Demand Responsive Transport and Transit Synchronization using Adaptive Large Neighborhood Search

Urban transit operations exhibit limited accessibility and connectivity to the critical transit stops from the suburbs and low-density areas. The shared vehicles that act as demand responsive transport (DRT) can improve system connectivity, while also reducing congestion. Shared vehicles typically address this problem by directing the unserved demand to a major transit stop, such as a bus or metro station. However, the first mile with shared vehicles from low density areas with specific time-windows is still a challenge from a combinatorial optimization perspective. The goal of this research is the synchronization of shared demand responsive transport with the transit system by using the adaptive large neighborhood search (ALNS) algorithm to iteratively explore wide regions surrounding potential solutions. We use ALNS to search for the most probable points as origin in a large region (low-density area), and we define the pick-up points as the origin. We further develop an optimizer that synchronizes the pick-up time at the origin point and the drop-off time at the transit stop by utilizing the optimum fleet size of shared vehicles to cater to the highest potential demand in the given area. The proposed approach is evaluated on the city of Paiania at the suburbs of Athens (Greece) which is strategically connected to the metro network, providing essential linkage to the city center.Results show that employing the proposed optimizer for synchronizing DRT with transit results in a significant reduction in the system's overall travel time and cost. Furthermore, the waiting time at the transit stop also decreases due to the proposed time windows-based approach. This methodology improves the efficiency and attractiveness of DRT for both users, policy maker and operators through optimized synchronization of two complementary mobility services.