Files
Abstract
Demand of using public transit has been tremendously increasing during the past decade. As the public transit system become more and more complicated both in volume and dimension, the need for trip planning assistance, especially trips involving transfers, is emerging. Traditional shortest path algorithms are not very suitable for trip planning tasks involving transfers. A new network representation is introduced inspired by the traditional connectivity network, only with multiple layers. Each layer represents one route in the network. An algorithm is designed and implemented to accommodate the network representation with relatively high computational efficiency. A series of matrices are introduced to store the information of least transfer paths between each pair of nodes in the network. A NYC subway case study is conducted to validate the algorithm and examine the efficiency of the algorithm. Potential application of the least transfer path matrices in planning supporting is discussed.