Fuzzy fitness functions applied to engineering design problems

Engineering design is concerned with the creation of devices and systems, and encompasses many activities, including the creative activity of conceiving or synthesizing new devices and systems, as well as the analysis, refinement and testing of those concepts. If all information available to an engineer were precise and deterministic, design would be straightforward. However, uncertainty of many forms is present throughout the design process. The uncertainty theories that are most relevant for engineering design are those that deal with a mixture of numerical, set or interval-valued and linguistic information. Populations of design alternatives can be generated using evolutionary methods [Antonsson and Cagan, Formal Engineering Design Synthesis, Cambridge University Press, Cambridge, 2001]. It has been shown that incorporating uncertainty, such as variations in the operating environment, simulated by variations in the evaluation of the fitness function, can synthesize designs that are robust to the variations. In this paper a class of fuzzy fitness functions ( F 3) and hybrid uncertainty fitness functions are introduced, which combine fuzzy and probabilistic types of uncertainties.