Tripsnek's optimizer is peer-reviewed work, published at GECCO 2026 — the Genetic and Evolutionary Computation Conference
Most automated trip planning scores destinations from crowd-sourced reviews. That has two problems. Reviews measure satisfaction relative to expectation, so a pleasant surprise can outrank a world-class sight. And reviews cannot express overlap: the châteaux of the Loire each deserve their rating individually, but the fifth one you tour teaches you much less than the first.
Tripsnek takes a different approach. It encodes the judgment of a recognized expert — Rick Steves — directly into the optimizer: how good each destination is, how many nights it deserves, and which places offer overlapping experiences. The paper describes that model, the memetic algorithm built on it, and the experiments validating both.
Destinations are rated and duration-scored against a consistent standard by someone who has assessed all of them — rather than averaged from reviewers who each saw one.
Cities within a destination share its reward, so "French Riviera: three nights" is one decision. Without this, an optimizer games the scoring by scattering single nights across neighboring towns.
The search is constrained to geographically sensible moves and pushed toward expert-recommended stay lengths — so routes stay smoothly paced with minimal backtracking, and good places aren't shortchanged.
The published work covers the optimization core. What the site adds on top of it matters just as much in practice: knowing what you need to lock in before you leave. Tripsnek flags which sights, transport, and rooms on your itinerary sell out or get more expensive, and how far ahead each one has to be booked. It also works out when to visit what — sequencing your days around opening hours, closure days, and the times worth avoiding, based on your actual travel dates.
Beyond that: budget and weather estimates for the trip you've built, and fully editable itineraries you can save, share, and export.
See the how-to guide for a tour of what the planner can do, about Tripsnek for why it was built this way, or browse sample itineraries and city-to-city routes.
The paper is open access under a Creative Commons Attribution license. If it is useful to your own work, please cite it:
@inproceedings{tuohy2026tripsnek,
author = {Tuohy, Daniel R.},
title = {TripSnek: A Memetic Algorithm for Expert-Guided Multi-Week European Travel Itinerary Optimization},
booktitle = {Proceedings of the Genetic and Evolutionary Computation Conference Companion},
series = {GECCO Companion '26},
year = {2026},
location = {San Jose, Costa Rica},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
doi = {10.1145/3795101.3805316}
}All of it runs live, for free, in your browser. Build a custom itinerary from your own dates and cities, or browse pre-generated examples.