Dr.-Ing. Felix Brandt

Abteilungsleiter

Werdegang

Felix Brandt studierte Informatik am Karlsruher Institut für Technologie (KIT) mit den Schwerpunkten Algorithmen- und Softwaretechnik. Von 2010 bis 2018 arbeitete er als wissenschaftlicher Mitarbeiter und Projektleiter am FZI im Forschungsbereich IPE mit dem Schwerpunkt Logistik und Supply-Chain-Optimierung. Seit 2018 leitet er die Abteilung Logistics Systems Engineering (LSE) am FZI.

Felix Brandt ist Experte für Planungs- und Optimierungsfragestellungen aus Produktion und Logistik. Im Rahmen seiner Tätigkeit als Projektmitarbeiter/Projektleiter entwickelt er in zahlreichen Projekten mit KMU und Großunternehmen (u.a. Lufthansa Cargo, PTV Group, Dillinger Hütte, Daimler TSS) innovative Lösungen und bringt diese in die Praxis.

2017 promovierte Felix Brandt am Karlsruher Institut für Technologie (KIT) summa cum laude zum Thema "The Air Cargo Load Planning Problem". In diesem Rahmen erarbeitete er Lösungen für verschiedenste Planungsprobleme in der Luftfracht, u.a. 3D-Ladeplanung, Weight&Balance, Operations Scheduling und Netzwerkplanung.

Seine derzeitigen Forschungsinteressen liegen in der Planung und Optimierung komplexer Systeme durch die Entwicklung effizienter Algorithmen und den Einsatz verschiedenster Methoden der künstlichen Intelligenz (kombinatorische Optimierung, Constraint-Programmierung, Bildverarbeitung). 

Publikationen

Zeitungs- oder Zeitschriftenartikel (3)

1. The air cargo load planning problem - a consolidated problem definition and literature review on related problemsInfoDetails

   Felix Brandt and Stefan Nickel, 2018

   This work describes the operational load planning problems of air cargo carriers. In the literature the term air cargo loading is used ambiguously for different subproblems that appear during the load planning process for a flight. In its entirety the problem has not yet been the subject of published research. This work introduces the overall problem, the planning decisions, the objectives of the involved stakeholders, and the restrictions occurring in practice. It reviews and categorizes related literature dealing with similar problems or subproblems. The typical problem setting in practice is analyzed to get insights into the real world problem and a large benchmark dataset is presented. Finally, challenges of the current planning practice and research gaps are highlighted.

2. Constraint-based large neighborhood search for machine reassignmentInfoDetails

   Felix Brandt and Jochen Speck and Markus Völker, 2014

   This paper addresses a process-to-machine reassignment problem arising in cloud computing environments. The problem formulation has been posed as the ROADEF/EURO challenge 2012. Our presented approach is basically a large neighborhood search that iteratively improves a given solution. In each iteration only a subset of processes is considered for reassignment and the newassignments are evaluated by a constraint program. In this paper we present our general solution approach. Furthermore, we evaluate different process selection strategies and other optimization means to improve the performance on larger instances. In addition, we present a simple way to compute tight lower bounds of the necessary costs.

3. A constraint programming-based approach to a large-scale energy management problem with varied constraintsInfoDetails

   Felix Brandt and Reinhard Bauer and Markus Völker and Andreas Cardeneo, 2013

   This paper addresses a large-scale power plant maintenance scheduling and production planning problem, which has been proposed by the ROADEF/EURO Challenge 2010. We develop two lower bounds for the problem: a greedy heuristic and a flow network for which a minimum cost flow problem has to be solved. Furthermore, we present a solution approach that combines a constraint programming formulation of the problem with several heuristics. The problem is decomposed into an outage scheduling and a production planning phase. The first phase is solved by a constraint program, which additionally ensures the feasibility of the remaining problem. In the second phase we utilize a greedy heuristic -- developed from our greedy lower bound -- to assign production levels and refueling amounts for a given outage schedule. All proposed strategies are shown to be competitive in an experimental evaluation.

Konferenzbeitrag (11)

1. Efficient Route Planning for Air CargoDetails

   Felix Brandt, 2018

2. Solving the Air Cargo Palletization Problem by Logic-based Benders DecompositionDetails

   Felix Brandt, 2017

3. The Air Cargo Load Planning ProblemDetails

   Felix Brandt, 2017

4. Generating real-world-like problem instancesDetails

   Felix Brandt, 2016

5. Training Planning for a Globally Dispersed WorkforceDetails

   Felix Brandt and Stefan Nickel and Brita Rohrbeck and Mirko Wichmann, 2015

6. Packing with Item Availability and Bin Due DatesInfoDetails

   Felix Brandt and Jan-Nikola Popratnjak and Enrico Schulz, 2014

   In this work, we present a new generalisation of the multiple heterogeneous knapsack problem (MHKP), by considering earliest availability dates of items and latest usage dates of bins. In the standard MHKP it is assumed that all the items to pack into bins are available and there are no restrictions about the time of bin assembly. In logistics however, especially in an air cargo environment this is not necessarily the case. Instead, the items for a flight arrive at the hub at different times, due to their inbound connection or delivery by the shipper. Furthermore, we assume there already exists a general production schedule for the bins of all flights. For a given flight to pack, we interpret this schedule as a sequence of possible due dates of its bins. In a solution each of these due dates can be assigned to at most one bin, i.e., the order of assembly of the bins is determined by the procedure. Accordingly, each bin can only be filled with items that are available before its due date. We propose an integrated problem formulation, from which we derive and solve a MIP model. Furthermore, we present our preliminary experiments to motivate the benefits of an integrated approach and compare its performance to simple decompositions into a sequence of standard MHKP.

7. A Multi-Temperature Truck Layout and Loading ProblemInfoDetails

   Felix Brandt and Anne Meyer, 2014

   In this work we consider a mixed vehicle layout and loading problem arising in the delivery of convenience products to stores by using multi-temperature vehicles. The vehicle's loading spaces have movable/retractable wall segments - both in longitudinal and lateral direction. The walls can be reconfigured for each tour into temperature zones, which are flexible in size and shape. For each tour, the order of stops and transport units per customer and temperature zone are given. We want to find a feasible wall setup and unit assignment, which results in a sequential unloading with a minimum number of unit reshuffles. In our talk, we present an integer programming model capturing the layout constraints as well as the reshuffling aspects of the problem. As the number of transport units per truck is naturally small and limited, the model can be solved to optimality within a reasonable amount of time. We give an overview of our experiments using real world data from our industrial partner PTV group. Our results show that there are significant time savings achievable - both during tour preparation and throughout the tour - compared to straight-forward manual loading strategies.

8. A Set-Covering Based Heuristic for Rich Vehicle Routing Problems with a Heterogneous FleetInfoDetails

   Felix Brandt and Nitin Ahuja and Werner Heid and Anne Meyer and Frank Radaschewski, 2013

   The Heterogeneous Vehicle Routing Problem (HVRP), as defined by Golden et. al, assumes a fleet that is composed of different types of vehicles. Usually, the vehicles are differing in their loading capacities as well as fixed or variable costs. The objective is to minimize the total cost of all vehicles used in the solution. One way to tackle the problem, presented by Taillard, is to solve a homogeneous fleet VRP for each vehicle type independently with an unlimited number of vehicles. Thereby, a lot of (partly overlapping) tours are generated. From these tours, the least cost subset visiting each customer once is found by solving a Set Partitioning Problem. In our work, we apply and extend the ideas of Taillard, to enable a commercial library for rich but homogeneous fleet VRPs to handle heterogeneous fleets. For the creation of tour plans, we rely on the VRP software package PTV xTour Server. Thereby, we are able to incorporate all of its supported types of restrictions from real world problems, like break and rest rules or multi-trips. We propose different strategies to create a number of preferably diverse solutions from the homogeneous fleet VRPs. To find the least cost subset of tours, we choose a Set Covering formulation. To make each customer visited exactly once, we introduce one additional step to repair the solution. The main idea of this step is to generate promising subtours out of the tours that contain customers, which are visited more than once in the current solution. These subtours are added to the Set Covering problem, which is solved repeatedly until no more duplicate visits exist. In this talk we present our solution approach, our insights of the experiment phase and from a real world application. Furthermore, we talk about some experience gained from implementing our approach into PTV xTour Server, a commercial vehicle routing solution.

9. Constraint-based large neighborhood search for machine reassignmentInfoDetails

   Felix Brandt and Markus Völker and Jochen Speck, 2012

   The presented approach is basically a large neighbourhood search that iteratively improves the given solution. In each iteration a subset of processes is released, while all other processes stay on their current machine. Released processes are then reassigned to the machines by a constraint program and the first improving solution is selected. We evaluate different neighbourhood selection strategies, branching techniques and optimization to improve performance on big instances. Furthermore, we present a network flow formulation that allows to compute lower bounds on the necessary costs.

10. Constraint-based construction heuristics for rich vehicle routing problemsInfoDetails

    Felix Brandt and Anne Meyer, 2012

    Constraint programming (CP) is successfully applied to rich vehicle routing problems. However, during the construction phase CP is used as a mere satisfiability checker, if at all, not using its power for finding an actual solution. Thus initial solutions might be poor or need additional repair procedures. In this work we want to show how constraint programming can be applied during a construction phase. We develop branching techniques analogous to established construction heuristics from literature. We present our first results and compare the performance of different approaches.

11. Merge-Savings - A Construction Heuristic for Rich Vehicle Routing ProblemsInfoDetails

    Felix Brandt and Andreas Cardeneo and Werner Heid and Alexander Kleff, 2010

    This contribution presents a new construction heuristic for general vehicle routing problems. The Merge-Savings algorithm is a generalization of the well-known Clarke & Wright Savings-algorithm. Moreover, our new approach implicitly includes the customer insertion operations found in the family of classical sequential and parallel insertion algorithms. We present the design goals of this algorithm, e.g. improved handling of time-windows and duration constraints compared to the Clarke & Wright algorithm, the optimization of planning problems with both depot-related as well as pickup and delivery transports in arbitrary proportions , and incremental planning, i.e. the extension of a given plan with unscheduled orders. The idea of Merge-Savings is, at each iteration, to merge two subtours into a new tour while maintaining each tour?s precedence relations. Algorithmically, this is realized by dynamic programming where each state represents the new tour constructed so far. The action set consists of choosing the next stop either from the first or the second tour. This algorithmic approach is fairly general and allows for the inclusion of constraints like multiple time windows, pickup and delivery transports, waiting time restrictions, and others as typically found in rich problems. We present computational results on real world instances realized with an implementation of Merge-Savings in the commercial vehicle routing package PTV Intertour and relate our approach to benchmark problems from the literature.

Thesis (2)

1. The Air Cargo Load Planning ProblemInfoDetails

   Felix Brandt, 2017

   A major operational planning problem in the air cargo industry is how to arrange cargo in an aircraft to fly safely and profitably. Therefore, a challenging planning puzzle has to be solved for each flight. Besides its complexity, the planning is mostly done manually today, which is a time consuming process with uncertain solution quality. The literature on loading problems in an air cargo context is scarce and the term is used ambiguously for different subproblems like selecting containers, packing items into containers, or loading containers into aircraft. All of the presented models only focus on certain aspects of what is in practice a larger planning problem. Additionally, some practical aspects have not been covered in the literature. In this work, we provide a comprehensive overview of the air cargo load planning problem as seen in the operational practice of our industrial partner. We formalize its requirements and the objectives of the respective stakeholders. Furthermore, we develop and evaluate suitable solution approaches. Therefore, we decompose the problem into four steps: aircraft configuration, build-up scheduling, air cargo palletization, and weight and balance. We solve these steps by employing mainly mixed-integer linear programming. Two subproblems are further decomposed by adding a rolling horizon planning approach and a Logic-based Benders Decomposition (LBBD). The actual three-dimensional packing problem is solved as a constraint program in the subproblem of the LBBD. We evaluated our approaches on instances containing 513 real and synthetic flights. The numerical results show that the developed approaches are suitable to automatically generate load plans for cargo flights. Compared to load plans from practice, we could achieve a 20 percent higher packing density and significantly reduce the handling effort in the air cargo terminal. The achieved costs of additional fuel burn due to aircraft imbalances and reloading operations at stop-over airports are almost negligible. The required runtimes range between 13 and 38 minutes per flight on standard hardware, which is acceptable for non-interactive planning. Cargo airlines can significantly profit from employing the developed approaches in their operational practice. More and especially the profitable last-minute cargo can be transported. Furthermore, the costs of load planning, handling effort, and aircraft operations can be significantly reduced.

2. Solving a Large-Scale Energy Management Problem With Varied ConstraintsInfoDetails

   Felix Brandt, 2010

   Diese Arbeit beschäftigt sich mit dem Problem der Wartungs- und Produktionsplanung eines großen konventionellen Kraftwerksparks unter Berücksichtigung vielfältiger Nebenbedingungen und Unsicherheit über einen Zeitraum von bis zu 5 Jahren. Die Aufgabenstellung entstammt dem ROADEF/EURO Challenge 2010, einem Wettbewerb der gemeinsam von der französischen Organisation für Operations Research und Decision Support (ROADEF) und der Europäischen Organisation für Operations Research (EURO) ausgeschrieben wurde.

Betreute Thesis (43)

1. Explorative Datenanalyse des ULD Break-Downs zur Entwicklung von Prognosemodellen zur ArbeitszeitDetails

   Moritz Schmid, 2019

2. Analysis of Handling Quality in Air FreightDetails

   Johannes Pesch, 2019

3. Forecasting non-compliant air freight shipmentsDetails

   Katja Habitzreither, 2019

4. Management and visualization of firewall rule sets using graph databasesDetails

   Jannik Lex, 2019

5. Weight and Balance Model for Aircraft Load PlanningDetails

   Anna Hess, 2019

6. Are air cargo market yields predictable?Details

   Celina Gödel, 2019

7. Forecast of Build-Up Relevant Cargo in the Air Cargo IndustryDetails

   Andreas Maier, 2018

8. Route-Choice Modelle in der LuftfrachtDetails

   Lukas Kienzler, 2018

9. Visual Analytics Tool für Operations Research VerfahrenDetails

   Frank Singh, 2018

10. Anwendung von Machine Learning zur Bestimmung des Frachtgewichts von Flugzeugen mit ADS-B Tracking-DatenDetails

    Timo Metzinger, 2018

11. Development of a routing algorithm for air cargo based on practice-oriented requirementsDetails

    Christoph Vierheilig, 2017

12. Entwicklung einer Heuristik zur Beladung von LuftfrachtcontainernDetails

    Dennis Wetzel, 2017

13. Implementierung eines Analysetools für Preisentwicklungen im LuftfrachtmarktDetails

    Hendrik Scherner, 2017

14. Analyse globaler Transportdauer von LuftfrachtDetails

    Oliver Vetter, 2017

15. The Internet-of-Things in the air cargo industryDetails

    Tobias Meyer, 2017

16. Marktmodellierung von LuftfrachtströmenDetails

    Konstantin Hendrich, 2017

17. Predicting Aircraft Fuel ConsumptionDetails

    Shengnan Xu, 2016

18. Analyse der Preiszuschläge für Spezialprodukte in der LuftfrachtDetails

    Sebastian Schäpertöns, 2016

19. Quantitative Analysis of Airfreight Charges on a Commodity BasisDetails

    Leonard Brauck, 2016

20. Reduktion von Container Loading Problemen auf relevante PackstückeDetails

    Felix Klenk, 2016

21. Validierung eines integrierten Modells zur Flugplanerstellung in der LuftfrachtDetails

    Jonas Elsen, 2016

22. Konzeption und Entwicklung einer Tool-Chain zur qualitativen und quantitativen Optimierung von Unit Load Devices im LuftfrachtverkehrDetails

    Marc Dannecker, 2016

23. Analyse der Netzwerkplanungsprozesse eines LuftfrachtunternehmensDetails

    Johannes Nehen, 2015

24. Entwicklung eines integrierten Modells zur Flugplanerstellung und Flugterminierung im kombinierten Fracht- und PassagiergeschäftDetails

    Kerstin Feurer, 2015

25. Managing Disruptions in the Airline IndustryDetails

    Lothar Weichert, 2015

26. Resource-Constrained-Scheduling zur Schulungsplanung am Beispiel der Trainingsplanung der Lufthansa CargoDetails

    Felix Götzinger, 2015

27. Frachtverfügbarkeit bei der Lufthansa Cargo AGDetails

    Marc Schmidt, 2015

28. Analyse der Verwendung von Ladehilfsmitteln am Beispiel der Lufthansa Cargo AGDetails

    Simon Burghartswieser, 2015

29. Stapelbarkeit in der 3D-LadeplanungDetails

    Jonas Hemlein, 2015

30. Modell zur Nachfrageprognose für Luftfracht unter Beachtung lokaler zeitlicher EinflüsseDetails

    Simon Schneckenberger, 2015

31. Algorithmische Grundlagen eines Routingdienstes für LuftfrachtDetails

    Julian Englert, 2015

32. Schulungsplanung als ZuordnungsproblemDetails

    Frauke Tabert, 2014

33. Models for Freighter Scheduling at Combination CarriersDetails

    Ulf Hansen, 2014

34. Modellierung und Analyse logistischer Prozesse im LuftfrachthandlingDetails

    Sebastian Scharf, 2014

35. Untersuchung von produktspezischen Preis-Mengen-Beziehungen in der Luftfracht am Beispiel der Lufthansa Cargo AGDetails

    Tonio Ihling, 2014

36. Visualization of load plansDetails

    Markus Behler, 2014

37. Vergabe von Langzeitverträgen für Niedrigpreis-Destinationen in der LuftfrachtDetails

    Jonas Elsen, 2014

38. Aufwandsminimale Umplanungen im laufenden Schulungsbetrieb bei der Lufthansa Cargo AGDetails

    Marc Janschekowitz, 2014

39. Konzeptionelle Bestimmung von Schulungsstandorten bei der Lufthansa Cargo AGDetails

    Mirko Wichmann, 2013

40. Analyse des Systems zur Flugplanevaluierung der Lufthansa Cargo AGDetails

    Maximilian Möller, 2013

41. Tourenplanung mit stochastischer korrelierter NachfrageInfoDetails

    Jacqueline Wirnitzer, 2012

    In practice, many parameters of truck scheduling problems are inherently uncertain. Thus, many extensions of the basic Vehicle Routing Problem (VRP) have been established. One of them is the VRP with Stochastic Demands (VRPSD). Almost all published models solving this optimization problem assume independency between demand distributions of customers, although this assumption does not apply for many cases in reality. In this work, a generalisation of the VRPSD is introduced, in which correlation between customers is taken into account. We call this problem VRP with Correlated Stochastic Demand (VRPCSD). The aim of this thesis is to develop a two-stage heuristic solution to the problem. A regular tour schedule is created on the basis of stochastic data before actual demands are announced. Afterwards, when demands are known, the tour schedule is checked for infeasibilities and adjusted by recourse strategies if necessary. The overall objective is to minimize total costs. We refrain from reoptimisation (replanning when demands are announced) to save costs by regularities in tours (see Gröer, Golden und Wasil (2009)). In this work various methods and recourse strategies for the VRPCSD are developed. An improved Savings heuristic is the basis of this process. To create a regular tour schedule, the only approach to the VRPCSD found in literature is implemented and refined (Yee and Golden (1979)). Both the original and the adapted approach turn out to be inefficient, therefore, new methods are developed. Likewise, new recourse strategies are developed besides the established Back-To-Depot strategy. Finally, the various methods are combined with the recourse strategies and experiments are conducted. As a benchmark we use the average reoptimisation results. The best combinations of methods and recourse strategies achieve results that differ on average less than 20% from the benchmark.

42. Algorithmische Packplanung für LuftfrachtInfoDetails

    Mathias Leu, 2012

    In dieser Arbeit wird ein Verfahren entwickelt, dass die praktische Anwendbarkeit von automatisierter Frachtplanung in der Luftfracht beweisen soll. Das entwickelte Verfahren basiert auf einem Packalgorithmus, der aus Frachtstücken, die in einer Reihenfolge übergeben werden, eine Packanordnung erstellt. Für die Ermittlung einer guten Reihenfolge wird eine Nachbarschaftssuche über verschiedene Nachbarschaften durchgeführt. Das Verfahren beinhaltet verschiedene praxisrelevante Restriktionen wie beispielsweise Stapelbarkeit, Stabilität, Priorisierung und Trennungs- bzw. Zusammenladbarkeitsvorschriften. Anhand eines praxisnahen Referenzdatensatzes sowie verschiedener Realdaten wird das Verfahren ausführlich getestet und bewertet. Abschließend wird ein Ausblick über mögliche weitere Entwicklungmöglichkeiten gegeben.

43. Entwicklung einer schnellen Heuristik für das Tourenplanungsproblem mit AlternativenDetails

    Martin Ziegler, 2011

Whitepaper (1)

1. Strategic Training Planning for a Globally Dispersed WorkforceInfoDetails

   Felix Brandt and Stefan Nickel and Brita Rohrbeck and Mirko Wichmann, 2017

   In this paper we introduce the Strategic Training Planning (STP) problem, which can be found in most large organisations operating at multiple locations. We consider a set of employees residing at different locations who have to attend on-site trainings. The problem is to determine an optimal set of training locations and to assign each employee to one of the training locations. Further practically relevant restrictions like capacities, languages spoken as well as cultural and legal issues have to be considered. The objective is to find a cost-optimal plan with respect to set-up, travel and opportunity costs. The STP turns out to be a variant of the warehouse location problem. We present the full problem description and an integer programming formulation of the problem. Furthermore, we demonstrate the relevance of the problem and the impact of STP solutions by a real-world case study. When compared to the found optimal training setup, straightforward solutions that are chosen by manual planners result in 60 to 80 percent higher cost. Accordingly, our results show that there are significant savings achievable by employing the STP, both in planning effort and cost reduction.

Export Suchergebnis .bib

Kontakt

Telefon: +49 721 9654-304
Fax: +49 721 9654-305
E-Mail: brandt@dont-want-spam.fzi.de

Zurück