IoT-Enabled Dynamic Planning in Smart Logistics
MARTIJN KOOT
PHD CANDIDATE OF INDUSTRIAL ENGINEERING, UNIVERSITY OF TWENTE.
Global supply chains have become more responsive towards disruptions due to recent IT advancements. For example, modern-day fleets are well equipped with wireless sensing, processing, and communication devices, which enables logistics operators to track their fleet’s performances in (near) real-time. Fleet coordinators can better anticipate on dynamic events by instantly reviewing centralized track-and-trace information already, but the rescheduling decision itself remains mainly dependent on human experience only. In other words, how should logistics operators incorporate real-time sensor data into their dynamic planning activities? In this presentation, we will briefly discuss how to improve logistics decision making by exploiting the real-time insights originating from the Internet of Things (IoT) in combination with data-driven optimization techniques. The main aim is to develop a resilient planning framework that enables logistics planners to better anticipate on disturbances by combining both human experiences, operations research techniques, and pattern recognition mechanisms (e.g., machine learning, data mining, etc.).
Martijn Koot is a PhD Candidate within the department Industrial Engineering and Business Information Systems at the University of Twente, Enschede. He holds a BSc and MSc degree in Industrial Engineering & Management from the University of Twente. His current research is devoted to improve logistics decision making by exploiting the enormous data sets originating from IoT networks in combination with Big Data Analytics. Besides his research activities, Martijn also supports various BSc courses related to research methodology.
Meeting Corporate Renewable Power Targets
DR. ALESSIO TRIVELLA
ASSISTANT PROFESSOR OF INDUSTRIAL ENGINEERING, UNIVERSITY OF TWENTE.
After briefly overviewing my research, I will focus on a work that studies the global trend of corporations publicly committing to renewable power purchase targets, which entails buying a percentage of power demand from renewable sources by a future date. Long-term financial contracts with renewable generators based on a fixed strike price, known as virtual power purchase agreements (PPAs), are popular to meet such a target. We formulate power purchases using a portfolio of PPAs as a Markov decision process accounting for price and supply uncertainties. Since computing an optimal procurement policy is intractable, we consider forecast-based reoptimization heuristics that vary the sourcing of different PPA types and the timing of new agreements, and propose a novel information-relaxation based reoptimization heuristic. Our computational study on realistic instances supports the effectiveness of PPAs for meeting a target, highlights the benefit of sourcing and timing flexibility in procurement decisions, and shows that dynamic PPA portfolios from our information-relaxation based procurement heuristic are near optimal, thus helping tie the knot between reducing energy costs and meeting renewable targets.
Dr. Alessio Trivella is an Assistant Professor of Operations Research at IEBIS since January 2022. Previously, he obtained a MSc degree in mathematics from the University of Milan, a PhD in Operations Research from the Technical University of Denmark, and has been a postdoc at ETH Zurich. His research focuses on developing optimization techniques for sustainable corporate operations within the energy, transport, and logistics sectors, using tools from Operations Research, Reinforcement Learning, and Large-scale optimization. His research has been published in journals including Management Science and M&SOM.