Artificial Neural Network Model for Affective Environmental Control System in Food SMEs

Abstract

This paper presents an affective environmental control system for Small and Medium-sized Enterprises (SMEs). The system is proposed as a technology innovation in appropriate information technology. It is defined that workplace environment set points could be controlled using worker workload. The research objectives are: 1) To design an affective environmental control model for SME; 2) To develop an Artificial Neural Network (ANN) model for predicting affective environment set points. The system consisted of 4 sub-systems as measurement, assessment, control and decision. An ANN model is developed for sub-systems of control. Training and validation data are acquired from 4 (four) samples of SME in Yogyakarta Special Region, Indonesia. The model has been developed successfully to predict temperature and light intensity set points using back-propagation supervised learning method. The research results indicated the satisfied performance of ANN with minimum error. ANN model indicated the closeness of R2 value between training and validation data. The research results could be applied to support the worker productivity in food SMEs by providing a comfort workplace environment and optimum worker workload.