Assembly Line Balancing using Artificial Neural Network: A Case Study of Tricycle Assembly Line

Abstract

This study reports the use of Artificial Neural Network in balancing an existing single-model assembly line of Boulous Enterprises Limited. A multilayer perceptron, with the help of online training was utilized, due to its ability to accommodate large dataset. The results obtained showed that standard cycle time of 576 seconds in the existing line was reduced to 526 seconds. Also, the average idle time was reduced from 105 seconds to 56 seconds, and the output of tricycles produced per day was increased from 50 to 55. The results clearly showed that a better balanced line was obtained with the use of Artificial Neural Network. Keywords: Line Balancing, bottlenecks, Idle Time, Efficiency