
provide more accurate and relevant recommendations.

Disadvantages:

 Personalized systems often require access to sensitive personal data, raising privacy
and data protection issues.
 For new users or destinations with limited data, the system may struggle to provide
relevant recommendations.
 Algorithms may inadvertently reinforce biases (e.g., always suggesting popular
destinations, ignoring niche interests).
2 Title: Conceptual Integration of AI for Enhanced Travel Experience

Authors: Rajeev Semwal, Dr. Nandita Tripathi, Ajay Rana, Anubhav Chauhan, Krshnakant
Gupta

Abstract: This paper delves into the captivating fusion of "Tourism 3.0" and artificial
intelligence(AI), exploring how their intersection reshapes the landscape of travel experiences.
Positioned within the framework of "Tourism 3.0," a progressive paradigm that embraces
technology for enhanced travel, AI emerges as a transformative catalyst. This synergistic
relationship marries the principles of AI, including machine learning and natural language
processing, with the ethos of tourism theories focused on experience design and traveler
behavior. Through conceptual models, the paper demonstrates how AI can seamlessly offer
personalized travel recommendations, intelligent itinerary planning, and real-time assistance.
These models emphasize continuous traveler engagement that spans pre-trip anticipation to
post-trip reflection. As this narrative unfolds, the benefits of AI integration, such as heightened
customer satisfaction and operational efficiency, interweave with ethical considerations,
encompassing transparency, fairness, data privacy, and preserving meaningful human
interactions. Looking ahead, the paper envisions AI's evolution in "Tourism 3.0," extending to
AI-powered virtual travel assistants and immersive experiences, and underscores the
transformative potential of this symbiotic relationship. In essence, this exploration highlights
how AI is fundamentally reshaping travel within "Tourism 3.0," promising a future of
personalized, immersive, and ethically grounded travel journeys.

Advantages:

 AI integrates multiple services (flights, hotels, activities) into a single interface,

simplifying the planning process.
 AI-powered chatbots and virtual assistants offer 24/7 support, resolving queries in real
time with minimal human intervention.
 AI tools like real-time translators bridge communication gaps, enabling smoother
interaction for international travelers.

Disadvantages:

 AI systems often require access to personal and behavioral data, posing privacy and
cybersecurity concerns.
  Full integration demands advanced infrastructure and technical expertise, which may
not be available in all regions or companies.

3 Title: Multi-Task Travel Route Planning With a Flexible Deep Learning Framework

Authors: Feiran Huang , Jie Xu , and Jian Weng

Abstract :Travel route planning aims to map out a feasible sightseeing itinerary for a traveler
covering famous attractions and meeting the tourist’s desire. It is very useful for tourists to plan
their travel routes when they want to travel at unfamiliar scenic cities. Existing methods for
travel route planning mainly concentrate on a single planning problem for a special task, but is
not capable of being applied to other tasks. For example, previous must-visit planning methods
cannot be applied to the next-point recommendation, despite these two tasks are closely related
to each other in travel route planning. Besides, most of the existing work do not consider the
important auxiliary information such as Point of Interests (POI) attributes, user preference, and
historical route data in their approaches. In this paper, we propose a flexible Multi-task Deep
Travel Route Planning framework named MDTRP to integrate rich auxiliary information for
more effective planning. Specifically, we first construct a heterogeneous network through the
relations between users and POIs and employ a heterogeneous network embedding method to
learn the features of users and POIs. Then we present an attention-based deep model to integrate
the auxiliary information and focus on important visited points for the prediction of next POIs.
Finally, a beam search algorithm is introduced to flexibly generate multiple feasible route
candidates for three types of planning tasks (next-point recommendation, general route
planning, and must-visit planning). We introduce six public datasets to conduct extensive
experiments, of which the results demonstrate the flexibility and superiority of the proposed
approach in travel route planning.

Advantages:

 The framework can handle various factors simultaneously (e.g., time, cost, user
preferences), leading to more balanced and efficient travel routes.
 Deep learning models can learn complex patterns and adapt to dynamic changes such
as traffic, weather, or user feedback.
 A flexible framework can scale easily to support a wide range of user types,
destinations, and travel modes without major redesign.
Disadvantage:

 Deep learning models require significant computing power, especially for real-time or
large-scale route planning tasks.
 These systems require large volumes of high-quality data (e.g., maps, traffic, user
history) to function effectively, which may not always be available.
 Deep learning models often act as "black boxes," making it difficult to understand how
a particular route or decision was generated.

4. Title: Cutting-Edge Travel Planner: Intelligent Route Recommendation System using

Enhanced Learning Scheme with AI Principles

Author: N.R Gladiss Merlin, Anandakumar D, S.Devilavanya, P. K. Sheela Shantha Kumari,

and T.Jayakumar

Abstract : - A route planner is crucial as it optimizes travel efficiency, minimizes time and
fuel consumption, and enhances overall navigation convenience and safety. This paper presents
the design and implementation of CollabRouteNet, an intelligent route recommendation system
that leverages collaborative filtering and reinforcement learning principles. The proposed
model aims to provide personalized and contextually relevant route recommendations tailored
to individual user preferences and dynamic environmental conditions. Collaborative filtering
techniques are employed to analyze user interactions and discern patterns within historical
route data. This involves constructing a user-item matrix and applying matrix factorization to
learn latent representations for users and routes. In parallel, reinforcement learning is utilized
to optimize route recommendations in real-time by defining the problem as a Markov Decision
Process (MDP) and training an agent to learn an optimal route selection policy. The model
balances exploration and exploitation to adaptively recommend routes that optimize user
satisfaction and navigation efficiency. Implemented in PyCharm, the CollabRouteNet model
demonstrates promising results in providing accurate and responsive route recommendations.
Through its integrated approach, CollabRouteNet offers a promising solution for enhancing
navigation experiences in diverse urban and rural settings.

Advantages:

 Enhanced AI learning schemes allow the system to make smarter, context-aware route
decisions based on real-time data and user behaviour.
  Routes can be adjusted in real time based on traffic, weather, events, or last-minute user
changes, ensuring optimal travel experiences.
 Enhanced learning algorithms optimize for time, cost, convenience, and user
preferences, providing well-balanced routes with minimal effort from the user.

Disadvantages:

 Implementing advanced AI learning schemes requires high expertise and a well-

structured system architecture, increasing development time and cost.
 Collecting and processing personal travel data raises ethical and legal issues regarding
data protection and user consent.
 Advanced AI models may produce opaque decision-making processes, making it
difficult to explain why a certain route was suggested.

5. Title: Intelligent travel information platform based on location base services to predict user
travel behavior from user-generated GPS traces

Author: Qasim Ali Arain, Hina Memon, Imran Memon, Muhammad Hammad Memon, Riaz
Ahmed Shaikh & Farman Ali Mangi

Abstract :In recent past, social networking has received great attention and millions of geo
tagged photos are placed online on different social networks websites like licker, Facebook and
Instagram. According to this, people used to share their photos and share their travel experience
over social media. However, photos themselves has a lot of hidden information to share like
position, time etc. We presented a new technique related to location travel for tourist by looking
on to their time and preference needs. We used to acquire its preference based on his or her
past time in one city and based on this information recommend another city. We executed
certain experiments to support our technique and acquire data-set from publicly available licker
database. Simulation results have revealed that our travel recommendation system has been
devised according to user needs and outperformed other previous travel and recommendation
methods.

Advantages:

 By analysing user GPS traces, the system can learn travel patterns and make
personalized recommendations aligned with the user's habits and preferences.
 The platform can adapt dynamically to the user’s current location, providing timely
suggestions or alerts (e.g., nearby attractions, traffic conditions).
  Location-aware services enhance user convenience by automating decisions like route
changes, stop suggestions, and travel duration estimates.

Disadvantages:

 Constant tracking of user location raises serious concerns about user consent, data
misuse, and potential breaches of personal privacy.
 Continuous GPS tracking can lead to significant battery consumption and data usage,
which may frustrate users.
 Real-time LBS platforms often require stable internet connections and GPS signal,
which may not be available in remote or rural areas.

6.Title-Travel Itinerary Generation using AI and recommendation systems

Author: Saransh Aggrawal , Dhruv Teotia, Yash, Seema Shukla

Abstract: The rapid growth of artificial intelligence (AI) in the travel and tourism sector has
revolutionized the way how tourists plans their trips. Personalized travel suggestions are
provided by Al-powered systems, which streamlines the usually laborious process of
organizing trips. This paper examines the integration of machine learning, natural language
processing, and recommendation systems with a focus on Al-based travel itinerary generators.
These programs can provide personalized Itineraries by analyzing user preferences, including
time, money, and hobbies. To improve personalization, methods including sentiment analysis
(Logistic Regression), content and collaborative filtering, and K-Means clustering are used.
Based on user's input, Point of Interests (POIs) are filtered and ranked. Time slots are assigned
within the itinerary and ordered to minimize travel time using K-means clustering algorithm.
A hybrid of collaborative and content filtering is applied based on user feedback on which
sentiment analysis is performed using logistic regression. The final itinerary is generated that
includes suggested POIs with suggested start and end times. This paper illustrates how current
developments in AI-driven travel solutions might improve user contentment and simplify travel
arrangements. Notwithstanding these developments, issues like data privacy and
personalization biases still need to be addressed in order to guide future studies.

Advantages:

 Personalized Planning – Generates tailored itineraries based on user preferences like

budget, interests, and dates.
  AI Integration – Uses NLP, sentiment analysis, and ChatGPT for smarter, context-aware
recommendations.
 User-Friendly Interface – A web-based interface simplifies input collection and
itinerary presentation.

Disadvantages:

 Data Dependency – Accuracy depends on the quality and completeness of user input
and dataset.
 Algorithm Limitations – The system may struggle with ambiguous or conflicting
preferences.
 No Real-Time Updates – Static recommendations may not reflect live travel updates
like weather or availability.

7 Title- Hybrid Artificial Intelligence Based Automatic Determination of Travel Preferences

of Chinese Tourists

Authors- Ibrahim Topal AND Muhammed Kurad Ucar

Abstract :Tourism and travel sector continues to grow by gaining an important place in the
world economy and many countries want to increase their share in this sector. At the same
time, it is known that todays consumer tourism and travel purchase decisions are influenced by
social media. By examining the data of consumers on social media, it is possible for businesses
to reach the right person and get more efficiency from high-cost promotion activities. The study
aims to analyse the historical data of users on TripAdvisor with artificial intelligence methods
to reveal a pro le of consumers who might prefer Turkey. Methods: In this context,
TripAdvisor, which is one of the best-known websites in the tourism sector, is an important
source of data for countries to increase their share in the tourism market. Inferences can be
made by using artificial intelligence methods and the data in TripAdvisor together. In this
study, as a case study, the potentials of Chinese tourists to prefer Turkey are dealt because
Turkey has increased its tourism targets ten folds for China and the year 2018 was declared as
Turkey Tourism Year in China. In this context, this study aims to determine the potentials of
Chinese tourists to prefer Turkey, by processing travel data histories obtained from
TripAdvisor with artificial intelligence methods. It is expected that the study will contribute to
the tourism sector as well as the academic literature. The study used the travel data history of
Chinese tourists taken from TripAdvisor. Significant travel histories were selected by the F-
score method. Depending on the selected and all travel histories of users, their travel
preferences (Turkey/France) were classified by artificial intelligence algorithms. The
developed model was tested with performance criteria. Results: At the end of the study, it was
ensured that the Chinese, who would prefer Turkey, were determined with an accuracy rate of
75.25% and sensitivity rate of 0.76. Conclusions: It was observed that it is possible to and the
tourists who will prefer Turkey by using the developed system. In other words, the study
revealed that the countries can reach the individual instead of masses in their promotional
activities.

Advantages:

 High Prediction Accuracy-The system determines Chinese tourists likely to prefer

Turkey with a 75.25% accuracy, which is considered satisfactory for tourism
applications.
 Targeted Marketing Efficiency-It enables reaching individual tourists instead of using
broad, costly promotional campaigns—saving marketing resources and improving
effectiveness.
 Robust AI Integration-Multiple machine learning models (like SVM, kNN, Neural
Networks) and ensemble methods are used for better adaptability and performance
across different data types.

Disadvantages:

 Accuracy Below Industry Ideal-While 75.25% is acceptable in tourism, it falls short of

the 80%+ benchmark desired in decision-critical fields like healthcare.
 Single-Source Data Limitation-The model relies solely on TripAdvisor data, which
may not capture the full spectrum of tourist behavior or preferences.
 Limited Feature Diversity-The model uses only past travel history to predict
preferences, excluding key factors like demographics, financial status, and personal
interests—limiting prediction depth.

8 Title- Intelligent Systems in Travel and Tourism

Authors-Hannes Werthner
Abstract- Travel and tourism is the leading application field in the b2c e-commerce, it
represents nearly 50% of the total b2c turnover. Already in the past travel applications were at
the forefront of Information Technology, i.e., the airline Computerized Reservation Systems in
the early 60s. The industry and its product have rather specific features which explain this
circumstance: the product is a confidence good, consumer decisions are solely based on
information beforehand; and the industry is highly networked, based on world-wide
cooperation of very different types of stakeholders. Consequently, this industry depends on
advanced IT applications. As such travel and tourism may serve as an example of what happens
and will happen in the emerging e-markets, pointing at structural changes as well as challenging
application scenarios. The paper provides an overview about the industry, describes ongoing
structural changes, outlines domain-specific requirements and discusses achievements and
challenges in the field, following an AI and e-commerce point of view. It finishes with
considerations regarding a future IT scenario.

Advantages:

 Enhanced Personalization: Intelligent systems allow users to customize travel

experiences based on individual preferences and decision styles.
 Efficient Service Integration: Intelligent platforms can bundle services (flights, hotels,
events) for a seamless experience.
 Market Transparency: Travelers benefit from easier price comparison and access to
real-time information via intelligent tools.

Disadvantages:

 Data Standardization Issues: Lack of universal data formats in tourism complicates

system integration and interoperability.
 Complex User Modelling: Accurately capturing user preferences and behaviours for
meaningful recommendations remains challenging.
 Mobile Limitations: Mobile devices still face constraints in interaction design and
performance compared to full web platforms.
9.Title- Personalised Travel Ratings using AI and Enhancing User Experience Through Multi-
Dimensional Analytics

Authors-Poornima S, A Krishnaveni, J Relin Francis Raj, R Divya, G Vinoth Rajkumar, I

Sakthidevi

Abstract: The rapid advancements in Artificial Intelligence (AI) have enabled the
development of intelligent systems for personalized travel recommendations. This study
proposes a novel framework, Generative Adversarial Networks with Real Time Analytics for
Travel Experience (GAN-RATE), designed to enhance user satisfaction by integrating
synthetic data generation and real-time feedback mechanisms. The GAN component addresses
the challenge of sparsely rated travel attributes by generating high-quality synthetic data, while
the Real-Time Analytics module evaluates user interactions and feedback to refine the
recommendation process dynamically. The proposed framework considers multiple
dimensions, including comfort, cleanliness, timeliness, entertainment amenities, and customer
feedback, ensuring a holistic approach to travel rating prediction. Simulation analysis is
conducted to compare the performance of GAN-RATE with existing algorithms. Metrics such
as accuracy, mean absolute error (MAE), precision, recall, and computational efficiency are
employed to evaluate system performance. Results indicate that GAN-RATE outperforms
baseline models in prediction accuracy and adaptability to sparse datasets while demonstrating
superior real-time response capabilities. The findings highlight the potential of GAN-RATE in
improving user experience through personalized, data-driven recommendations, setting a new
benchmark for AI-based travel recommendation systems.

Advantages:

 Improved Accuracy: The GAN-RATE framework significantly improves prediction

accuracy, reducing mean absolute error (MAE) compared to existing models.
 Real-Time Adaptability: The system dynamically updates recommendations based on
user interactions, providing faster and more relevant results.
 Handles Sparse Data: GANs generate synthetic data for poorly rated attributes,
addressing the common problem of missing or sparse user feedback.
Disadvantages:

 Complex Architecture: The integration of GANs and real-time analytics increases

system complexity, requiring careful tuning and resource management.
 High Computational Cost: Training GANs and processing real-time feedback
simultaneously demand significant computational resources.
 Data Dependency: The effectiveness of recommendations still relies on the quality and
diversity of initial user data and feedback.

10. Title- Research on Personalized Travel Route Recommendation AI Algorithm Based on

Simulated Annealing

Authors Jingchao Leng, Xue Yu

Abstract: In order to improve the insufficient diversified demand for tourist itinerary selection,
a travel itinerary selection method based on consumer needs is provided. By collecting scenic
spot images and travel time shared by consumers and considering the popularity of the scenic
spot, consumer preferences, travel time, and consumer requirements, highly rated locations can
be recommended to consumers. On this basis, a travel path selection model is constructed to
integrate cost and time constraints. Based on this model, a simulated annealing solution is
proposed with embedded chaotic perturbation and variance determination rules. Empirical
evidence shows that this method meets the travel needs of different consumers, enhances the
authenticity of tourist attractions selection, and provides new methods and means for selecting
travel routes.

Advantages:

 User-Centered Personalization: The model considers user preferences, travel time, and
costs to generate tailored travel routes.
 Enhanced Optimization: The improved simulated annealing algorithm with chaotic
perturbation increases accuracy and avoids local optima.
 Higher Recommendation Accuracy: Compared to existing methods (like TRR), it
achieves better accuracy, recall, and F-measure values for route selection.
Disadvantages:

 Increased Algorithm Complexity: Embedding chaotic mechanisms and variance criteria

makes the system more complex to implement and tune.
 High Data Dependency: The approach relies heavily on detailed and accurate user input
and travel data for effectiveness.
 Limited Real-Time Adaptation: The method is designed for pre-planning and may not
adapt well to dynamic changes during travel.

11.Title- Travel Demand Forecasting: A Fair AI Approach

Authors- Xiaojian Zhang, Qian Ke, Xilei Zhao

Abstract- Artificial Intelligence (AI) and machine learning have been increasingly adopted for
travel demand forecasting. The AI-based travel demand forecasting models, though generate
accurate predictions, may produce prediction biases and raise fairness issues. Using such biased
models for decision-making may lead to transportation policies that exacerbate social
inequalities. However, limited studies have been focused on addressing the fairness issues of
these models. Therefore, in this study, we propose a novel methodology to develop fairness-
aware, highly-accurate travel demand forecasting models. Particularly, the proposed
methodology can enhance the fairness of AI models for multiple protected attributes (such as
race and income) simultaneously. Specifically, we introduce a new fairness regularization term,
which is explicitly designed to measure the correlation between prediction accuracy and
multiple protected attributes, into the loss function of the travel demand forecasting model. We
conduct two case studies to evaluate the performance of the proposed methodology using real-
world ridesourcing-trip data in Chicago, IL and Austin, TX, respectively. Results highlight that
our proposed methodology can effectively enhance fairness for multiple protected attributes
while preserving prediction accuracy. Additionally, we have compared our methodology with
three state-of-the-art methods that adopt the regularization term approach, and the results
demonstrate that our approach significantly outperforms them in both preserving prediction
accuracy and enhancing fairness. This study can provide transportation professionals with a
new tool to achieve fair and accurate travel demand forecasting.

Advantages:

 Fairness Across Multiple Attributes: The proposed method corrects prediction

disparities for several protected groups (e.g., race, income, education) simultaneously.
  Model-Agnostic Flexibility: The fairness correction can be applied to any AI model
(e.g., MLP, GRU, STGCN) without modifying the model architecture.
 Preserved Prediction Accuracy: The fairness-aware method enhances fairness while
maintaining—or only slightly reducing—forecasting accuracy.

Disadvantages:

1. Complex Implementation: Incorporating fairness regularization and tuning

hyperparameters (like λ) increases model complexity.
2. Limited Generalization to All Travel Modes: The study focuses only on ridesourcing
data; it may not generalize directly to transit or other travel modes.
3. Assumption of Linear Correlation: The fairness metric assumes linear relationships
between prediction error and protected attributes, which may not always hold true

12.Title- Travel With Generator AI: A Novel Approach to Itinerary Creation

Authors- Mishka Gupta, Rish Dias, Netra Jain, Ruhani Rai Dhamija, Ranjeet Vasant Bidwe,
Sashikala Mishra, Ganesh Deshmukh

Abstract- The usual problems that arise are in the efficiency of itinerary planning of the
Travelers. The problem with all travel options and destinations is the challenge of organizing
a personal itinerary according to personal preferences, constraints, and interests. Traditional
methods of itinerary planning are both inflexible and time-consuming, which produces many
generic recommendations that do not take into consideration the preferences of the travelers or
address the comprehensive travel experience. These problems are solved by our app: it
incorporates user preferences into the itinerary generation, allowing the input of user
preferences by the users. Machine learning algorithms then analyze these preferences and
generate personalized recommendations based on them. This way, each itinerary is specifically
tailored to your needs, making the experience that much more enjoyable and fulfilling.

Advantages:

 Personalized Itineraries: The system generates travel plans tailored to user preferences,
such as budget, interests, and constraints.
 User-Friendly Interface: The app design includes simple navigation, interactive forms,
and accessible features that improve usability.
  Faster and More Detailed Results: Compared to general-purpose tools like ChatGPT,
the app delivers more structured and detailed itineraries.

Disadvantages:

 Subscription Requirement: Full itinerary access requires a ₹500 subscription, limiting

features for free users.
 API Rate Limiting Issues: External services (like OpenAI, Claude) have call
limitations, which can affect system responsiveness.
 Complex Backend Management: Managing multiple APIs, secure payment processing,
and user data requires robust infrastructure and maintenance.

13 .Title- AI-Driven Personalized Travel Planning: Enhancing Tourist Experiences in

Uzbekistan

Authors- Dr. Manik Arora, Dr. Naina Chaudhary, Dr. Mohit Bhandwal, Dr. Tanveer Baig, Dr.
Pratap Patil

Abstract- The tourism sector in Uzbekistan, an ancient Silk Road hub replete with historical
landmarks and rich cultural traditions, is primed for transformation through Artificial
Intelligence (AI). This paper presents an in-depth analysis of the applications and implications
of AI-driven personalized travel planning focused specifically on enhancing the overall tourism
experience in Uzbekistan. Drawing from existing AI models and a comprehensive
understanding of Uzbekistan’s unique travel offerings, a bespoke framework is introduced.
This framework integrates travel choices, historical data, and real-time local context to craft
customized itineraries for tourists. This research demonstrates that AI driven approaches not
only streamline travel processes but also amplify cultural immersion and appreciation.

Advantages:

 Personalized and Dynamic Itineraries: AI adapts plans in real time using user
preferences and local events, providing tailored and immersive travel experiences.
 Enhanced Tourist Satisfaction: 90% of beta testers preferred AI-generated itineraries
over traditional ones due to their relevance and detail.
 Cultural and Historical Integration: The system promotes cultural immersion by
suggesting lesser-known attractions aligned with user interests.
Disadvantages:

 Privacy and Data Use Concerns: Despite safeguards, users expressed a need for greater
transparency in how their personal data is used.
 Dependence on High-Quality Data: The system relies heavily on accurate, diverse data
inputs; poor data can reduce recommendation quality.
 Implementation Complexity: Integrating multiple data sources, real-time updates, and
ethical constraints increases system complexity and development effort.

14.Title- Travel Agent: An AI Assistant for Personalized Travel Planning

Authors- Aili Chen, Xuyang Ge, Ziquan Fu, Yanghua Xiao, Jiangjie Chen

Abstract- As global tourism expands and artificial intelligence technology advances,

intelligent travel planning services have emerged as a significant research focus. Within
dynamic real world travel scenarios with multi-dimensional constraints, services that support
users in automatically creating practical and customized travel itineraries must address three
key objectives: Rationality, Comprehensiveness, and Personalization. However, existing
systems with rule-based combinations or LLM-based planning methods struggle to fully satisfy
these criteria. To overcome the challenges, we introduce TravelAgent1, a travel planning
system powered by large language models (LLMs) designed to provide reasonable,
comprehensive, and personalized travel itineraries grounded in dynamic scenarios.
TravelAgent comprises four modules: Tool-usage, Recommendation, Planning, and Memory
Module. We evaluate TravelAgent’s performance with human and simulated users,
demonstrating its overall effectiveness in three criteria and confirming the accuracy of
personalized recommendations.

Advantages:

 Highly Personalized Itineraries: TravelAgent uses large language models (LLMs) and
memory modules to tailor recommendations based on real-time inputs and user history.
 Comprehensive Travel Planning: It integrates tools for flights, hotels, attractions, and
budget planning—covering all major aspects of a travel itinerary.
 Proven Effectiveness: Outperforms baseline models (like GPT-4+) in rationality,
comprehensiveness, and personalization through both human and simulated user
evaluations.
Disadvantages:

 Data Reliability Dependency: The system relies on real-time external data sources,
and inaccuracies (e.g., outdated flight info) can affect the quality of recommendations.
 System Complexity: Managing multiple modules (tool usage, memory,
recommendation, planning) increases implementation and maintenance effort.
 Limited Personalization Scope: Current personalization may not capture deeper user
needs or support all activity types, requiring future enhancement.

15. Title- A Multi-Criteria Multi-Modal Predictive Trip Planner: Application on Paris

Metropolitan Network

Authors-Pascal Benchimol, Ahmed Amrani, Mosthepha Khouadija

Abstract- Public transport route planning is of growing interest in smart cities and especially
in metropolitan areas where congestions and traffic jams are frequently recorded. The
availability of multiple data sources, such as passenger load in trains or ticketing logs, provides
an interesting opportunity to develop decision support tools to help passengers better plan their
trips around the city and to enhance their travel experience. We present, in this paper, a multi-
criteria journey planner that incorporates train load predictions as criteria. To this end, on the
one hand, we enrich the proposed routes with predictive indicators of passenger flow such as
the load on board the trains. These indicators are computed for each section of the itinerary
using machine learning algorithms. On the other hand, we design a journey planner that
incorporates the predicted load in its search criteria.

Advantages:

 Multi-Criteria Optimization: The planner considers various factors such as time,

comfort, and number of transfers, enabling more user-relevant route suggestions.
 Real-Time Prediction: Incorporates predictive modeling to estimate delays and improve
route accuracy before travel.
 Multi-Modal Integration: Combines multiple transport modes (e.g., metro, bus) into a
single route, improving travel flexibility across the network.
Disadvantages:

 Complex System Design: Handling multi-criteria and multi-modal data increases

system complexity and computational demand.
 Dependence on Data Quality: Accurate predictions rely heavily on the availability and
reliability of real-time and historical transit data.
 User Preference Modeling Limitations: Limited personalization; the system may not
fully capture individual user priorities or adapt dynamically during trips.

16. Title- E-Tourism: Mobile Dynamic Trip Planner

Authors- Hamzah Alghamdi , Shiai Zhu and Abdulmotaleb El Saddik

Abstract- —In this paper, we propose an algorithm called the Balanced Orienteering Problem,
to design trips for tourists. This algorithm, combined with a recommender system for tourism
suggestions, create the infrastructure for the mobile application of the tourism guide we
developed. A comparison study between some of the current algorithms and our proposed one
were performed and the initial results illustrate that our proposed algorithm yields comparable
results to existing system, yet it outperforms them in the average execution time.

Advantages:

 Efficiency: The Balanced Orienteering Problem (BOP) algorithm reduces execution

time by 5x compared to traditional methods, making it suitable for mobile devices with
limited resources.
 Personalization: The system combines user preferences with POI popularity, offering
tailored trip recommendations that adapt to weather conditions and real-time changes.
 Scalability: Clustering POIs into walkable areas simplifies computation and enables
dynamic re-planning during trips.

Disadvantages:

 Lower Precision: BOP sacrifices slight accuracy (1% drop in F1-score) for speed,
potentially missing some optimal POIs.
 Data Dependency: Reliance on external datasets (e.g., Flickr metadata) may limit
functionality in areas with sparse geotagged data.
 Initial Setup: Requires offline processing (e.g., clustering, database updates), delaying
real-time use for new locations.
17. Title- TripPlanner: Personalized Trip Planning Leveraging Heterogeneous Crowdsourced
Digital Footprints

Authors: Chao Chen, Daqing Zhang, Bin Guo, Xiaojuan Ma, Gang Pan, and Zhaohui Wu

Abstract: —Planning an itinerary before traveling to a city is one of the most important travel
preparation activities. In this paper, we propose a novel framework called TRIPPLANNER,
leveraging a combination of location-based social network (i.e., LBSN) and taxi GPS digital
footprints to achieve personalized, interactive, and traffic-aware trip planning. First, we
construct a dynamic point-of-interest network model by extracting relevant information from
crowdsourced LBSN and taxi GPS traces. Then, we propose a two-phase approach for
personalized trip planning. In the route search phase, TRIPPLANNER works interactively with
users to generate candidate routes with specified venues. In the route augmentation phase,
TRIPPLANNER applies heuristic algorithms to add user’s preferred venues iteratively to the
candidate routes, with the objective of maximizing the route score while satisfying both the
venue visiting time and total travel time constraints. To validate the efficiency and
effectiveness of the proposed approach, extensive empirical studies were performed on two
real-world data sets from the city of San Francisco, which contain more than 391900 passenger
delivery trips generated by 536 taxis in a monthand110214check-ins left by 15680 Foursquare
users in six months.

Advantages:

 Dynamic Traffic Awareness: The system leverages real-time taxi GPS data to estimate
time-varying transit times between POIs, ensuring routes are optimized for current
traffic conditions.
 Personalization: Combines user preferences (from check-in history) with POI
popularity to tailor recommendations, enhancing relevance for individual users.
 Scalability: Uses a two-phase approach (route search + augmentation) to efficiently
handle large datasets, making it practical for city-scale trip planning.

Disadvantages:

 Data Dependency: Relies heavily on external data sources (e.g., Foursquare check-ins,
taxi GPS), which may be sparse or unavailable in certain regions.
 Complexity: The NP-hard nature of route augmentation increases computational
overhead, though mitigated by heuristics.
  Limited Adaptability: Assumes static user preferences and stay times at POIs, lacking
real-time adjustment during trips (e.g., unexpected delays or changing interests).

18.Title-Design and Implementation of an Open Data Assisted Real-Time Trip Planner

Authors-Tien-Yu Chu, Kun-Che Hsu, Jenq-Shiou Leu

Abstract-Since the government data is credible, plentiful, and relevant to citizen, the
government disclosures part of its data as government open data to stimulate the development
of public applications [1]. The trip planning application providing as the best path for the trip
[2] draws the most attention for people. In this article, we proposed a real-time trip planner
system assisted with government open data, and presented a path selection scheme based on
the provided data. Besides, we designed different user interfaces- WebUI and AppUI for
stationary users and mobile users, respectively. Based on the requirements from users, the trip
planner system would reply all the candidate paths in a descendant traveling time order back
for users’ references.

Advantages:

 Real-Time Information Integration: Incorporates live bus data from government open
data sources to provide accurate travel and wait time estimates.
 Multi-Interface Accessibility: Offers both Web and App user interfaces, catering to
different user devices and preferences.
 Enhanced Path Planning Options: Supports direct, transfer, and transfer-after-walk
paths, improving flexibility and relevance of suggested routes.

Disadvantages:

 Limited to Bus Transit Only: Current system supports only bus data; integration with
other transit modes like trains or MRT is planned but not yet implemented.
 Data Update Interval Constraints: Real-time data updates every 30 seconds, which may
still lag behind actual transit conditions in fast-changing environments.
 No User Personalization: The system does not incorporate user preferences or historical
behaviour into path selection, limiting personalized recommendations.
19.Title-Budget and experience based travel planner using collaborative filtering

Authors-Gokul Krishna M, Mohammed Haseeb, Mohammed Siyad B, P.A Mohamed Zameel,

S Vyshnav Raj

Abstract-Travellers in the modern era face many challenges. Budget and time are the major
constraints for a traveller. So a good travel plan should be satisfying these constraints and
provide maximum enjoyment to the traveller. The proposed ’Budget and Experience-based
Travel Planner’ allows a traveller to create a travel plan that satisfies the above-mentioned
constraints as well as other factors such as the experience of the traveller. The collected data
from different sources are filtered and places that have missing fields that are relevant such as
tags, district, coordinates, rating, reviews, etc are removed along with places having irrelevant
tags. After filtering, the database contains over 600 destinations with 77 unique tags. Using
suitable machine learning algorithms considering the constraints, the recommendation engine
recommends places. We have created a routing algorithm which minimizes the distance
covered to reach the different destinations considering the various cases that could happen.
Suitable algorithms are used for scheduling the travel itinerary. The time, budget and
destination are provided by the user and a suitable travel plan is provided back to the user.

Advantages:

 Personalized Recommendations: Uses collaborative filtering and user preferences to

suggest travel destinations suited to individual interests and budget.
 Optimized Routing: Implements a routing algorithm that minimizes travel distance,
enhancing efficiency and saving time.
 Budget and Time Constraints Handling: Integrates user-input budget and time to
generate realistic and feasible itineraries.

Disadvantages:

 Geographic Limitation: The system is currently limited to destinations within Kerala,

reducing broader applicability.
 Cold Start Problem: New users without prior data receive less accurate
recommendations due to lack of historical preferences.
 Data Quality Dependency: Relies heavily on scraped data from Google Travel and
Maps, which may introduce errors or inconsistencies.
20.Title-Intelligent Trip Planning with Integrated Street View: A Seamless AI-driven
Approach

Authors-Siva Sankar A, Nirmal Kumar K, Sidharth M Dinesh, S Abhishek, Anjali T

Abstract-The advent and widespread adoption of AI-powered trip planners have revolutionized
how travelers engage in travel planning. These intelligent systems leverage artificial
intelligence algorithms and data to provide users with personalized, efficient, and immersive
planning experiences. In this research the indispensable role played by immersive street view
experiences within AI trip planners. Integrating immersive street view experiences within AI
trip planners empowers travelers to make well-informed choices, as they can virtually navigate
and assess potential destinations, accommodations, and attractions. These intelligent systems
leverage artificial intelligence algorithms and data to provide users with personalized, efficient,
and immersive planning experiences. The findings reveal that immersive street view
experiences significantly elevate the effectiveness of travel planning and decision-making
processes.

Advantages:

 Immersive Street View Integration: Users can virtually explore destinations through
360° images, enhancing decision-making and engagement.
 Personalized Itinerary Generation: The system uses AI and collaborative filtering to
recommend destinations and create optimized trip schedules.
 Efficient Budget Optimization: The planner considers user constraints to allocate
budget effectively across travel, accommodation, and activities.

Disadvantages:

 Limited Customization Options: User feedback indicates the need for more flexible
preferences and settings in itinerary customization.
 Reliance on External APIs: Heavy dependence on APIs (OpenAI, Google Street View)
may lead to performance or access issues.
 Scalability and Real-Time Data Integration: Current implementation may face
challenges when scaling up or incorporating live travel data and updates.