Evolutionary Artificial Neural Networks for Concrete Mix Design Modelling

Concrete is a composite of materials that each material has complex characteristic, therefore it generate a variety of complicated models.Concrete designer experts are not convinced by the mathematical formula of concrete mix design.According to them, one mix design can only be applied to one mixture.Therefore in every production which involves variations in material characteristics that are slightly different, new mixture design is required.Usually it is done manually with a variety of mixed composition and then test it by using destructive test which has some drawbacks, such as expensive material cost, unpredictable results, and environmentally unfriendly.Moreover concrete mix design nowadays are limited to a few component such as cement, sand, gravel, and water.Therefore it needs to be developed in such way to accommodate recent concrete development.Computational model which is developed in this study uses several parameters such as cement, sand, gravel, water, coal ash (fly ash), and blast furnace slag with Evolutionary Artificial Neural Network approach that based on Genetic Algorithm.Our results proved that the proposed model has lower error rate than Artificial Neural Network model thus help reduce drawbacks that are exist by designing manually.