L. Amorosi, P. Dell'Olmo, G. L. Giacco
This paper describes a new method for automatically generating a text to customers in a concise and clear way with a calendar represented by a boolean vector as its input. Our attention is focused on problems arising from calendars in the transportation field, in particular railway services. The aim is to verify if it is possible to optimize the quality of outcomes in terms of intelligibility with a mathematical programming approach. Two mathematical formulations for this challenging task have been proposed. The first one consists in using a specific set covering model with embedded generation of sets out of the ground set (integrated model). The second one combines a set covering model with a parallel vector generation algorithm. A Mixed Integer Programming solver is adopted to solve the two models, while a parallel algorithm for the vector generation problem is designed in the second approach. Both models have been tested on several real railway timetables, always obtaining optimal solutions which compare favorably with those produced using current practices. Moreover, an enhanced formulation for the integrated model with better performances is presented. Computational times show that both approaches are applicable in practical contexts.
Mathematical Models, Optimization, Transportation Services
Tipo di pubblicazione: