Copyright © IFAC Automatic Systems for Building the Infrastructure in Developing Countries, Istanbul, Republic of Turkey, 2003 ELSEVIER IFAC PUBLICATIONS www.elsevier.com/locale/ifac THE MODELLING AND SIMULAnON OF ELEVATOR GROUP CONTROL SYTEMS FOR PUBLIC SERVICE BUILDINGS C.Erdem Imrak l , Mustafa OzkJnm 2 /Istanbul Technical University, Department of Mechanical Engineering GumlLSsuyu, 34439 Istanbul, Republic of Turkey Fax:2I2 2450795... E-mail: imrak@itu.edu.tr 2 Istanbul Technical University, Department of Mechanical Engineering GumlLSsuyu, 34439 Istanbul, Republic of Turkey Fax:212 24507 95... E-mail: ozkirimmu@itu.edu.tr Abstract: Elevator traffic control systems have become more and more complicated due to their nature of intelligence. Artificial intelligence methods employing neural networks have been proved to be successful in many fields, such as process modeling, pattern recognition and classification problems. They have also been applied to basic problems in elevator traffic control systems, such as the prediction and control of elevator movements. In particular, neural networks can offer better solutions to the passenger call allocation process when compared to the classical traffic control methods. Elevator control algorithms utilizing neural networks aims at distributing the most suitable cars to the floors by considering the passenger service demand. Neural networks can dynamically learn the behavior of an elevator system and predict the next floors to stop, based on what has been learnt. In this paper the neural network approach has been applied to DuplexfTriplex group control systems for improving passenger waiting time and a lift simulation software has been developed and implemented in order to assess the learning capability by measuring the performance of the control algorithm. The lift traffic analysis have been carried out by examining the simulation results obtained. Copyright © 2003 IFAC Keywords: control system design, neural network, simulation, backpropagation, learning algorithms. I. INTRODUCTION Computer based traffic control systems can assign cars more effectively than the classical traffic control systems, there exists, however, a limit to what can be achieved. The main limit is the finite capacity of the underlying equipment to handle the traffic demands. Hall and car calls are often allocated to suitable cars by taking into account of the minimum cost concept that operates by allocating the call to the car with the lowest cost amongst all the cars that are available to serve. The criteria for determining a suitable cost function depends on either quantity of service and/or quality of service. The quality of service is a measure of the lift capacity consumed to serve a specific set 145 of calls, indicated by total journey times of all the cars (Barney, 2003). An excellent service capability of a lift system in any modem public service building (PSB) is needed for the efficient functioning of the building. In the design of lift traffic and calculating the performance of a lift system, the traditional method is to calculate the round trip time (RTT), which relies on calculating the average number of stops made (S), the average highest reversal floor (H) and the average number passengers carried (P) (Jenkins, 1992). Then the number, car capacity and the speed of the elevators can be derived in order to provide a reasonable interval (RTT / number of lifts) and five minute handling capacity (300 x P / interval). The