No label defined (Q4936002)

The Ph.D. thesis illustrates the approach to 4 concrete practical case studies using combinatorial optimization techniques. Besides the interesting topics of the 4 cases, it is also a very self-critical thesis showing also the negative aspects of modelling and providing solutions -- all in all it underlines the importance of evaluating the model in each solution step. In the introduction the solution techniques are explained and the four case studies are introduced: Case 1 is a production scheduling problem to minimize maximum lateness which can be modelled by a generalized job shop scheduling problem. The solution technique used is tabu search. Case 2 is a frequency assignment problem. This case is used to compare different solution approaches. Case 3 is a problem of statistical disclosure control -- the aim is to measure statistical data in a way to prevent identification of an individuum from these data. Despite the efforts the author made to tackle this practical problem, he is not satisfied with his approach and he explains his main criticism. Case 4 is a problem of industrial cutting of certain rectangles out from huge rectangular blocks. The heuristic implemented for this case proved to be superior to former manually generated solution approaches. A very praxis-oriented report is presented for all 4 cases, including all the experiences during the real-life projects. The essential conclusions of the author are summarized in 4 solution phases: Modelling -- solving -- implementation -- maintenance.