ULTRAOPTYMAL - Urban Logistics and sustainable TRAnsportation: OPtimization under uncertainTY and MAchine Learning

The global aim of ULTRAOPTYMAL is to model and solve realistic and timely urban goods transportation problems, while explicitly considering their uncertain dimensions, such as Customer Demand (CD), Travel Times (TT), Presence of Customers (PC). Accounting for costs, carbon emissions, and customer service levels, our models will facilitate the identification and design of sustainable local authority policies aimed at supporting livability, connectivity, and sustainability. We pursue 3 main research objectives:

• O1. Optimizing urban distribution planning: path selection for VRPs taking into account uncertain CDs and correlated TTs. A Stochastic Optimization framework including CD forecasting and a new traffic flow predictive model, exploiting the dynamic dimension of the available information will be developed. 

• O2. Integrating sustainable transport modes and delivery options: integration of sustainable transport modes and different delivery options into problems for efficient last-mile delivery. We will adopt Robust Optimization approaches to study 2E-LRPs to better plan urban deliveries by taking into account the uncertainty in PC and TT.

• O3. Solving reverse logistics problems: routing and location problems in the area of reverse logistics by considering uncertain CD and PC. We will use Stochastic Optimization methodologies, by evaluating the inclusion of different risk measures, and we will develop efficient algorithms capable of optimizing the resulting models.

The objectives will be achieved by developing adequate Mathematical Programming models along with efficient algorithms integrating Machine Learning and optimization for several classes of difficult combinatorial problems considering logistics operations under cost and sustainability goals.

The website is always updated with new materials... Stay tuned!