Modelling a Multi-car Elevator System using Witness

Elevators are considered a necessity in buildings with multiple floors. In tall buildings there is a need for efficient control of the elevator system in order to obtain satisfactory service levels. To achieve an efficient elevator control strategy a variety of diverse and sometimes conflicting constraints has to be resolved. To investigate the effectiveness of a vertical transportation system, we have developed a simulation model to analyze the call strategy of a multi-car elevator system. In this paper the main steps in the methodology for modelling elevator performance and passenger traffic in a public building are explained. Elevator dynamics and control, call allocation and passenger traffic modelling are described resulting in a Witness model. Witness simulation software has been used as a test-bed for model building, simulation and experimentation. Three different strategies were devised and experimented on the model and the results were compared to find the best average waiting time. Results indicate that the average waiting time of passengers was slightly improved when using the shortest distance strategy. In addition to passenger waiting time it has been realised that elevator performance depends on the passenger traffic patterns and, during heavy traffic, service level is greatly affected by the call allocation algorithm. The research concludes with the need to carefully select the data set as the basis for simulation comparisons.