The application of neural networks to lift traffic control

The applicability of neural network to improve the traffic performance in lifts is analyzed. The use of neural network to learn the likely destinations from one floor to another and allocate cars to call using the information obtained is proposed. The system is described to be of little use during up-peak and down-peak traffic patterns but useful during interfloor traffic encompassing a random pattern of movements. The comparison of the results to other algorithms established a better performance of the neural network system than the fixed sectoring systems but a poorer performance against the computer-based traffic control systems.