Complexity reduction for automated design problems of building automation systems

Modern large buildings like office buildings include a great number of building automation systems like heating, cooling or lighting. To compose them, the design process should be automated, because no system designer have an overall knowledge about all the different available systems on the market. An important problem is the complexity of the optimization of the device selection: many different systems should be integrated, the costs and the number of integrated devices should be minimized, etc. Well-known population-based algorithms are suited to such complex multi-criterial optimization processes. But the most important drawback of these algorithms is their disadvantageous time performance by a high number of available devices or by many criteria. This paper gives a solution to handle this problem by restricting the device domain to provide a good scalability of the device selection.