Plenary Presentations
Prof Leo Kroon
Biography
Leo Kroon is a professor of quantitative logistics at the Department of Decision and Information Sciences, Rotterdam School of Management, Erasmus University (RSM). Leo Kroon is also a logistics consultant in the department of Process quality and Innovation Netherlands Railways (NS), the main operator of passenger trains in the Netherlands. His main research interest is the development of decision support tools for the planning and real-time operations control of logistic systems, in particular public transport systems. His research involves close cooperation with practice and has resulted in numerous published articles in journals including Transportation Science, Transportation Research B, and Interfaces.
Professor Kroon was a member of the NS team that won the prestigious INFORMS Edelman Award 2008 for its model-based contributions to the development of the 2007 NS timetable. He is the recipient of 2008 ERIM Impact Award. He was a coordinator of a Dutch team in the EU funded research projects AMORE and ARRIVAL. Currently, he is one of the project leaders of EUR team in the EU funded research project ON TIME. This project focuses on robust planning and effective recovery of railway systems. Currently he is also the project leader of the NWO funded Complexity in Public Transport project. He is an Associate Editor of the journal Transportation Science. Professor Kroon received his Master's in mathematics cum laude from the Free University in Amsterdam and his PhD from the Erasmus University.Talk Title - Passenger oriented railway disruption management by adapting timetables and rolling stock schedules
In passenger railway operations, unforeseen events require railway operators to adjust their timetable and their resource schedules. The passengers will also adapt their routes to their destinations. When determining the new timetable and rolling stock schedule, the railway operator has to take passenger behavior into account. The capacity of trains for which the operator expects more demand than on a regular day should increase. Furthermore, at locations with additional demand, the frequencies of trains serving that station could be increased. This talk describes a real-time disruption management approach which integrates the rescheduling of the rolling stock and the timetable by taking the changed passenger demand into account. The timetable decisions are limited to additional stops of trains at stations at which they normally would not call. Several variants of the approach are suggested, with the difference in how to determine which additional stops should be executed.
Prof Ben Paechter
Biography
Prof Ben Paechter wrote his first paper on automated timetabling in 1994. After meeting at the conference where that paper was presented, a group of UK researchers in the same field got together to organise the first international conference on the practice and theory of automated timetabling which took place in Edinburgh in 1995 with Ben as the Chair of the local organising committee. That conference was the start of the PATAT series which is now in its tenth iteration. Ben served as Treasurer and member of the PATAT Steering Committee until 2010, and has been part of the Programme Committee throughout the series. With a specialism in Evolutionary Computation Ben has been involved in a number of research projects funded by national and European bodies including coordinating an early project distributing the processing of algorithms across a peer-to-peer network. Another project compared the performance of various metaheuristic algorithms across a range of combinatorial optimisation problems including timetabling. As part of this project Ben led the organisation of the first International Timetabling Competition in 2002. This competition led to 2nd and 3rd competitions in 2007 and 2011 and Ben led a stream in the second competition. More recently Ben has led and been involved in projects that coordinate research across Europe, organising research agendas, summer-schools, interactions with industry, exchanges of research ideas, and the development of new funding programmes. Ben likes to concentrate on producing algorithms which solve real-world problems, and his “Tatties” automated tabling system was used in his institution, Edinburgh Napier University for seven semesters, timetabling all of the classes of the university using twelve competing objectives. Ben is now an Assistant Dean at the institution, responsible for Research and Innovation. He continues to publish in the field of evolutionary computing, combinatorial optimisation and related areas and is an Associate Editor of the journal Evolutionary Computation (MIT Press).Talk Title - “Mine’s better than yours” – comparing timetables and timetabling algorithms
Automated timetabling papers, along with papers about all types of optimisation methods are full of claims that a particular algorithm, or adaption to an algorithm “out-performs the state-of-the-art”. If science is going to move the world forward then we need to be sure what we mean when we make claims such as this – and have ways of verifying that the claim is indeed the case. Taking a tutorial approach, this talk begins by looking at how we might decide how good a particular timetable solution might be – or how we might, at least, compare one solution with another. Ways of considering the hard and soft constraints are examined, along with methods of dealing with multiple objectives. Three classifications of soft constraints are defined. The need to work with the person using the automated system to fully understand what is “good” about a timetable is underlined. This might include non-obvious criteria, such as the need to reduce the chance that users of the timetable will be able to suggest variations which lead to improvement.
Once we understand how we can compare two solutions to a particular timetabling problem, we can then look at how we might compare two algorithms trying to find good solutions to problems. Factors that might be taken into account are, for example, speed, reliability, closeness to optimum, or the chances of a really super result once in a while. Non-obvious criteria might be for example the extent of the ability for the user to change their mind about what they care about (in terms of either timetable or algorithm quality) in the middle of producing the timetable. Again, in order to be really useful to the world, the emphasis here needs to be on finding out what the users of the algorithm really want from it. The use of standard problem instances is examined along with analysis of the conditions that make this is useful or not. The talk argues that, in order to be useful, problem instances need to be accompanied by one or more standard sets of criteria on which solutions will be measured and, just as importantly, one or more sets of criteria on which algorithms will be measured. Even when there are clear criteria for judging between algorithms on a specified set of problem instances there can still be problems interpreting the results meaningfully. For example we have to be careful that an algorithm is not just good at solving the particular problem instances, i.e. over-fitted to those instances. One way to try to solve this problem is to have public competitions where algorithms are developed on one set of problem instances, and then compared on another set. Competitions also help to encourage work in a particular area, and work which is genuinely comparative. The talk will discuss the author’s experience of running competitions, and what makes a successful competition. The problem of competition entrants using different hardware, compilers, interpreters and operating systems is also discussed. Competitions have largely concentrated on hidden problem instances having the same user criteria. The author will argue that the future may lie in competitions where hidden problems change the user requirements in some way, so as to encourage the development of algorithms which are more generally useful.Prof Kate Smith-Miles
Biography
Kate Smith-Miles is a Professor in the School of Mathematical Sciences at Monash University in Australia, where she was Head of School from 2009-2014. She has recently commenced as the inaugural Director of MAXIMA (the Monash Academy for Cross & Interdisciplinary Mathematical Applications). She has previously held Professorial positions (chairs) in Engineering at Deakin University (where she was Head of the School of Engineering and Information Technology from 2006-2008) and in Information Technology at Monash University, where she worked from 1996-2006. Her third Chair (in Mathematical Sciences), demonstrates her multi-disciplinary breadth. Kate obtained a B.Sc(Hons) in Mathematics and a Ph.D. in Electrical Engineering, both from the University of Melbourne, Australia. She has published 2 books on neural networks and data mining applications, and over 220 refereed journal and international conference papers in the areas of neural networks, combinatorial optimization, intelligent systems and data mining. She has supervised to completion 22 PhD students, and has been awarded over AUD$10 million in competitive grants, including 11 Australian Research Council grants and industry awards. From 2007-2008 she was Chair of the IEEE Technical Committee on Data Mining (IEEE Computational Intelligence Society). She was elected Fellow of the Institute of Engineers Australia (FIEAust) in 2006, and Fellow of the Australian Mathematical Society (FAustMS) in 2008. She was awarded the Australian Mathematical Society Medal in 2010 for distinguished research. In 2012, she received the Vice-Chancellor’s Award for Excellence in Postgraduate Supervision. In addition to her academic activities, she also regularly acts as a consultant to industry in the areas of optimisation, data mining, and intelligent systems.She currently serves as an Area Editor for Computers & Operations Research.
Talk Title - Visualising the diversity of benchmark instances and generating new test instances to elicit insights into algorithm performance
Objective assessment of optimization algorithm performance is notoriously difficult, with conclusions often inadvertently biased towards the chosen test instances. Rather than reporting average performance of algorithms across a set of chosen instances, we discuss a new methodology to enable the strengths and weaknesses of different optimization algorithms to be compared across a broader instance space. Results will be presented on timetabling, graph colouring and the TSP to demonstrate: (i) how pockets of the instance space can be found where algorithm performance varies significantly from the average performance of an algorithm; (ii) how the properties of the instances can be used to predict algorithm performance on previously unseen instances with high accuracy; (iii) how the relative strengths and weaknesses of each algorithm can be visualized and measured objectively; and (iv) how new test instances can be generated to fill the instance space and provide desired insights into algorithmic power.Prof Konstantinos G. Zografos
Biography
Konstantinos G. Zografos is Chair Professor at the Department of Management Science, Lancaster University. His professional expertise, research and teaching interests include applications of Operations Research and Information Systems in Transportation and Logistics. His current work is focused on vehicle routing and scheduling, itinerary planning, facility location, airport planning and operations, emergency response logistics, supply chain management, and project management. He has published more than 60 papers in refereed academic journals and edited volumes. He has been a member of the Editorial Board of Transportation Research Part C: Emerging Technologies, Operational Research: An International Journal, Simulation Modeling Practice and Theory, International Journal of Logistics Economics and Globalization, and Journal of Aerospace Operations, and has served as co-editor of the special issue of Transportation Science on Hazardous Material Transportation. He has been involved as a principal investigator in more than 60 R&D projects funded by national and international organizations and companies in USA, Europe, and Greece. He has acted as consultant to projects funded by governmental agencies, companies, and international organizations, including the European Commission, United Nations Economic Commission for Europe (UNECE), and EUROCONTROL.
Professor Zografos has received: i) the ENO Foundation for Transportation award in 1986 ii) the "Excellence in Teaching Award for 2003-2004" of the MBA International Program of the Athens University of Economics and Business, iii) the 2005 President's Medal Award of the British Operational Research Society and iv) the Edelman Laureate Honorary Medal of the Institute of Operations Research and the Management Sciences (INFORMS) in 2008 for significant contributions to Operations Research.Talk Title - Pushing the Envelope: the role of slot scheduling in optimising the use of scarce airport resources
The rapid growth of demand for air transport services coupled with political, physical and institutional constraints for building new airport capacity has resulted in acute airport congestion in UK and across Europe. Imbalances between traffic and capacity generate serious undesirable externalities for air transport and the society at large. Increasing complications for expanding capacity render a pure supply-side solution both expensive and practically difficult to implement. In effect, a more sustainable approach being able to better cope with the congestion problem with existing resources is called for. Solutions aiming to manage congestion through the optimum allocation of scarce airport capacity have received a great deal of consideration from the airport community, policy makers, and researchers. Capacity at congested airports is expressed in slots. A slot identifies a time interval, specific date and time, during which a carrier is permitted to use the airport infrastructure for landing or take-off at a slot-controlled (coordinated) airport. Therefore, slot scheduling and the setting of optimum capacity levels are closely interdependent and both lie at the heart of optimising the allocation and use of scarce airport resources. Slot scheduling signifies a challenging stream of research due to its potential to generate quick and drastic capacity utilisation improvements and the complexity and size of the resulting mathematical problems. The objective of this presentation is to provide an overview of the evolution of slot scheduling modelling, identify open research issues, and underline the potential of slot scheduling in optimising the allocation and use of scarce airport resources. We first identify slot scheduling requirements arising from the decision making environment of congested airports and we characterise the resulting slot scheduling problems. We discuss the implications of misuse and mismanagement of airport capacity due to poor slot scheduling decisions, and the lack of rationalising the definition and setting of airport declared capacity. We also review, classify, and comparatively discuss existing slot scheduling models and we identify gaps between slot scheduling requirements and the capabilities of existing slot scheduling models. We conclude the presentation with directions for future research for developing the next generation of slot scheduling models for closing the gap between decision making needs and available decision support capabilities.Paul Harrington and Geoffrey Forster
Scientia Ltd
Biographies
Paul Harrington is the Head of Product Development at Scientia. He has over 15 years of experience of developing scheduling software.
Geoffrey Forster is a Director of Scientia Resource Management and was the founder of Scientia in 1989. He led the development of Syllabus Plus. Scientia is the global market leader in the provision of resource management solutions to the further and higher education sectors.
Talk Title - Scheduling in an unknown, diverse consumer world
This talk will look at how the needs of higher education for timetabling systems have changed since Syllabus Plus was first released in 1991 to today when it is used by nearly 500 of the largest Universities in the world, in 6 continents and 34 countries. It will describe our observations of how the sector is undergoing fundamental transformation in its role, funding sources, management, processes, technology and culture thereby impacting the way institutions operate around the globe. It will conclude that the current approaches to timetabling will need to be transformed to meet the future needs of institutions. It will offer a glimpse of Scientia's latest innovative approaches to timetabling as we put the finishing touches to our next generation of intelligent software applications.
