arXiv 2510.21329

TripTide: A Benchmark for Adaptive Travel Planning under Disruptions

By Priyanshu Karmakar, Soumyabrata Chaudhuri, et al.

Published 2025-10-24

Mindmap

Browse the paper's core ideas, clusters, and relationships in a structured outline.

Recent efforts like TripCraft and TravelPlanner have advanced the use of Large Language Models ( LLMs) for personalized, constraint aware travel itinerary generation. Yet, real travel often faces disruptions. To address this, we present TripTide, the first benchmark evaluating LLM's ability to revise itineraries under realistic disruptions. TripTide models key dimensions such as disruption severity and traveler tole…

View the original paper on arXiv