Traveler-oriented multi-criteria decision support for multimodal itineraries

In recent years, the variety of mobility services has increased strongly. Travelers expect a diverse set of combined mobility services according to multiple individual preferences. Due to the competing characteristics of these preferences (e.g. travel time, price, and the number of transfers), several Pareto-optimal itineraries representing trade-offs arise. While there are efficient approaches for finding multimodal shortest paths, the full set of Pareto-optimal itineraries cannot be determined efficiently when multiple traveler preferences are considered in a large multimodal network. However, this would be required to provide travelers with more relevant choices in the light of available options and complex solution space characteristics. In this work, we propose smart ways to approximate the Pareto front of multimodal itineraries efficiently. The core idea is to apply solution space sampling systematically. We focus on the scalability of the sampling framework with respect to multiple traveler preferences as well as identifying interesting characteristics of itineraries to enable travelers to take wellinformed decisions. The framework is evaluated with a large amount of real-world data of mobility services. To this end, we analyze long-distance trips between major cities in Germany, taking up to five most prevalent traveler preferences into account. In addition, we examine the Pareto-optimal solutions and derive characteristics of potential interest for the traveler that can help to make the search more transparent and explainable and thus shape the traveler's choice set.