Touristic route planning problem: The case of Eskişehir

According to the Turkish Statistical Institute, tourism holds significant importance for the Turkish economy. Turkey's tourism revenues increased by 16.9% in 2023 compared to the previous year, reaching 54.3 billion USD. Eskişehir has become a prominent city in the tourism sector due to its cultural and archaeological values, historical sites, museums, events, festivals, and natural attractions. Eskişehir has become a city that hosts millions of tourists every year in both domestic and foreign tourism. The Traveling Salesman Problem with Hotel Selection (TSPHS) was first studied in the literature by Vansteenwegen (2012) and is an extension of the Traveling Salesman Problem. This problem aims to minimize the total travel time and the number of days of the trip. There are several restrictions within the scope of the model, such as each trip starting and ending at the same hotel and each point of interest (POI) being visited only once. In this study, a tourist route plan including 15 POI and 5 hotels was designed for the cities of Eskişehir and Samsun, Turkey based on the CSVG (Castro, Sörensen, Vansteenwegen, Goos) Model. However, in this study, the restriction that the trip should start and end at a certain hotel was removed and time window constraints were added to the model. Thus, a more flexible shortest traveler route was designed that allows the start and end hotels to be determined by the model. In addition, the Single Objective Model with Time Window was created by considering the opening and closing times of the POI. Then, a second objective was added and the Multi Objective Model with Time Windows was created. The Single Objective Model with Time Windows aims to minimize the total travel time. The Multi Objective Model with Time Windows aims to minimize the total travel distance and maximize the total satisfaction score of the tourists at the same time. The Pareto efficient frontier was drawn for each problem using the ε -constraint method and a set of Pareto optimal solutions were obtained.