Decision-making system for a route recommendation app incorporating user preferences

Abstract

This thesis explores whether incorporating weather data and user preferences into route planning can improve recommendations for semi-autonomous vehicles in Finland. Traditional routing systems typically rely on shortest-path algorithms that do not account for dynamic environmental conditions or individual comfort preferences, factors particularly important in regions with challenging weather. After reviewing various decision-making approaches capable of integrating multiple criteria, the Analytic Hierarchy Process (AHP), a Multi-Criteria Decision Analysis (MCDA) technique, was selected and implemented to personalize route recommendations. A local Open-Source Routing Machine (OSRM) instance generated multiple alternative routes, and temperature data from the Open-Meteo API was associated with route midpoints to represent weather conditions. Routes were evaluated based on travel distance, duration, and temperature, reflecting user priorities. The results showed that incorporating temperature preferences influenced route selection in approximately one-third of cases, indicating MCDA’s capacity to adapt recommendations beyond traditional shortest-path solutions. Logistic regression analysis identified temperature as the strongest predictor of route changes, alongside travel time and distance.

While real-time traffic data was initially considered, it was excluded due to integration challenges, limiting the study to free-flow traffic conditions. The thesis discusses implementation challenges, methodological choices, and outlines future opportunities to incorporate more granular weather data, dynamic traffic information, and personalized user feedback. Overall, this work contributes evidence that MCDA-based decision-making can effectively incorporate environmental preferences into semi-autonomous vehicle route planning, paving the way for safer and more comfortable personalized navigation experiences.