Under The Hood
An ideal trip planner should plan everything from destinations to sightseeing to activities to accommodation to transport all together and that too based on your personal taste. Why, almost all the websites talk about trip planning yet they offer much less than a trip planner!
When somebody tells you about holiday planning, you can come up with hundreds of websites that can help you do that. But if you think carefully, you’ll find that the holiday planning is much more than just travel guides, hotel reviews and the flight bookings.
We talk about AI advancements and fear about the singularity yet a true trip planner is hard to find. Why?
The reason is, planning a trip is one of the hardest decision-optimization problems and falls in the NP-hard category.
We, at Tripigator, have come up with an Intuitive Planning Engine (IPE) which heuristically solves the Trip Planning problem considering more than 50 major parameters including weather, budget, popularity, reviews, transport, comfort, duration etc.
Finding destinations
Every component, be it a destination, an activity, a hotel, a route etc., is viewed as a feature vector which keeps on updating at a regular interval.
When you enter your preferences, IPE uses fuzzy matching to come up with a list of prospective destinations and creates multiple sets. It may sound straightforward but it takes care of the feasibility as well. Consider the case when you are looking for a 3-day plan and it shows a trip to some place 1000kms from your starting location and makes it a hectic one. And what if you have pockets to spend money on flight tickets then despite the place to be far away, it may come out to be one of the best trips. However, assuming you have a budget for airline travel, even a short trip can turn out to be very fulfilling, and in fact, one of the best trips you take.
Allocating appropriate time at destinations
Based on the preferences you entered, the list of activities & sightseeings of your interests, the optimal duration at each of the places is allocated. Hopping from one place to another requires time and energy. Whether you want a relaxed trip or an active one, the system keeps a sufficient buffer time to make your holiday nothing less than the best one.
Creating routes
Finding the optimal route from place A to place B is always tough. It’s a function of price, duration, departure time, mode of transport and many more. And in case, if direct connectivity is not there, you may need to consider the transit timings. You need to keep a sufficient buffer time to choose the next option for that matter. IPE is designed in a way that it keeps a sufficient buffer and calculates the optimal routes considering all these parameters.
But wait, it’s not done yet. What if you go to a place C from B and you don’t find a good transport option? What if the option to go to C first, followed by B could lead to a better trip? Our greedy-tracking approach makes a good call here and except in a few rare instances, it turns out to be the best one. To know about the complexity of this problem read here.
Churning out sightseeing & activities at every destination
Once information of your arrival and departure at each of the destination, the mode of transport and the duration is known, it knows how tiring your journey would be, therefore, how much time you need to relax before you start exploring the place. Then, IPE allocates appropriate activities, sightseeings to every destination based on your interests making it more rejuvenating.
Selecting accommodation
All the ratings & reviews, dynamic pricing and availability are taken care of to find you a best hotel matching your budget. Whether you arrive early morning or you are on a short visit, IPE knows whether you would need an accommodation or not.
Stitching the pieces together to make every plan close to the perfect one
Once everything is in place, all the pieces are stitched together and itineraries are created. The experience rating is calculated based on the input preferences. A few trips are rejected if they do not qualify a certain threshold score.
Our system takes care of most of the user-personas and is being improved on a regular basis. We may not be 100% there but we are not far behind. If you want to share some ideas, throw it in comments.
