|
Abstract and Bio Speakers NGB/LNMB Seminar
Back to school,
learn about the latest developments in Operations Research
Martin Bichler (Technical University of Munich)
Short Bio:
Martin Bichler received his MSc degree from the Technical University of Vienna, and his Ph. D. as well as his Habilitation from the Vienna University of Economics and Business. He was working as a research fellow at UC Berkeley, and as research staff member at the IBM T. J. Watson Research Center, Yorktown Heights, New York. Since 2003 he is full Professor at the Department of Informatics of the Technical University of Munich (TUM), and also a faculty member at the TUM School of Management. He was a visiting scholar at the University of Cambridge (2008), at HP Labs Palo Alto (2008), and at Yale University (2016). Martin is responsible for the Master's program in Information Systems at TUM and a faculty and board member of the Bavarian Elite Master program "Finance and Information Management". He is a fellow of the Agora Group on Market Design at the University of New South Wales, Australia.
(Website Martin Bichler)
Title: About Allocation Problems and Equilibrium Bidding Strategies in Procurement Auctions
Abstract: Equilibrium bidding strategies in first-price combinatorial auctions are not well understood. Ex-post split-award auctions are a widespread form of combinatorial procurement auctions in which the demand for some quantity to be procured is split into two or more shares. Similar to the seminal article by Anton and Yao (1992) who analyze markets with two bidders, we focus on diseconomies of scale, but with an arbitrary number of bidders. These markets are practically relevant and strategically challenging because bidders must coordinate to achieve an efficient outcome. Dutch combinatorial auction formats have received little attention so far. We show that, unlike in single-object auctions, first-price sealed-bid and the Dutch combinatorial auctions are not strategically equivalent. While the former exhibits a coordination problem for bidders, the latter has a unique and efficient equilibrium. Motivated by industry practice, we also analyze a combination of a Dutch and a first-price sealed-bid format that has attractive properties, and the efficiency and purchasing costs of these auctions are compared in a welfare analysis. The price information revealed during the Dutch auction formats avoids equilibrium selection problems and this also helps bidders coordinate in the lab. Overall, the theoretical predictions explain important aspects of the empirical bidder behavior remarkably well, and first-price auctions with more than two bidders achieve high levels of efficiency also in the lab.
Frans Cruijssen (ArgusI)
Short Bio:
Frans is founder and partner of Argusi, a quantitative supply chain advisory with its offices in Breda. Prior to Argusi, Frans worked at TNT Express where he was responsible for the development and global roll-out of optimization tools and solutions. After his studies on Econometrics and Operations Research, Frans conducted a PhD study (Cum Laude) on the topic of horizontal cooperation in transport and logistics. He has published papers in international academic journal on the topics of, amongst others, horizontal cooperation, network design, new logistics concepts, and benchmarking. At Argusi he specializes at projects on horizontal collaboration and tactical and strategic network design.
Title: Game Theory in supply chains
Abstract:
Horizontal collaboration is a rising topic in supply chain logistics. Overall, trucks are currently not filled to their maximum capacities. Looking at weight capacity, the average load factor of cross-border road transport in Europe is actually below 50%. With thousands of individual shippers and thousands of logistics service providers, this is a problem that is obviously difficult to solve. One of the possible answers could be to collaborate much more intensively and bundle flows of multiple companies. European governments heavily support this horizontal collaboration and some successful cases have been realized in recent years. One of the hurdles for lasting collaboration however is the issue of fair gain sharing. Who gets which share of the savings that are realized by the group? Cooperative game theory can be of very much use here.
Herbert Hamers (Tilburg University)
Short Bio:
Herbert Hamers is Professor of Game Theory and Operations Research at the Department of Econometrics and Operations Research of Tilburg University.
One part of his game theoretical research is concerned with studying the interaction between operations research problems and cooperative game theory. In particular, solution concepts and properties of cooperative transferable utility games that correspond to operations research problems are investigated. The following Operation Research games are enclosed in his research: coloring games, sequencing games, Chinese postman games, traveling salesman games, assignment games, lot sizing games, spanning tree games.
Another part of his game theoretical research is devoted to applications of game theory to practice: cost allocation of ATM machines (prescribing an allocation of the costs taking into account location and number of users of different banks), insinking (a method that creates synergetic behaviour by costumers in an logistic environment by tariff determination), development economy (stimulating cooperation in rural households in Nigeria), cooperation in transportation networks and structures. However, the most prominent application of game theory is devoted to terroristic networks. Here structure of such networks as importance of members in these networks is investigated.
Currently, he is also involved in non-game theoretical research projects that are focused to detect waste in hospital billing more efficient and in optimizing market driven portfolios of universities.
He published among others in: Mathematical Programming, European Journal of Operational Research, Social Networks and Games and Economic Behavior.
Some recent publications:
Husslage, B.G.M., Borm, Peter, Burg, T., & Hamers, Herbert (2015). Ranking terrorists in networks: A sensitivity analysis of Al Qaeda's 9/11 attack. Social Networks, 42, 1-7.
Hamers, H.J.M., Miquel, S., & Norde, H.W. (2014). Monotonic stable solutions for minimum coloring games. Mathematical Programming, 145(1-2), 509-529.
Josune Albizuri, M. & Hamers, H. J. M. (2014) Graphs inducing totally balanced and submodular Chinese postman games. Discrete Applied Mathematics, (172), 98-103.
(Website Herbert Hamers)
Title: Terrorism using game theory
Abstract:
All over the world intelligence services are collecting data concerning possible terrorist threats. This information is usually transformed into network structures in which the nodes represent the individuals in the data set and the links possible connections between these individuals. Unfortunately, it is nearly impossible to keep track of all individuals in the resulting complex network. Therefore, Lindelauf et al. (2013) introduced a methodology based on a game theoretic centrality measure, which is innovative in the sense that it takes into account not only the structure of the network but also individual and coalitional characteristics of the members of the network.
Husslage et al. (2015) perform a sensitivity analysis on the rankings derived from this centrality measure for the case of Al Qaeda’s 9/11 attack. In this sensitivity analysis we consider firstly the possible additional information available about members of the network, secondly, variations in relational strength and, finally, the absence or presence of a small percentage of links in the network. We also introduce a case specific method to compare the different rankings that result from the sensitivity analysis and show that the new centrality measure is robust to small changes in the data.
The calculation of the Shapley value plays an important role in this methodology. Since the Al Qaeda 9/11 network can be represented by relatively small networks, the determination of Shapley value is not time consuming. However, in larger networks the Shapley value may not be determined in polynomial time. A small improvement to approximate the Shapley value is discussed.
Emiel Maasland (Erasmus University Rotterdam)
Short Bio:
Dr. Emiel Maasland (Founder of Auctiometrix) is Managing Consultant at Auctiometrix and Affiliate Researcher at the Department of Applied Economics of the Erasmus School of Economics. He holds a PhD in Auction Theory. He has published in international academic journals, such as Journal of Political Economy, Games and Economic Behavior, Economic Theory, Journal of Institutional and Theoretical Economics, and Telecommunications Policy. He has over 15 years of experience in scientifically based contract research. He advised European ministries/regulators on the design and implementation of various (CCA) auctions, in particular spectrum auctions for mobile communication and radio broadcasting. He also has practical experience in advising major telecom operators and broadcasters on their optimal bid strategy in numerous spectrum auctions in Europe and the USA (among which CCA auctions). Recently, he was hired by Agentschap Telecom (the Dutch Radiocommunications Agency) to perform an independent verification on CCA auction software. In the past, he was involved in large scale auction evaluation studies, including an evaluation of the Dutch UMTS-auction for the Dutch Parliament and an evaluation of the auction of petrol station locations along the Dutch highway for the Dutch Ministry of Finance. Furthermore, he advised private companies in antitrust cases in sectors such as the market for mobile telephony, motor fuels, bicycles, and liberal professions.
Title: Game Theory in Spectrum Auctions
Abstract:
Auctions (in particular spectrum auctions) are considered to be the clearest success story in the application of game theory to economics. Since 1994, auction theorists have designed spectrum auctions for governments world-wide. In some of these auctions, the amount of money raised was almost beyond imagination. E.g., in 2000, the British and German governments, together, raised almost 100 billion euros for licenses for third generation mobile telecommunications. The Simultaneous Multi Round Auction (SMRA) and the Combinatorial Clock Auction (CCA) are the two most used auction formats to sell spectrum licenses nowadays. In this talk I will discuss the pros and cons of these formats, provide several SMRA and CCA real-life examples in which firms have gamed the auction, and explain which design features could be held responsible for this gaming behaviour.
Guido Schäfer (CWI/VU University Amsterdam)
Short Bio:
Guido Schäfer is a senior researcher in the Networks and Optimization group at the Centrum Wiskunde & Informatica. He also holds a (part-time) professorship on Algorithmic Game Theory in the Department of Econometrics and Operations Research at Vrije Universiteit Amsterdam. His research interests are in combinatorial optimization and algorithms in general, and in algorithmic game theory in particular.
He pursued his PhD studies at the Max Planck Institute for Informatics in Saarbrücken (2000-2004). After this, he was a postdoctoral research fellow at the Sapienza University of Rome (2004/2005). Before joining CWI in 2009, he was the head of the Independent Research group "Discrete Optimization" at the Institute for Mathematics at the Technical University Berlin (2005-2008). He got his Habilitation from the Technical University Berlin (2009). He was a visiting professor at Carnegie Mellon University (2005), University of Waterloo (2008), Sapienza University of Rome (2011) and University of Chile (2014).
In his current research, he focusses on the study of the existence, computability and inefficiency of equilibria in games. He also works intensively on algorithmic mechanism design, in particular on the the development of coordination mechanisms to reduce equilibrium inefficiencies and on cost sharing mechanisms for combinatorial optimization problems. He is also involved in several cooperations with industry partners.
(Website Guido Schäfer)
Title: Inefficiency of Multi-Unit Auctions
Abstract:
We study two standard multi-unit auction formats for allocating multiple units of a single good to multi-demand bidders. The first one is the Discriminatory Auction, which charges every winner his winning bids. The second one is the Uniform Price Auction, which determines a uniform price to be paid per unit. Variants of both formats find applications ranging from the allocation of state bonds to online sales over the internet. For these formats, we consider two bidding interfaces: (i) standard bidding, which is most prevalent in the scientific literature, and (ii) uniform bidding, which is more common in practice.
We evaluate the economic inefficiency of both multi-unit auction formats for both bidding interfaces by deriving upper and lower bounds on the Price of Anarchy for pure Nash equilibria and mixed Bayes-Nash equilibria. Our developments improve significantly upon bounds that have been obtained for submodular valuation functions. Also, for the first time, we consider bidders with subadditive valuation functions under these auction formats. Our results signify near-efficiency of these auctions, which provides further justification for their use in practice.
Katerina Stankova (Maastricht University)
Short Bio:
Dr. Katerina Stankove is an assistant professor in game theory, with additional expertise in optimization, optimal control, mathematical biology, and adaptive dynamics. She specializes on modeling complex real-world systems and predicting their temporal and spatial behavior, using mathematical techniques.
Her PhD research (at Delft Institute of Applied Mathematics, TU Delft, 2005-2009) focused on Stackelberg and inverse Stackelberg games and on application of these games in traffic control and in energy markets. During her postdoctoral fellowships at INRIA Grenoble - Rhone-Alpes and Delft Center for Systems and Control, TU Delft (2010-2012) she draw her attention to modeling complex biological systems, such as arthropod predator-prey systems or tri-trophic systems plant-herbivore-carnivore, in collaboration with colleagues from The Netherlands, France, and Brazil. She continues working on these research areas in her current position at the Department of Data Science and Knowledge Engineering, Maastricht University (since 2012.
Since 2015 she has been working on modeling cancer growth and treatment, using spatial evolutionary game theory, in collaboration with the Moffitt Cancer Center, Florida, US. She is very much interested in understanding both spatial and temporal characteristics of cancer, using advanced analytical and computational tools to find optimal treatment, and validating the predicted behaviors via histological data.
Title: Games of Cancer
Abstract:
Cancer can be viewed as an evolutionary process and as such can be modeled and explained using evolutionary game theory. In the first part of the talk we will focus on hallmarks of cancer and various models of cancer as a game (e.g. nonspatial model based on differential equations, spatial model using diffusion equations, spatial in-silico model).
Subsequently, we will focus on the cancer treatment. The current standard of care for cancer therapy is to kill the largest possible number of tumor cells by applying the maximum tolerated dose. While this approach is often initially successful at reducing tumor burden, it inevitably fails due to evolution and proliferation of resistant cancer phenotypes. Moreover, this highest tolerable dose regimen is typically very aggressive to the patient and very expensive. Shifting the goal of therapy from complete elimination of tumor burden to instead controlling the tumor burden (with much lower treatment doses) for a maximal period of time - the so-called adaptive therapy - can change the approach to therapy dramatically. We will show that (inverse) Stackelberg game theory can be very helpful in designing such an adaptive treatment.
Onno Zoeter (Booking.com)
Short Bio:
Dr. Onno Zoeter is principal data scientist at Booking.com where he works with a very strong data science team on designing and improving prediction, filtering and ranking models. Before joining Booking.com, Dr. Zoeter led the research on transportation demand management at Xerox's European Research Lab and worked on large scale click- through rate prediction problems at Microsoft Research in Cambridge. His research received numerous awards. Among them the OECD International Transport Forum Innovation Award and the International Parking Institute Award of Excellence. The large scale system for advertisement relevance prediction he co-developed won Microsoft's adPredict competition and was adopted as part of the Bing search engine. Fortune selected him among their 20 Big Data All-Stars in 2014.
Title: E-commerce platforms and optimization: examples and challenges from Booking.com
Abstract:
E-commerce platforms, due to their digital nature, allow (i) for the recording of very detailed customer feedback ("big data") and (ii) fast and cheap experimentation with variants of the customer experience ("A/B testing"). These two ingredients together allow for the systematic optimization of the platform. In this talk I will describe this optimization process and will introduce several open challenges.
|
