In this talk, I will present our recent work on pickup-and-delivery problems from two perspectives. First, focusing on a city application, I will present the Senior Transportation Problem. A number of organizations in Toronto and the surrounding area provide on-demand transportation services for members of the elderly population to attend doctor's appointments, engage in social events, and do everyday activities such as grocery shopping. More formally, a fixed fleet of volunteer-operated, heterogeneous vehicles from multiple depots must satisfy as many door-to-door transportation requests as possible within a fixed time horizon. All requests consist of a pickup and delivery with time windows, maximum ride-time limits, and limited vehicle capacities. I will present a number of optimization models for this problem. Second, inspired by the single vehicle version of the senior transportation problem, I will present work on the use of decision diagrams as an optimization substrate for the multi-commodity pickup-and-delivery traveling salesperson problem. Decision diagrams are a recent addition to the stable of tools for optimization and I will show how they can be used in combination with Lagrangian relaxation to achieve state-of-the-art performance on these hard logistics problems.

Joint work with Chang Liu, Margarita Castro, Dionne Aleman, and Andre Cire.