Entwicklung eines Instrumentes zur Loesung grosser energiesystemanalytischer Optimierungsprobleme durch Dekomposition und verteilte Berechnung

Having only a small number of alternatives, decisions can primarily be made on a basis of experience. However, if the problems become more and more complex it is in most cases preferable to use a formal approach like methods from operations research. Systems analysis requires an integrated assessment and should model reality as 'real as possible'. Due to the huge amount of information - and hence of input data - model applications which either cannot be solved by the currently available algorithms or which require a huge amount of computer time are created easily. These applications are called large-scale systems. In order to combine the mathematical formulation of the decision problem with the necessary algorithms algebraic modeling languages have been developed. It is the objective of this thesis to develop an instrument to decompose general linear optimization problems for either a sequential or concurrent calculation (to achieve solvability at all) and to inquire if a speed-up in the process of solving can be achieved with a distributed calculation. The integration into an algebraic modeling language allows for the application of the instrument to general linear optimization problems. Using this instrument for a distributed calculation it is in principle possible to use it either in the environment of a local area network or on other computers connected by the Internet. However, our experiments revealed that distributed calculations should be avoided of linear problems when possible. The method of decomposition that is described in this paper can also be used for a further development to link different models and for decomposing non-linear problems. The use and application of the developed instrument is demonstrated in an example in the area of energy systems analysis. In this example the energy system of the Federal Republic of Germany is analyzed until the year 2050 by calculating different scenarios. (orig.)SIGLEAvailable from TIB Hannover: RO 9712(84) / FIZ - Fachinformationszzentrum Karlsruhe / TIB - Technische InformationsbibliothekDEGerman