System for Selecting Tourist Routes Based on Genetic Programming

Abstract

One of the most important tasks in tourism is helping tourists plan a trip to a given place. Since it is impossible to visit all places, tourists try to be rational and choose what they find most acceptable and attractive. Each plan has limitations (e. g., length of tour, limited budget) and preferences (e. g., art, culture, historical sites, architecture, modernism) that should be considered when planning your travel trip. This is a special case of the Orientation problem — Tourist trip design problem. The goal is to maximize the total score achieved within a given tour duration limit. This paper focuses on developing a recommendation system that considers users limitations during tour planning and their preferences. Since this problem is an NP-hard problem, heuristic algorithms such as the genetic algorithm (GA) are well suited for solving it. However, this algorithm can take a very long time to find the optimal solution, so this article focuses on developing a greedy strategy of GA to find optimal or near-optimal solutions. Instead of random genetic transformations, the algorithm consciously modifies optimal routes in order to find the desired tour for the users in a shorter time. After this, the developed algorithm was compared with a GA using various generated user profiles. Over 700 locations in Novosibirsk were used as a dataset for making recommendations. These modifications made it possible to obtain optimal routes faster than the standard implementation of the genetic algorithm.