This paper presents how the capacity of mass transit can be optimally operated to alleviate congestion pressure during the evacuation of busy urbanareas. The proposed model extends the traditional vehicle routing problem(VRP) to include: Multiple Depots to better distribute the transit fleet,Time Constraints to account for the evacuation time window, and constraints for Pick-up and Delivery locations of evacuees. The evacuation problem is hereafter defined as a Multi-Depot Time Constrained Pick-up Delivery Vehicle Route Problem (MDTCPD-VRP). A framework, using Constraint Programming (CP), is developed to model and solve the MDTCPD-VRP evacuation problem.An Optimal Spatoi-Temporal Evacuation (OSTE) model is performed first to optimize the evacuation of the background vehicular traffic, generating transit travel cost (i.e. link travel times) as an input to the MDTCPD-VRP. We apply our methodology on a case study of a hypothetical evacuation event in the busiest core of the downtown area of the City of Toronto, Ontario, with 65% of total evacuees are transit-dependent. The results show the optimal scheduling and routing plan for transit-vehicles as the solution tothe evacuation problem defined as MDTCPD-VRP. The equilibrium mode-split between traffic and transit (at which the total vehicle-travel time is equal for vehicular drivers and transit users) is found to be at 25% and 75%,respectively.
Abstract