Date of Award

6-30-2022

Publication Type

Thesis

Degree Name

M.Sc.

Department

Computer Science

Keywords

Cultural algorithm;Itinerary recommendation;Optimization

Supervisor

Pooya Moradian Zadeh

Abstract

The tourism industry plays a vital role in today’s world. Many people travel around the world to visit and explore other places. However, planning an itinerary is one of the most challenging and time-consuming tasks for many travelers. It could be even more complicated when they travel as a group with different constraints and various choices of points of interest (POIs). The problem of group itinerary recommendation is an extension of the orienteering problem and is NP-hard, which can be defined as an optimization problem. This research will address the problem by proposing a personalized group itinerary recommendation algorithm using cultural algorithms. Cultural algorithms are evolutionary algorithms that use knowledge to guide the search direction during the evolution process. The main objective of our proposed model is to maximize the group’s satisfaction by optimizing the number of visiting POIs, while considering the interests of all users, travel time, visit duration, and budget. The performance of our proposed model is evaluated on real-world datasets and compared with the existing methods. The results revealed that our proposed algorithm outperforms alternative baselines on both datasets in most of the experiments, which means our final solution had better quality compared with other algorithms. Furthermore, non-parametric tests demonstrated that this approach generates consistent results in various situations and is notably different from existing algorithms.

Download

Share

COinS