Bi-objective optimization of transit network and frequencies design using Pareto genetic algorithm

This study solves the simultaneous planning problem of the transit network and frequency design (TNFD) for suburban bus systems. A Pareto genetic algorithm (PGA) in which the fitness function is designed based on the Pareto relation and the population diversity is first applied to better tackle the bi-objective nature of the problem. The real size network of the suburban residential area is conducted to show that our PGA is effective in generating better Pareto optimal solutions for both conflicting objectives when compared with the existing bus network design, since solutions with very high percentage of direct trips, lower in-vehicle travel time and fewer buses could be found. Other two experiments are set up to demonstrate the good robustness of our proposed algorithm.