Autumn 2023
9 November 2023
https://www.maastrichtuniversity.nl/events/dutch-day-optimization
Tapijnkazerne F0.019
International speakers:
Rebecca Reiffenhäuser (UvA) - Combinatorial Online Selection with Minimal Priors
Clara Stegehuis - Optimal structures in random graphs
Jean Cardinal (Brussels) - Flip Graphs and Matroids
12 December 2023 -13.00 - 15.00 PM TS 53 A1.23
OR lunch seminar
Speaker: Karol Węgrzycki - Max Planck Institute for Informatics (Saarbrücken) - "Parameterized Approximation for Maximum Weight Independent Set of Rectangles and Segments"
2 November 2023 - - 16:00-17:00, room TS53 A0.23
Speaker: Lukas Hoesch - Vrije Universiteit Amsterdam - Locally Robust Inference for Non-Gaussian SVAR Models
9 November 2023- 16:00-17:00, room TS53 A0.23
Speaker: Nabil Bouamara - Université catholique de Louvain - A Stepwise Cauchy Combination Test for Multiple Testing Problems with Financial Applications
16 November 2023 - 16:00-17:00, room TS53 H0.04
Speaker: Julian Ashwin - Maastricht University - A Method to Scale-Up Interpretative Qualitative Analysis, with an Application to Aspirations in Cox's Bazaar, Bangladesh
23 November 2023 -16.00-17.00 PM
Speaker: Johannes Lederer- Ruhr-University Bochum
30 November 2023 -17.00-18.00 PM - online Seminar in the Virtual Time Series Seminar (VTSS) Series
Speaker : Stephan Smeekes - Maastricht University - Inference in Non-stationary High-Dimensional VARs
More information on the talk and how to register for the zoom session is available on https://sites.google.com/view/vtss/home
7 December 2023- 16.00-17.00 PM TS53 H0.04
Speaker: Julien Hambuckers – University of Liège - Efficient estimation in extreme value regression models of hedge fund tail risks. For more details: https://arxiv.org/abs/2304.06950
14 December 2023- 16.00-17.00 PM TS53 A0.23
Speaker: Lucas Harlaar - Maastricht University - Statistical Early Warning models with applications
Workshop on noncausal models
(post Francesco Giancaterini’s morning PhD defense)
04 December 2023- 14.00 - 17.00 PM TS53 SBE H0.04
Programme:
2:00-2:15: Alain Hecq (UM). Opening and intro: MAR for Dummies.
2:15- 2:45: Daniel Velasquez (UM), From Past to Future: A Study on noncausal Forecasting for Bitcoin’s Realized Volatility
2:45-3:15: Gabriele Mingoli (VU), Observation-Driven filters for Time Series with Stochastic Trends and Mixed Causal Non-Causal Dynamics
3:15-3:30 : cake/coffee break
3:30-3:45: Tomas Del Barrio Castro (Universitat de les Illes Balears): A note on seasonal MARs
3:45-4:15: Arthur Thomas (Universite Paris Dauphine), Learning the predictive density of mixed causal ARMA processes
4:15-4:45: Sean Telg (VU) Fractional Integration in Mixed Causal-Noncausal Models
Winter 2023 Spring 2024
MLSE is a (mostly) bi-weekly seminar to foster cooperation between the Department of Microeconomics and Public Economics and the Department of Quantitative Economics. It aims to give researchers the opportunity to present their ongoing work and to facilitate cooperation
23 April 2024 -12:30-13:30 PM -TS53 A1.23
Authors: Andries Vermeulen (joint with János Flesch, Chris Kops, and Anna Zseleva)
Title: A general definition of perfect equilibrium
Abstract:
We propose a definition of perfect equilibrium that can be applied to a wide class of games in strategic form. The two key features in the definition are a) using nets instead of sequences, b) using a new concept of completely mixed nets of strategies, based on a more detailed interpretation of the notion of a carrier of a strategy. In the case of finite action sets our definition coincides with the original definition of Selten (1975), and in the compact-continuous case our definition yields a nonempty and compact set of perfect equilibria, which are all weak perfect equilibria according to the definition of Simon and Stinchcombe (1995).
We present several examples which both motivate and illustrate our definition, including games with discontinuous payoffs as well as games played with finitely additive strategies where our definition has a bite.
06 February 2024 -12.30 - PM TS 53 Room H0.06
Speaker: Marc Schröder, Maastricht University, a whiteboard talk on "Minimizing total completion time with machine-dependent priority lists"
Abstract:
We consider a natural, yet challenging variant of the parallel machine scheduling problem in which each machine imposes a preferential order over the jobs and schedules the jobs accordingly once assigned to it. We study the computational complexity of the problem of minimizing the total completion time. Joint work with Vipin Ravindran Vijayalakshmi and Tami Tamir.
Operations Research Seminars
I4CS THREE INVITED TALKS
All colleagues from KE are invited to attend these talks. However, given the capacity of the rooms, there are limited seats.
Wednesday June 11 - 9:15-10:15 AM TS53 A0.23
Anna Wilbik - Maastricht University- "Responsibility and Explainability in Automated Systems"
Thursday June 12 -12:45-13:45 PM TS53 A.023
Klaus-Dieter Rest -University of Natural Resources and Life Sciences, Vienna- " Decision Support for Home Health Care Services in Urban Regions"
Friday June 13 - 11:30 AM -12:30 PM TS53 A.024
Rob van der Mei - CWI/VU Amsterdam- "Saving Lives with Mathematics: the Bumpy Road from Mathematical Modeling to a Successful Company"
Abstracts can be found on http://www.i4cs-conference.org/
Friday June 13 - 12:00 AM -13:30 PM TS53 H0.06
Whiteboard talk by Leo Krull - UM. - "Minimizing Tardy Processing Time on a Single Machine in Near-Linear Time"
Abstract:
this work, we revisit the elementary scheduling problem $1||\sum p_j U_j$. The goal is to select, among $n$ jobs with processing times and due dates, a subset of jobs with maximum total processing time that can be scheduled in sequence without violating their due dates. This problem is NP-hard, but a classical algorithm by Lawler and Moore from the 60s solves this problem in pseudo-polynomial time $O(nP)$, where $P$ is the total processing time of all jobs. With the aim to develop best-possible pseudo-polynomial-time algorithms, a recent wave of results has improved Lawler and Moore's algorithm for $1||\sum p_j U_j$: First to time $\widetilde O(P^{7/4})$ [Bringmann, Fischer, Hermelin, Shabtay, Wellnitz; ICALP'20], then to time $\widetilde O(P^{5/3})$ [Klein, Polak, Rohwedder; SODA'23], and finally to time $\widetilde O(P^{7/5})$ [Schieber, Sitaraman; WADS'23]. It remained an exciting open question whether these works can be improved further.
In this work, we develop an algorithm in near-linear time $\widetilde O(P)$ for the $1||\sum p_j U_j$ problem. This running time not only significantly improves upon the previous results, but also matches conditional lower bounds based on the Strong Exponential Time Hypothesis or the Set Cover Hypothesis and is therefore likely optimal (up to subpolynomial factors). Our new algorithm also extends to the case of $m$ machines in time $\widetilde O(P^m)$. In contrast to the previous improvements, we take a different, more direct approach inspired by the recent reductions from Modular Subset Sum to dynamic string problems. We thereby arrive at a satisfyingly simple algorithm.
Joint work with Nick Fischer
09 January 2024 -13.00 - 15.00 PM TS 53 Room A1.23
Speakers: both from the Operations Research group at Maastricht University
Ashkan Safari - A k-swap Local Search for Make pan Scheduling"
Arman Rouhani - An improved bound for the price of anarchy for related machine scheduling
Abstracts:
Talk 1
Local search is a technique for tackling challenging optimization problems, offering significant advantages in terms of computational efficiency and exhibiting strong empirical behavior across a wide range of problem domains.
In this talk, we address a scheduling problem on two identical parallel machines with the objective of makespan minimization. For this problem, we consider a local search neighborhood, called k-swap, which is a more generalized version of the widely-used swap and jump neighborhoods.
The k-swap neighborhood is obtained by swapping at most k jobs between two machines in our schedule.
First, we propose an algorithm for finding an improving neighbor in the k-swap neighborhood which is faster than the naive approach, and prove an almost matching lower bound on any such an algorithm.
Then, we analyze the number of local search steps required to converge to a local optimum with respect to the k-swap neighborhood. For the case k = 2 (similar to the swap neighborhood), we provide a polynomial upper bound on the number of local search steps, and for the case k = 3, we provide an exponential lower bound.
Finally, we provide computational experiments on various families of instances, and we discuss extensions to more than two machines in our schedule.
This talk is based on a joint work with Lars Rohwedder and Tjark Vredeveld.
Talk 2
In this paper, we introduce an improved upper bound for the efficiency of Nash equilibria in utilitarian scheduling games on related machines. The machines have varying speeds and adhere to the Shortest Processing Time (SPT) policy as the global order for job processing. The goal of each job is to minimize its completion time, while the social objective is to minimize the sum of completion times. Our main result provides an upper bound of $2-\frac{1}{2\cdot(2m-1)}$ on the price of anarchy for the general case of $m$ machines. We improve this bound to 3/2 for the case of two machines, and to $2-\frac{1}{2\cdot m}$ for the general case of $m$ machines when the machines have divisible speeds.
07 March 2024 - 13.00- 14.00 PM TS49A 0.008-0.009
Speaker: Lissa Melis, Maastricht University - "Delivering efficiency: the parcel locker puzzle"
Abstract:
The presentation will explore ongoing work on the implementation of parcel lockers as a solution for urban last-mile delivery challenges. To effectively implement parcel lockers, decisions must be made regarding the number of lockers, their optimal locations, and internal configurations. Existing research has not fully integrated these decisions with vehicle routes and flexible delivery options, i.e., home or nearby parcel locker delivery.
To address this gap, we will define the parcel locker location-configuration-routing problem with flexible deliveries as an optimization challenge. The solution approach combines a population-based incremental learning algorithm with an embedded adaptive large neighbourhood search heuristic, considering both carrier and customer perspectives to achieve a balanced solution.
We will discuss the algorithm’s performance as well as a sensitivity analysis of the different cost components of the problem’s solution. I will end the presentation with some open research questions regarding the problem.
14 March 2024 - 13.00- 14.00 PM TS53 C-1.03
Speaker: Moritz Buchem, TU Munich, formerly Maastricht University - "Improved approximation algorithms for (variants of) the minimum sum of radii problem"
Abstract:
Clustering is a fundamental problem setting with applications in many different areas. For a given set of points in a metric space and an integer k, we seek to partition the given points into k clusters. For each computed cluster, one typically defines one point as the center of the cluster. A natural objective is to minimize the sum of the cluster center's radii, where we assign the smallest radius r to each center such that each point in the cluster is at a distance of at most r from the center. In this talk, I will present improved approximation algorithms for variants of this problem. Our algorithms are primal-dual algorithms that use fundamentally new ideas to compute solutions and to guarantee the claimed approximation ratios.
This is joint work with Katja Ettmayr, Hugo Kooki Kasuya Rosado and Andreas Wiese.
09 April 2024 - 13.00- 14.00 PM TS 53 Room A1.23
Speaker: Aida Abiad , TU Eindhoven - "eigenvalue bounds for the independence and chromatic number of graph powers and its applications"
Abstract:
Spectral graph theory looks at the connection between the eigenvalues of a matrix associated with a graph and the corresponding structures of a graph. In this talk we will show how spectral methods provide a handy tool for obtaining results concerning the structure of graphs.
In particular, we will derive eigenvalue bounds on several NP-hard distance-type graph parameters such as the independence number and the distance chromatic number of graph powers (where the k-th graph power is a graph which has the same vertex set as the original graph G, with two vertices adjacent if and only if they are at distance at most k in G).
We will then see how to use polynomials and mixed integer linear programming in order to optimize such bounds. Finally, some applications to other fields like coding theory will be discussed.
Tuesday 7 May 2024 - 13.00 -14.00 PM TS53 H0.06
Speaker: Lars Rohwedder - UM - "Sensitivity, Proximity and FPT Algorithms for Exact Matroid Problems"
Abstract:
With Kruskal's algorithm one can easily find a maximum (or minimum) weight spanning tree, but finding a spanning tree of exactly a given weight proves to be much harder. Since the problem becomes NP-hard we assume that weights are small integers. We present an FPT algorithm parameterized by the size of the weights even for the more general problem of finding an exact weight basis of an arbitrary matroid. This extends to m linear constraints if we parameterize also on m. Our main technical contributions are new proximity and sensitivity bounds for matroid problems, independent of the number of elements. These bounds imply FPT algorithms via matroid intersection. Prior knowledge on matroids is not required.
This is joint work with Friedrich Eisenbrand and Karol Węgrzecki.
Econometric Seminars
08 February 2024 15.00 - 16.00 PM TS53 room H0.04
Speaker: Julia Schaumburg - Vrije Universiteit Amsterdam - Vector autoregressions with dynamic factor coefficients and conditionally heteroskedastic errors
Abstract:
We introduce a new and general methodology for analyzing vector autoregressive (VAR) models with time-varying coefficient matrices and conditionally heteroskedastic disturbances. The proposed approach is transparent and simple to implement. Our model allows us to derive well-defined impulse response functions that rely on the overall stability of the system. Besides illustrating the good finite sample properties of the model in a simulation study, we provide an empirical illustration analyzing possibly time-varying relationships between U.S. industrial production, inflation, and bond spread. We empirically identify a time-varying linkage between economic and financial variables which are effectively described by a common dynamic factor. The impulse response analysis points towards substantial differences in the effects of financial shocks on output and inflation during crisis and non-crisis periods. The results also illustrate how fixing the VAR coefficients in the derivation of the impulse responses leads to a sizable underestimation of the impact of a financial shock on output and inflation during some of the crises in our sample.
22 February 2024 15.00 - 16.00 PM TS53 room H0.04
Speaker: Daniel Velasquez -Maastricht University - Conditional Heteroscedasticity in Noncausal Autoregressive Models: Estimation and Identification
Abstract:
This paper explores the integration of conditional heteroskedasticity into noncausal autoregressive (AR) models, enhancing their ability to capture extreme values even when the underlying error distribution has light tails—a critical feature for accurately modeling financial time series. This model diverges from the conventional AR-GARCH by acknowledging the misalignment between the sigma fields of the observed variable and the innovations. This divergence stems from the model's structure, where conditional volatility depends on past values, but the autoregressive process depends on future values. Faced with the absence of a straightforward estimation method within this semi-parametric setting, we propose a novel estimator inspired by the Quasi-Maximum Likelihood (QML). However, the standard proofs for consistency do not apply. In this way, our investigation delineates the necessary conditions for achieving strong consistency and asymptotic normality. Our analysis suggests that under the symmetry condition of innovations, the estimator is consistent, and it achieves asymptotic normality when the sixth moment of innovations and the fourth moment of the GARCH process are finite. These theoretical advancements are empirically validated through an extensive Montecarlo study, obtaining unbiased estimated parameters and a decreasing variance as the sample size increases. This work expands the econometric toolkit for modeling financial time series with noncausal AR models and provides valuable methodological insights for applying these models in empirical research.
07 March 2024 15.00 - 16.00 PM TS49A room 0.008 and 0.009
Speaker: Jonas Lembrechts - Universiteit Antwerpen - Microclimate change - the long-overlooked key to understanding the impacts of climate change on nature
Abstract:
Recent research has underscored the complexity of climate change impacts on nature, revealing unexpected delays and shifts in species distributions. While several mechanisms have been proposed to explain these mismatches, one crucial aspect has remained largely overlooked: the role of microclimate change. Unlike macroclimate change, which is commonly monitored through weather station data, microclimate change at ground level or beneath vegetation can vary significantly and remains poorly understood.
In this talk, we address the critical need for relevant baseline data on microclimate change to accurately assess its impacts on nature. I argue that existing climate change data are simply inadequate for this task, as organisms respond primarily to microclimate changes. However, the rate at which this microclimate changes remains largely a mystery, as it is influenced by both climate change and land use changes at the same time.
We introduce the SoilTemp database, comprising over 75,000 in-situ measured microclimate time series, as a valuable resource for improving global microclimate products. By bridging the gap between climate research and ecology, this talk aims to highlight the urgency of understanding microclimate change for climate modelers and ecologists alike. Only by accurately modeling climate where it matters most for nature can we generate precise estimates of climate change impacts on biodiversity, essential for adapting biodiversity management strategies to a rapidly changing world.
14 March 2024 15.00 - 16.00 PM TS53
Speaker: Arthur Thomas - Paris Dauphine University - Path prediction of aggregate alpha-stable moving averages using semi-norm representations.
Co-author: Sébastien Fries
Abstract: This paper studies the conditional distribution of future paths of a two-sided alpha-stable moving average, given a piece of the observed trajectory and when the process is far from its central values. In this framework, vectors of the observed trajectory (t-m to t) and the forecast path up to a horizon h are multivariate alpha-stable, and the dependence between the past and future components is encoded in their spectral measures. A new representation of stable random vectors on unit cylinders and an appropriate semi-norm are proposed to describe the tail behaviour of the vectors when only the first m+1 components are assumed to be observed and large in norm. Not all stable vectors admit such a representation, and the process must be <<anticipative enough>> to admit one. The conditional distribution of future paths can then be explicitly derived using the regularly varying tails property of stable vectors and has a natural interpretation in terms of pattern identification. The approach is extended to processes resulting from the linear combination of stable moving averages, which have much richer dynamics, and applied to several examples. We propose a series of MC simulations to ensure the reliability of the formula, and an application to the prediction of El Nino/La Nina occurrence is proposed.