Institution: Graduate School der Fakultät Wirtschafts- und Sozialwissenschaften der Universität Hamburg
Dozent/in: Stefan Voß
3-Tage-Block ca. Ende Nov./Anfang Dez. 2012
Plus Metaheuristics-Tagung am 28.02.-01.03.2013
Zielgruppe: Whoever is interested. PhD or advanced master students in business administration, economics, information systems, mathematics, civil engineering and alike.
Metaheuristics support managers in decision making with robust tools providing high quality solutions to important problems in business, engineering, economics and science in reasonable time horizons. While finding exact solutions in these applications still poses a real challenge despite the impact of recent advances in computer technology and the great interactions between computer science, management science / operations research and mathematics, (meta-) heuristics still seem to be the methods of choice in many (not to say most) applications. In this course we give some insight into the state of the art of metaheuristics. This focuses on the significant progress regarding the methods themselves as well as the advances regarding their interplay and hybridization with exact methods.
Basically, a metaheuristic is a top-level strategy that guides an underlying heuristic solving a given problem. That is, a metaheuristic is an iterative master process that guides and modifies the operations of subordinate heuristics to efficiently produce high-quality solutions. It may manipulate iteratively a complete (or incomplete) single solution or a collection of solutions. The subordinate heuristics are, e.g. high- (or low-) level procedures, simple local search, or just a construction method. Metaheuristics may use learning strategies to structure information in order to find optimal or near-optimal solutions efficiently.
Assuming a given problem, the goal is to find an optimal or at least a high-quality solution.
Each problem is associated with a solution space containing all feasible solutions according to the restrictions of the problem. One way of finding an optimal solution could be the search through all solutions and selecting the best one. Unfortunately, the size of the solution space is too large to accomplish all comparisons within a realistic time span and, therefore, heuristic methods have to be applied that limit the search on interesting areas of the solution space.
Well-known examples of metaheuristics are, e.g., simulated annealing, genetic algorithms, ant systems and tabu search.
In this course we learn about the basics of metaheuristics (What are metaheuristics? Why do we need them? What are the basic ingredients? How could we implement them? How could they be applied in different settings? etc.). Moreover, we are going to discuss the use of metaheuristics in different problem domains, let it be in computational logistics, telecommunications, production planning, marketing plus other domains the students might be interested in. Interest given, we could extend into hybridization with other types of methods like simulation, mathematical programming and alike.