Solving the Transit Network Design problem with Constraint Programming

Alexandre Barra (1) , Luis Carvalho (2) , Nicolas Teypaz (3) , Van-Dat Cung (3) , Ronaldo Balassiano (4)

1 Alumnus, PET/COPPE
2 Division of Applied Mathematics
3 G-SCOP_ROSP - Recherche Opérationnelle pour les Systèmes de Production
4 PET/COPPE-UFRJ - Programa de Engenharia de Transportes

Transit network design (TND) is a complex combinatorial optimization problem. The goal is to obtain an optimized public transport network, taking into consideration main characteristics of the system, passenger's desires, budget constraints and several possible levels of service. The author proposes a comprehensive nouvel model, based on Constraint Satisfaction, with essential and complementary constraints. Using Constraint Programming (CP), some tests are made using the proposed model in an attempt to open new perspectives for solution alternatives. The tests show that the chosen CP package is apparently unable to process large instances, although it works well with small ones. Other CP packages and search strategies should be tried by CP specialists in order to comprehensively validate the conclusions of the present work.

Domaines: Recherche opérationnelle [math.OC]

Alexandre Barra, Luis Carvalho, Nicolas Teypaz, Van-Dat Cung, Ronaldo Balassiano. Solving the Transit Network Design problem with Constraint Programming. 11th World Conference in Transport Research - WCTR 2007, Jun 2007, Berkeley, United States. ⟨hal-00189690⟩