Optimization of Bus Stops Locations Using GIS Techniques and Artificial Intelligence

Optimization of public transportation network in terms of reducing travel time and providing access to areas currently without sufficient access to the service facility would certainly motivate private car owners to use public transport. The reduction of vehicle numbers on the roads will undoubtedly lead to minimizing traffic congestion and reducing air pollution due to lesser exhaust emissions. This study used Geographic Information System (GIS), Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) to model the location of bus stops in Amman city in Jordan to find optimal travel time and serviceability of stops. The GIS modeling provided significant reduction of travel time for streets with redundant bus stops located at irregular distances, e,g. Zahran Street, where travel time was reduced by (23.25%) during off-peak hours and by 19.74% during peak hours, while PSO provided (28.95%) and 39% reductions, respectively. GA also provided significant reduction in travel time (47.96 %) for Zahran Street. For streets with insufficient number of bus stops, travel time went up due to increase in the number of stops, e.g. Al-Quds Street, where travel time increased from 12.25 min per route to 50.71 min due to increase in the number of stops from 6 to 25. In contrast, this increase in travel time significantly reduced the walking distance to the bus stop, from over 2000 m to approximately 400 m. Moreover, demand for bus stops serviceability significantly more than capacity. Average demand per bus stop on Al-Quds street, for example, was 10456 persons with a capacity of 1465.Comparison of the models by applying them to roads not included in the study showed consistent results that further confirmed the reliability of the models. The developed PSO algorithm and GA are easy to use for urban planning, where they can be applied to any existing network or planned development. (C) 2020 The Authors. Published by Elsevier B.V.