IMPLEMENTATION OF DUAL RESPONSE APPROACH TO CONCRETE MIX DESIGN BASED ON HYBRID NEURAL NETWORK-GENETIC ALGORITHMS (PRELIMINARY STUDY)

Abstract

Process parameters prediction in robust design is very important. If the predictions results are fairly accurate thenthe quality improvementprocess will save time and reduce cost. The concept of dual response approach based on response surface methodology has widely investigated. Separately estimating mean and variance responses, dual response approach may take advantages of optimization modeling for finding optimum setting of input factors.A sufficient number of experimentations are required to improve the precision of estimations. Research on optimization of concrete mix design to achieve adequate strength is of considerable importance. In a concrete mix design, as the number of mix variables increases the number of experiments also increases. This study proposed an alternative dual response approach without performing experiments. Neural Network – Genetic Algorithmscan be appliedto model relationships between responses and input factors. Using empirical process data, process parameter can be predicted without performing real experimentations. This is a wise solution to minimize the environmental impacts due to using of raw material wasted.An optimization of brick with concrete mix design has been investigated to demonstrate the procedures and applicability of the proposed approach.
Keywords: dual response approach, algorithms