Non linear and multidimensional models of an interdisciplinary nature are needed to study a broad spectrum of tasks related to design of mechanical engineering systems. Oversimplified and therefore inapplicable results are frequently obtained as the final result of model development. A human way of making decisions is not based on numbers but on common-sense reasoning. Qualitative modeling brings additional options in the modeling of complex systems. Qualitative variables are quantified using three values only - positive (increasing), zero (constant) and negative (decreasing). The classical quantitative tools cannot deal with such variables. Typical areas suitable for qualitative modeling are design related tasks with a large number of variables. A presented case study an optimization of a solar collector, is given in details. and Obsahuje seznam literatury