Intelligent Personalised Tourist Route Generation
The main objective of Personalised Electronic Tourist Guides (PETs)is to help tourists in personalised route generation tasks, offering added-value functionalities. PETs provide an integrated solution for route generation based on the profile and constraints of tourists, and up-to-date Points Of Interest (POIs) and destination information. The main objective of this thesis is the implementation of advanced algorithms to improve the personalised route generation functionality of PETs, including public transportation. This work proposes different approaches to solve new simplified versions of the Tourist Trip Design Problem (TTDP)in real-time. On the one hand, an efficient heuristic for the Multi Constrained Team Orienteering Problem with Time Windows (MCTOPTW), which adds multiple constraints to previous existing models, is proposed. Moreover, this heuristic is tested with related approaches from the literature and with a new test set we propose for the MCTOPTW. On the other hand, two different solution approaches to tackle the time-dependent travel times for the Time Dependent Team Orienteering Problem with Time Windows (TDTOPTW), an extension of previous models including public transportation, are presented. All of the approaches are tested with a test set based on the city of San Sebastian. Finally, a prototype developed to test the viability of the TDTOPTW solution algorithms in a real world tourist environment is described. Moreover, the results of the validation performed with the Local Tourist Office (LTO) staff and tourists visiting the city are shown.