LATERITE-CEMENT COMPONENT MIXTURE SELECTION USING GENETIC ALGORITHM

Abstract

Meeting specification requirements at first attempt in mixture proportioning are difficult to achieve without employing a carefully designed process. Such a method is regarded to, as an experimental design process. A Genetic Algorithm (GA) seamless process is carried out here for a laterite-cement brick component selection process for use, particularly in mass housing project. It involved firstly, by developing a prediction equation described here as the objective/cost function achievable within a constrained design domain for component mixture quantities using the Scheffe mixture design process. Secondly, the relationship of component mixtures which yields responses such as cost were modelled as equality or inequality constraints. The choice of the GA procedure employed was based on the feasibility of the solution and convergence rate to obtain solutions meeting the specification requirement desired. This method enabled obtaining combinations of mixture quantities in an effortless way to meet the desired user requirement.