5. Design of optimization algorithms

Preliminary work for the test and evaluation of several algorithms was conducted. Based on that, a set of algorithms to estimate OD matrices from entry-only Automatic Fare Collection data were developed. Accordingly, different Trip Chaining Models were implemented by considering different sets of assumptions, along with the identification of transfers made by passengers. Moreover, it included the development of algorithms to identify Public Transportation networks that best benefit from synchronization timetabling optimization.

Additionally, the potential of a Demand Responsive Transport system for people with reduced mobility in an urban environment was explored. For this purpose, the Dial-A-Ride model was implemented using a multivariable minimisation approach. In this approach, an algorithm for “assigning requests to vehicles” was used to obtain an initial solution. Then a Multi-Objective Tabu Search algorithm was applied to minimize the total travelled distance, the deadheading distance and the number of vehicles. The performance of the model was computed combining using six different parameters.

At the end of this task, a set of optimization models will be obtained to improve the quality of intermodality services provided by Urban Public Transport. They will address the optimized synchronization of schedules and routes from distinct transport modes, whilst encompassing the requirements from passengers (service quality, efficiency and reliability) and from operators (operation costs). Due to their combinatorial nature, metaheuristics combined machine-learning approaches will be used.