Real time hybrid installation line balancing and manufacturing application

Keywords – Line balancing, neural network, simulation. Variations, in the manufacturing system, are hard tackle. Breakdowns of machines, variable labour time, poor quality production, and labour mistakes can cause the variation of manufacturing time. The variation of time may change the cycle time, and increase the loss of balance in manufacturing line. To avoid the problem, a dynamic line-balancing method should be applied. The time the line-balancing method intervenes the manufacturing line is crucial for prevention against loss of balance. If the re-balancing process is preformed at the time when a problem occurs, only slack-time of next stations can be used. However, when the problem is predicted, the total slack-time of the manufacturing line can be taken advantage of. If the method can achieve to complete manufacturing in slack-time, which makes the method real-time, there will be no loss. In the study, a real-time assembly line balancing method is developed. An artificial intelligence model and statistical estimations are integrated to the model for prediction of the problem occurrence and duration.