Evolutionary Approach to Function Model Synthesis: Development of Parameterization and Synthesis Rules

Abstract

The goal of this paper is to develop the groundwork for automated synthesis of function models. To this end, an evolutionary algorithm based framework has been developed. A parameterization method that can completely describe any given function models has been proposed. The parameterization makes the function models compatible for use within the evolutionary algorithm framework. Validation of the parameterization method is carried out by using an evolutionary algorithm to synthesize the function models for five different electromechanical products. The algorithm converged in each case, indicating that the method is satisfactory and that function models can actually be synthesized using an evolutionary framework. In addition, the adaptation of several a priori rules for use in this framework has been proposed. These rules are categorized as grammar, logical and feature based rules. An updated evolutionary framework that incorporates these rules is also presented.