RAW MATERIAL OPTIMIZATION WITH NEURAL NETWORK METHOD IN CONCRETE PRODUCTION ON PRECAST INDUSTRY

Abstract

The development of construction is presently experiencing rapid growth in Indonesia, leading to the requirement of the right materials for infrastructural enhancements. From the existing infrastructure, concrete innovations such as precasts are needed with good quality materials, for the quick completion of construction. This is because the need for good quality and smooth material helps to determine the success of a building project, with the use of technology through precast being a problem-solving process. Therefore, this study aims to analyze the patterns by which inventory procurement predictions produce precasts with good quality, using the e-readiness framework concept of the neural network through appropriate decision-making processes. It also focuses on innovating technological products used in the Indonesian precast industry. The Methodology Neural Network was used to produce the best target quality time and precast commodities. The result indicated two outputs from 2 neural network models, using five similar input-value variables. Based on the Adaline neural network, the outputs were observed as the highest sales-cost predictions for precast products, which often occurred in 1, 5, 6 and 9 months. Besides this, production activities were also normally operated at level (1), with profit optimization being highly considered before months 1, 5, 6 and 9. For the LVQ neural network, the result was a predictive classification of class intensity levels, where fast decision-making processes occurred in months 1, 6 and 9. Cost optimization was also carried out by ordering raw materials several months in advance, considering the trend in material prices and logistics.