An evolutionary algorithm with a fuzzy fitness evaluation module for the configuration of personal computers

In the artificial intelligence area, the knowledge-based configuration of personal computers might serve as a reference example for configuration-type expert systems. Nevertheless, we use, as an alternative approach, evolutionary algorithms. This is because of at least three reasons to make experiments with such types of algorithms in the design and configuration domain: real world configuration problems (in particular in the engineering field) become very big because of the component structure and because of the large variety of alternatives for each of the components; even if it would be possible to model the multi-criteria optimization problem, which is related to the configuration task, appropriately, it would be almost impossible to find globally optimal or good configurations by the sequential-type of configuration algorithms, which are typically found in the knowledge-based approaches; and user requirements are imprecise rather than crisp. To deal with this imprecision, we are required to generate a set of alternatives which should be ranked by a fuzzy-MADM method afterwards