@misc{AndersonHendelLeBodicetal.2019, author = {Anderson, Daniel and Hendel, Gregor and Le Bodic, Pierre and Viernickel, Jan Merlin}, title = {Clairvoyant Restarts in Branch-and-Bound Search Using Online Tree-Size Estimation}, issn = {1438-0064}, url = {http://nbn-resolving.de/urn:nbn:de:0297-zib-72653}, year = {2019}, abstract = {We propose a simple and general online method to measure the search progress within the Branch-and-Bound algorithm, from which we estimate the size of the remaining search tree. We then show how this information can help solvers algorithmically at runtime by designing a restart strategy for Mixed-Integer Programming (MIP) solvers that decides whether to restart the search based on the current estimate of the number of remaining nodes in the tree. We refer to this type of algorithm as clairvoyant. Our clairvoyant restart strategy outperforms a state-of-the-art solver on a large set of publicly available MIP benchmark instances. It is implemented in the MIP solver SCIP and will be available in future releases.}, language = {en} } @misc{GleixnerBertholdMuelleretal.2016, author = {Gleixner, Ambros and Berthold, Timo and M{\"u}ller, Benjamin and Weltge, Stefan}, title = {Three Enhancements for Optimization-Based Bound Tightening}, issn = {1438-0064}, doi = {10.1007/s10898-016-0450-4}, url = {http://nbn-resolving.de/urn:nbn:de:0297-zib-57803}, year = {2016}, abstract = {Optimization-based bound tightening (OBBT) is one of the most effective procedures to reduce variable domains of nonconvex mixed-integer nonlinear programs (MINLPs). At the same time it is one of the most expensive bound tightening procedures, since it solves auxiliary linear programs (LPs)—up to twice the number of variables many. The main goal of this paper is to discuss algorithmic techniques for an efficient implementation of OBBT. Most state-of-the-art MINLP solvers apply some restricted version of OBBT and it seems to be common belief that OBBT is beneficial if only one is able to keep its computational cost under control. To this end, we introduce three techniques to increase the efficiency of OBBT: filtering strategies to reduce the number of solved LPs, ordering heuristics to exploit simplex warm starts, and the generation of Lagrangian variable bounds (LVBs). The propagation of LVBs during tree search is a fast approximation to OBBT without the need to solve auxiliary LPs. We conduct extensive computational experiments on MINLPLib2. Our results indicate that OBBT is most beneficial on hard instances, for which we observe a speedup of 17\% to 19\% on average. Most importantly, more instances can be solved when using OBBT.}, language = {en} } @misc{GleixnerEiflerGallyetal.2017, author = {Gleixner, Ambros and Eifler, Leon and Gally, Tristan and Gamrath, Gerald and Gemander, Patrick and Gottwald, Robert Lion and Hendel, Gregor and Hojny, Christopher and Koch, Thorsten and Miltenberger, Matthias and M{\"u}ller, Benjamin and Pfetsch, Marc and Puchert, Christian and Rehfeldt, Daniel and Schl{\"o}sser, Franziska and Serrano, Felipe and Shinano, Yuji and Viernickel, Jan Merlin and Vigerske, Stefan and Weninger, Dieter and Witt, Jonas T. and Witzig, Jakob}, title = {The SCIP Optimization Suite 5.0}, issn = {1438-0064}, url = {http://nbn-resolving.de/urn:nbn:de:0297-zib-66297}, year = {2017}, abstract = {This article describes new features and enhanced algorithms made available in version 5.0 of the SCIP Optimization Suite. In its central component, the constraint integer programming solver SCIP, remarkable performance improvements have been achieved for solving mixed-integer linear and nonlinear programs. On MIPs, SCIP 5.0 is about 41 \% faster than SCIP 4.0 and over twice as fast on instances that take at least 100 seconds to solve. For MINLP, SCIP 5.0 is about 17 \% faster overall and 23 \% faster on instances that take at least 100 seconds to solve. This boost is due to algorithmic advances in several parts of the solver such as cutting plane generation and management, a new adaptive coordination of large neighborhood search heuristics, symmetry handling, and strengthened McCormick relaxations for bilinear terms in MINLPs. Besides discussing the theoretical background and the implementational aspects of these developments, the report describes recent additions for the other software packages connected to SCIP, in particular for the LP solver SoPlex, the Steiner tree solver SCIP-Jack, the MISDP solver SCIP-SDP, and the parallelization framework UG.}, language = {en} } @misc{BorndoerferLindnerRoth2019, author = {Bornd{\"o}rfer, Ralf and Lindner, Niels and Roth, Sarah}, title = {A Concurrent Approach to the Periodic Event Scheduling Problem}, issn = {1438-0064}, url = {http://nbn-resolving.de/urn:nbn:de:0297-zib-71907}, year = {2019}, abstract = {We introduce a concurrent solver for the periodic event scheduling problem (PESP). It combines mixed integer programming techniques, the modulo network simplex method, satisfiability approaches, and a new heuristic based on maximum cuts. Running these components in parallel speeds up the overall solution process. This enables us to significantly improve the current upper and lower bounds for all benchmark instances of the library PESPlib.}, language = {en} } @misc{ShinanoAchterbergBertholdetal.2015, author = {Shinano, Yuji and Achterberg, Tobias and Berthold, Timo and Heinz, Stefan and Koch, Thorsten and Winkler, Michael}, title = {Solving Open MIP Instances with ParaSCIP on Supercomputers using up to 80,000 Cores}, issn = {1438-0064}, doi = {10.1109/IPDPS.2016.56}, url = {http://nbn-resolving.de/urn:nbn:de:0297-zib-56404}, year = {2015}, abstract = {This paper describes how we solved 12 previously unsolved mixed-integer program- ming (MIP) instances from the MIPLIB benchmark sets. To achieve these results we used an enhanced version of ParaSCIP, setting a new record for the largest scale MIP computation: up to 80,000 cores in parallel on the Titan supercomputer. In this paper we describe the basic parallelization mechanism of ParaSCIP, improvements of the dynamic load balancing and novel techniques to exploit the power of parallelization for MIP solving. We give a detailed overview of computing times and statistics for solving open MIPLIB instances.}, language = {en} } @misc{MunguiaOxberryRajanetal.2017, author = {Munguia, Lluis-Miquel and Oxberry, Geoffrey and Rajan, Deepak and Shinano, Yuji}, title = {Parallel PIPS-SBB: Multi-Level Parallelism For Stochastic Mixed-Integer Programs}, number = {ZIB-Report 17-58}, issn = {1438-0064}, doi = {10.1007/s10589-019-00074-0}, url = {http://nbn-resolving.de/urn:nbn:de:0297-zib-65517}, year = {2017}, abstract = {PIPS-SBB is a distributed-memory parallel solver with a scalable data distribution paradigm. It is designed to solve MIPs with a dual-block angular structure, which is characteristic of deterministic-equivalent Stochastic Mixed-Integer Programs (SMIPs). In this paper, we present two different parallelizations of Branch \& Bound (B\&B), implementing both as extensions of PIPS-SBB, thus adding an additional layer of parallelism. In the first of the proposed frameworks, PIPS-PSBB, the coordination and load-balancing of the different optimization workers is done in a decentralized fashion. This new framework is designed to ensure all available cores are processing the most promising parts of the B\&B tree. The second, ug[PIPS-SBB,MPI], is a parallel implementation using the Ubiquity Generator (UG), a universal framework for parallelizing B\&B tree search that has been successfully applied to other MIP solvers. We show the effects of leveraging multiple levels of parallelism in potentially improving scaling performance beyond thousands of cores.}, language = {en} } @masterthesis{Noeckel2018, type = {Bachelor Thesis}, author = {N{\"o}ckel, Celine}, title = {Bidirectional A* Search on Time-Dependent Airway Networks}, pages = {51}, year = {2018}, abstract = {This thesis deals with a new algorithm for finding Shortest Paths on Airway Networks. It is about a Bidirectional A* Search, a Greedy algorithm exploring a network from two sides instead of one. We will use it to solve the so-called 'Horizontal Flight Trajectory Problem', where one searches for an aircraft trajectory between two airports of minimal costs on an Airway Network. The given network will be modeled as a directed graph and in order to reflect reality we concentrate on the dynamic version. Here a timedependent cost function for all arcs is integrated, that shall represent the winds blowing. This way we model the Horizontal Flight Trajectory Problem mathematically as a Time-Dependent Shortest Path Problem. The basic algorithm idea derives from the algorithm presented in 'Bidirectional A* Search on Time-Dependent Road Networks' [1], where a similar setting is elaborated for road networks. The algorithm procedure bears on a modified generalization of Dijkstra's algorithm, made bidirectional and improved in several aspects. As for the backwards search the arrival times are not known in advance, the reversed graph it occurs on has to be weighted by a lower bound. Contrary to the static case the forwards search still has to go on, when they 'meet' in one node. In the static case, the shortest path would have been found at this point. For road networks the TDSPP is well-studied, for airway networks cannot be found as much in literature. In order to test efficiency, we implement Dijkstra's algorithm, unidirectional A* Search and Bidirectional A* Search. We draw up how potential functions for the static case could look like and that with a suitable potential A* Search with works approx. 7 times faster than Dijkstra in the dynamic case. Our computations lead also to the result, that the unidirectional A* Search works even better on the network than our new bidirectional approach does. On average it labels fewer nodes and also yields 1,7 times faster to the solutions. For assessing the efficiency of the different algorithms we compare the running times and to exclude processor characteristics we consider also the set labels relative to the labels on the resulting optimal path. In addition, we present examples of routes visually and explain shortly why there appear local differences regarding performance of A* Search and Bidirectional A* Search.}, language = {en} } @masterthesis{Vornberger2018, type = {Bachelor Thesis}, author = {Vornberger, Leo}, title = {Approximation von Windkomponenten in der Luftfahrt durch lineare Interpolation}, url = {http://nbn-resolving.de/urn:nbn:de:0297-zib-71570}, pages = {49}, year = {2018}, abstract = {Das Wind-Interpolation-Problem (WIP) ist ein bisher selten diskutiertes Problem der Flugplanungsoptimierung, bei dem es darum geht, Wind-Komponenten auf einer Luftstraße zu approximieren. Anhand von Winddaten, die vektoriell an den Gitterpunkten eines den Globus umspannenden Gitters vorliegen, soll bestimmt werden, wie viel Wind entlang der Luftstraße und quer zu ihr weht. Thema dieser Arbeit ist ein Spezialfall des WIP, n{\"a}mlich das statische WIP auf einer Planfl{\"a}che (SWIPP). Dazu wird zuerst ein Algorithmus besprochen, der das SWIPP zwar l{\"o}st, aber einem Ansatz zugrunde liegt, der bei genauerem Hinsehen nicht sinnvoll erscheint: hier wird Wind zwischen vier Punkten interpoliert, wozu es keine triviale Methode gibt. {\"A}hnlich zu diesem Algorithmus, der heute als State-of-the-Art gilt, wird als Ergebnis dieser Arbeit ein neuer Algorithmus vorgestellt, der das SWIPP akkurater und schneller l{\"o}st. Hier wird deutlich seltener auf die Interpolation zwischen vier Punkten zur{\"u}ckgegriffen - stattdessen wird fast immer linear zwischen zwei Punkten interpoliert. Die Algorithmen zum L{\"o}sen des SWIPP werden auf ihre Genauigkeit, asymptotische Laufzeit und Geschwindigkeit untersucht und verglichen. Als Testareal dienen zum einen echte Wetterdaten sowie das Luftstraßennetz, das die Erde umspannt, und zum anderen ein eigens generiertes Windfeld und fiktive Luftstraßen. Es wird gezeigt, dass der hier vorgestellte Algorithmus die State-of-the-Art-Variante in allen genannten Aspekten {\"u}bertrifft.}, language = {de} } @masterthesis{Akil2018, type = {Bachelor Thesis}, author = {Akil, Fatima}, title = {Lineare Gleichungssysteme modulo T}, url = {http://nbn-resolving.de/urn:nbn:de:0297-zib-71560}, pages = {52}, year = {2018}, abstract = {Mit dem Voranschreiten der Technologie erhalten die {\"o}ffentlichen Verkehrsmittel eine gr{\"o}ßere Bedeutung. Die Bef{\"o}rderung mehrerer Personen er{\"o}ffnet der Gesellschaft viele M{\"o}glichkeiten, unter Anderem den Vorteil der Zeitersparnis. Die Dauer des Verkehrswegs mit {\"o}ffentlichen Verkehrsmitteln ist h{\"a}ufig geringer, als die mit individuellen Verkehrsmitteln. Jedes {\"o}ffentliche Transportmittel ist mit einem Fahrplan versehen. Dieser bietet Passagieren, die {\"o}ffentliche Verkehrsmittel {\"o}fter nutzen, eine Strukturierung und Planung ihrer Zeit. Dabei lassen sich Taktfahrpl{\"a}ne aufgrund ihres periodischen Verhaltens leicht einpr{\"a}gen. Dieses periodische Verhalten ist durch mathematische Modellierungen darstellbar. Das pers{\"o}nliche Nutzverhalten vieler B{\"u}rger im Personenverkehr ist auf die {\"o}ffentlichen Verkehrsmittel beschr{\"a}nkt. Diese beinhalten im Gegensatz zum individuellen Verkehrsmittel eine Wartezeit. Dabei stellt sich die Frage, ob man anhand mathematischer Modelle diese Wartezeit minimieren kann. Eine bekannte mathematische Modellierung dieses Problems ist das Periodic Event Scheduling Problem (PESP). Die optimale Planung eines periodischen Taktfahrplanes steht im Vordergrund. W{\"a}hrend ich dieses Problem betrachtet habe, wurde ich auf das Rechnen mit linearen Gleichungssystemen modulo T aufmerksam. Bei periodischen Taktfahrpl{\"a}nen wird ein einheitliches zeitliches Muster, welches sich nach T Minuten wiederholt, betrachtet. Das dabei zu betrachtende L{\"o}sungsproblem er{\"o}ffnet ein Teilgebiet der Mathematik, welches bislang nicht im Vordergrund stand: Das L{\"o}sen linearer Gleichungen modulo T, wobei T f{\"u}r die Zeit in Minuten steht und somit 60 ist. Da 60 keine Primzahl ist, kann - wie im Laufe der Arbeit pr{\"a}sentiert - das lineare Gleichungssystem nicht mehr {\"u}ber einen K{\"o}rper gel{\"o}st werden. Lineare Gleichungssysteme werden nun {\"u}ber Nicht-K{\"o}rpern betrachtet. Die Literatur weist sowohl im deutschsprachigem als auch im englischsprachigen Raum wenig Umfang bez{\"u}glich linearer Gleichungssysteme {\"u}ber Nicht-K{\"o}rper auf. Der Bestand an Fachliteratur bez{\"u}glich den Themen lineare diophantische Gleichungssysteme, Hermite- Normalform und Smith-Normalform ist zurzeit gering, dennoch erreichbar, beispielsweise in [1], welches in dieser Bachelorarbeit genutzt wurde. Insbesondere wurde ich bei der Suche nach geeigneter Literatur zu linearen Gleichungssystemen {\"u}ber Restklassenringe, die keinen K{\"o}rper bilden, nicht f{\"u}ndig. Dabei recherchierte ich sowohl in den Universit{\"a}tsbibliotheken als auch in webbasierenden Suchmaschinen. Aufgrund dem geringen Bestand an Fachliteratur in diesem Kontext, war ich gezwungen, an vielen Stellen eigene logische Verkn{\"u}pfungen zu konzipieren und zu beweisen. Dies brachte viele Schwierigkeiten mit sich, die mit bestm{\"o}glichem Verst{\"a}ndnis bearbeitet wurden. Abseits der Zug{\"a}nglichkeit der Literatur, finde ich es sehr {\"u}berraschend, dass sich viele Professoren der Mathematik mit diesem Themenbereich nicht besch{\"a}ftigten. Insbesondere gingen von den Dozenten, die ich um Literaturempfehlung bat, kein Werk aus. Damit wurde das Thema "Lineare Gleichungssysteme Modulo T" einerseits eine große Herausforderung, andererseits eine große Motivation, da ich mit dieser Bachelorarbeit vielen Interessenten der Mathematik als Sekund{\"a}rliteratur dienen kann.}, language = {de} } @misc{BorndoerferKarbsteinLiebchenetal.2018, author = {Bornd{\"o}rfer, Ralf and Karbstein, Marika and Liebchen, Christian and Lindner, Niels}, title = {A Simple Way to Compute the Number of Vehicles That Are Required to Operate a Periodic Timetable}, issn = {1438-0064}, doi = {10.4230/OASIcs.ATMOS.2018.16}, url = {http://nbn-resolving.de/urn:nbn:de:0297-zib-69688}, year = {2018}, abstract = {We consider the following planning problem in public transportation: Given a periodic timetable, how many vehicles are required to operate it? In [9], for this sequential approach, it is proposed to first expand the periodic timetable over time, and then answer the above question by solving a flow-based aperiodic optimization problem. In this contribution we propose to keep the compact periodic representation of the timetable and simply solve a particular perfect matching problem. For practical networks, it is very much likely that the matching problem decomposes into several connected components. Our key observation is that there is no need to change any turnaround decision for the vehicles of a line during the day, as long as the timetable stays exactly the same.}, language = {en} } @misc{BorndoerferReutherSchlechteetal.2011, author = {Bornd{\"o}rfer, Ralf and Reuther, Markus and Schlechte, Thomas and Weider, Steffen}, title = {A Hypergraph Model for Railway Vehicle Rotation Planning}, issn = {1438-0064}, doi = {/10.4230/OASIcs.ATMOS.2011.146}, url = {http://nbn-resolving.de/urn:nbn:de:0030-drops-32746}, number = {11-36}, year = {2011}, abstract = {We propose a model for the integrated optimization of vehicle rotations and vehicle compositions in long distance railway passenger transport. The main contribution of the paper is a hypergraph model that is able to handle the challenging technical requirements as well as very general stipulations with respect to the ``regularity'' of a schedule. The hypergraph model directly generalizes network flow models, replacing arcs with hyperarcs. Although NP-hard in general, the model is computationally well-behaved in practice. High quality solutions can be produced in reasonable time using high performance Integer Programming techniques, in particular, column generation and rapid branching. We show that, in this way, large-scale real world instances of our cooperation partner DB Fernverkehr can be solved.}, language = {en} } @misc{Oleynikova2018, type = {Master Thesis}, author = {Oleynikova, Ekaterina}, title = {Mathematical optimization of joint order batching and picker routing problems}, pages = {60}, year = {2018}, abstract = {In this thesis we study order picking optimization problems for a two-blocks rectangle warehouse layout. We present combinatorial formulations and linear programming models based on the Steiner graph representation for order batching, picker routing, and joint order batching and picker routing problems. A special case of the latter is considered. This case assumes that each order contains exactly one item and each item can be picked from different possible locations in a warehouse. The underlying optimization problem is called joint multi-location order batching and picker routing problem (JMLOBPRP). Since having only one-item orders turns the JMLOBPRP into a special case of a capacitated vehicle routing problem, we suggest to implement algorithmic approaches for those to solve the JMLOBPRP. In particular, we define the JMLOBPRP as a generalization of the resource constrained assignment problem, for which a regional search method exists. The intention of the thesis is to investigate how a relaxation of the JMLOBPRP, a so-called group assignment problem (GrAP), can be solved following the ideas of regional search. We present a mathematical model of the GrAP and prove that it is NP-hard. Furthermore, we propose a novel heuristic algorithm for the GrAP. We call this method a network search algorithm, as it is based on a Lagrangian relaxation of the GrAP, which is solved by the network simplex method. On each its iteration network search examines a solution region suggested by the network simplex algorithm and improves the incumbent solution. Numerical experiments are conducted to assess a performance of the network search method. We create more realistic problem instances. The proposed algorithm is compared to the integer optimal solution of the GrAP and optimal fractional solution of its linear relaxation. Both computed using the commercial linear solver Gurobi. Our experiments show that the developed network search algorithm leads to the hight-quality solution within a short computing time. The results obtained testing large problem instances which cannot be solved by Gurobi within a reasonable computing time, show that the network search method provides a solution approach which can be used in practice.}, language = {en} }