Wolfgang Domschke

The Job Shop Scheduling Problem: Conventional and new Solution Techniques

A job shop consists of a set of different machines that perform operations on jobs. Each job has a specified processing order through the machines, i.e. a job is composed of an ordered list of operations each of which is determined by the machine required and the processing time on it. There are no precedence constraints among operations of different jobs. Operations can not be interrupted (non-preemption) and each machine can handle only one job at a time. Consequently, each job can be performed on only one machine at a time. The operation sequences on the machines are unknown and have to be determined in order to minimize the makespan, i. e. the time required to complete all jobs.

Pattern Based Vocabulary Building for Effectively SequencingMixed-Model Assembly Lines

In this paper, the problem of sequencing mixed-model assembly lines in case of fixed rate launching and closed stations is considered. The problem consists of finding an intermixed sequence of different models of a basic product, which are jointly produced on an assembly line, such that customer demands are fulfilled and total work overload is minimized. For solving this problem an informed tabu search procedure with a pattern based vocabulary building strategy is developed. Computational tests demonstrate that considerable improvements are obtained by comparison to methods which do not incorporate such an approach.

Multiple center capacitated arc routing problems: A tabu search algorithm using capacitated trees

Abstract In the capacitated arc routing problem with multiple centers (M-CARP) the objective is to find routes starting from the given depots or centers such that each required arc is served, capacity (and usually additional) constraints are satisfied and total travel cost is minimized. In this paper we consider a heuristic transformation of the M-CARP into a multiple center capacitated minimum spanning tree problem with arc constraints, that we call arc-constrained CMST. An algorithm for determining initial feasible solutions as well as an improvement procedure for this problem are described and the “re-translation” of CMST solutions into the CARP context is explained. It is shown that the objective function value of the obtained CARP solution is easily derived from the respective value of the corresponding heuristic CMST solution. Furthermore, the possibility of including side constraints and the consideration of additional objective functions is discussed. Computations on real-world benchmark problems compare the results of the tabu search and simulated annealing metastrategies embedded in the improvement procedure.

Schedule synchronization for public transit networks

SummaryThe problem of minimizing waiting times of passengers who want to change communication routes at transfer stations is considered. The paper contains heuristics (regret-methods, improvement algorithms, simulated annealing) as well as a branch and boundalgorithm and provides computational results, some of them gained with practical data.ZusammenfassungBetrachtet wird das Problem der Fahrplangestaltung im öffentlichen (Personen-)Nahverkehr mit dem Ziel der Minimierung der Summe der Wartezeiten für Umsteiger. Es werden heuristische Lösungsverfahren (Regret-Methoden, Verbesserungsverfahren, Simulated Annealing) sowie ein Branch-and- Bound-Algorithmus beschrieben. Die Verfahren wurden u. a. anhand von Daten aus der Praxis getestet.

ADD – heuristics' starting procedures for capacitated plant location models

Abstract This paper is concerned with heuristics for capacitated plant location models where locations have different capacities. In this case ADD-heuristics normally lead to bad solutions. We present some starting procedures (priority rules) in order to overcome this difficulty. Finally, we report numerical results, including comparisons between ADD-heuristics with starting procedures and DROP-heuristics.

Tabu search techniques for the quadratic semiassignment problem

Assigning items to sets such that a quadratic function is minimized may be referred to as the quadratic semi-assignment problem. This problem indeed is a relaxed version of the well known quadratic assignment problem. Applications may be found in floor layout planning, in certain median problems with mutual communication, and in the problem of schedule synchronization in public transit networks. The same problem also arises in certain scheduling problems where the deviation from due dates is penalized by a quadratic function.

Zwei Verfahren zur Suche negativer Zyklen in bewerteten Digraphen

ZusammenfassungDas erste der angegebenen Verfahren basiert auf dem Dijkstra-Algorithmus [3], das zweite auf dem Algorithmus von Ford [9] (bzw. Moore [12]). Beiden wird eine mehrmals auszuführende “Prozedur” hinzugefügt. Dadurch wird für den Dijkstra-Algorithmus erreicht, daß er aufbeliebig reell bewertete Digraphen anwendbar ist und daß vorhandene negative Zyklen entdeckt werden können; mit Hilfe des erweiterten Ford-Algorithmus können negative Zyklen mit geringerem Rechenaufwand als mit dem ursprünglichen Verfahren erkannt werden.AbstractThe first of the two methods developed is based on the Dijkstra-algorithm [3] the second on the algorithm of Ford [9] (respectively Moore [12]). To both of them a “procedure” is added which must be carried out several times. By this means one attains that the Dijkstra-algorithm can be applied toany real valued digraph and that existing negative cycles can be detected, by the aid of the extended Ford-algorithm negative cycles can be detected with less computing time than with the original method.

Modelle und Verfahren zur Bestimmung betrieblicher und innerbetrieblicher Standorte – Ein Überblick

ZusammenfassungEs wird ein Überblick über die in der Literatur behandelten Modelle zur Bestimmung von betrieblichen und innerbetrieblichen Standorten gegeben. Für die meisten Modelle werden exakte und/oder heuristische Lösungsverfahren beschrieben.SummaryThis paper presents a survey of plant-location and plant-layout models discussed in the literature. For most of the models exact and/or heuristic solution methods are described.

Kapazitätsplanung in Netzwerken

ZusammenfassungDie Netzplantechnik zählt zu den für die Praxis wichtigsten Gebieten des Operations Research. Insbesondere die Modelle und Methoden der Struktur- und Zeitplanung werden in vielen Fällen mit Erfolg angewendet. Vorgehensweisen der Kosten- und vor allem der Kapazitätsplanung haben jedoch bislang kaum größere praktische Verbreitung gefunden — trotz oder gerade wegen einer Fülle neuerer theoretischer Konzepte. Der vorliegende Beitrag gibt einen Überblick über deterministische Modelle und Methoden zur Kapazitätsplanung in Netzwerken, welche zur Projektplanung entwickelt wurden. Im Mittelpunkt der Ausführungen stehen Modelle mit Zeit-/Kosten-/Ressourcen-Tradeoffs bei beschränkter Perioden- und Gesamtkapazität und (nicht-) unterbrechbaren Vorgängen.SummaryProject scheduling is one of the most important disciplines of Operations Research with respect to applications. Especially the simple methods associated with CPM or MPM are successfully applied rather oftenly in practice. However, cost and especially resource-oriented project scheduling methods are in practical use only rarely — despite a lot of new and promising concepts. The paper surveys recent advances in modelling and solving deterministic resource-constrained project scheduling problems suitable for project planning. The main topics are nonpreemptive time/cost/resource-tradeoff models with renewable, nonrenewable and doubly-constrained resources as well as preemptive models which may be treated by linear programming.

Two new algorithms for minimal cost flow problems

Two new algorithms for minimal cost flow problems in lower and upper bound capacitated flow networks and an ALGOL-procedure for one of them are given.On the average both methods need less than half of the computing time of the Out-of-Kilter-algorithm.ZusammenfassungZwei Algorithmen zur Bestimmung kostenminimaler Flüsse in Netzwerken mit unteren und oberen Kapazitätsbeschränkungen sowie ein ALGOL-Programm zu einem der Verfahren werden angegeben.Beide Verfahren erfordern durchschnittlich weniger als die Hälfte der vom Out-of-Kilter-Algorithmus benötigten Rechenzeit.