Banner
Home      Log In      Contacts      FAQs      INSTICC Portal
 

Keynote Lectures

The e-VSP Problem: Planning Electric Buses and Drivers
Han Hoogeveen, Universiteit Utrecht, Netherlands

Dynamic Pricing under Competition: Challenges and Opportunities
Rainer Schlosser, Hasso Plattner Institute, Germany

Two Approaches to Just-in-Time Scheduling
Joanna Józefowska, Institute of Computing Science, Poznan University of Technology, Poland

 

The e-VSP Problem: Planning Electric Buses and Drivers

Han Hoogeveen
Universiteit Utrecht
Netherlands
 

Brief Bio
I work as a senior researcher and lecturer at the department of Computer Science at Utrecht University, where I am a leading researcher in the AI & Mobility Lab. Moreover, I hold a joined position with Dutch Railways (NS) within the KickstartAI program. I started my academic career by working on theoretical scheduling problems, but after some years I discovered that it is both challenging and rewarding to solve non-trivial problems that occur in real-life. Together with many master students and PhD-students, I have done research projects from very diverse areas of logistics, ranging from planning public transport to blood transfusion, and from making rosters for nurses to making plans for cutting flowers. In this way, I started a project with NS that has led to the development of the first method that is capable of automated planning of service sites (shunting yards with service facilities) for realistic instances. Our method will be taken into operation by NS throughout the Netherlands in the near future. Another successful project is on planning electric buses (and drivers) with Qbuzz, which provides the public bus transport in large parts of the Netherlands. In my presentation, I will present results of the latter project.


Abstract
For the operation of Public (Bus) Transport it is key to get buses and drivers at the right position at the right time. This leads to a huge planning puzzle. To find the best plan, we use techniques from Artificial Intelligence and Operations Research. In this presentation I will discuss a real-life project concerning the planning of buses and drivers at Qbuzz. To make the Public Transport more sustainable, Qbuzz uses more and more electric buses. Since the batteries are not big enough to let the bus operate all day, this introduces the need for recharging the batteries, which must be included in the driving plans of the buses. Furthermore, each bus needs a driver, and due to the labor rules planning the drivers is a difficult problem, even if the bus schedules are known. I will first discuss a new approach, combining Local Search and ILP, to plan the electric buses, and next I will extend it to solve the integrated bus and driver planning problem.



 

 

Dynamic Pricing under Competition: Challenges and Opportunities

Rainer Schlosser
Hasso Plattner Institute
Germany
 

Brief Bio
I am a senior researcher at the Hasso Plattner Institute, University of Potsdam, Germany, leading the research group “Data-driven decision support”. I graduated in Mathematics as well as Business Administration and received a PhD in Operations Research from Humboldt University of Berlin, Germany. My research focuses on automated decision-making in the areas of Revenue Management and beyond using quantitative methods of operations research and data science. Our research has been published in over 60 peer-reviewed publications including renowned OR Journals (EJOR, JEDC, JCLP, IJPE, IJPR, COR, DGAA, JRPM), distinguished data science conferences (KDD, IJCAI, RECSYS), and leading computer science venues (VLDB, ICDE, EDBT, DAPD). Further, I serve as a reviewer for over 50 Journals in the area of operations management and data science.


Abstract
Online markets have become highly dynamic and competitive. Many sellers use automated data-driven strategies to estimate demand and to update prices frequently. Further, notification services offered by marketplaces allow to continuously track markets and to react to competitor's price adjustments instantaneously. To derive successful automated repricing strategies is challenging as competitors' strategies are typically not known. In this talk, we look at automated repricing strategies in theory and in practice. We discuss data-driven price anticipations, self-learning strategies, and their interaction. Further, we consider dynamic pricing agents when applied to raising recommerce markets, which offer the opportunity to trade in and resell used products.



 

 

Two Approaches to Just-in-Time Scheduling

Joanna Józefowska
Institute of Computing Science, Poznan University of Technology
Poland
www.cs.put.poznan.pl/jjozefowska
 

Brief Bio
Joanna Józefowska is a full Professor at the Institute of Computing Science, Poznan University of Technology (PUT). Graduated in Mathematics at Adam Mickiewicz University in Poznan and Management Engineering at PUT, Ph.D. and habilitation at the Faculty of Electrical Engineering, PUT. Post-doc at Memorial University, Newfoundland Canada, visiting professor at Brandenburg University of Technology and University of Siegen, Germany. Currently Vice-president of EURO (The Association of European Operational Research Societies) and one of two Coordinators of the EURO WG on Project Management and Scheduling. Major research areas: scheduling (project, machine, production), artificial intelligence – in particular knowledge representation systems. Author and co-author of 8 books and more that 100 papers in major professional journals and conference proceedings.


Abstract
Scheduling, defined as allocation of resources to tasks over time, belongs to the basic optimization problems in production engineering. The theory of scheduling evolves with the development of manufacturing and control systems. Among others, the concept of just-in-time scheduling led to formulation of new models and algorithms addressing the resulting production requirements. In this talk two approaches to just-in-time scheduling will be discussed. The first one is based on the product rate variation concept applied in Toyota and the second one consists in minimization of a total earliness and tardiness cost. It may be shown that each of them relates to slightly different production conditions. Both lead to interesting theoretical results.



footer