Personal Rapid Transit network design using Genetic Algorithm and Ant Colony System hybridization

This study proposes a new hybrid meta-heuristic to address the guideway network (GN) design problem of personal rapid transit (PRT). PRT is a novel transportation paradigm, which operates a number of driverless vehicles over an elevated GN. Since the GN interconnects many stations, designing an efficient GN is a challenging problem even for a moderate-sized PRT system. To solve the GN design problem effectively, we propose a hybrid meta-heuristic of a genetic algorithm (GA) and a local search ant colony system (ACS). The proposed hybrid meta-heuristic uses a special representation technique named cycle representation, which encodes a GN candidate as an aggregation of cycles (circulators). The GA searches for the best combination of the cycles via special genetic operators dedicated to the cycle representation. On the other hand, the local search ACS accelerates the search by fine-tuning the cycles using the pheromone matrix maintaining past search history. The empirical tests performed for realistic GN design problems verify the effectiveness and efficiency of the proposed hybrid meta-heuristic.