Multiobjective Optimization in Elevator Group Control

Modern elevator systems in high-rise buildings consist of groups of elevators with centralized control. The group control allocates hall calls to the most suitable elevators by optimizing a cost function. This problem can be viewed as a combination of online scheduling, resource allocation, and stochastic control. The usual performance criterion to be optimized when scheduling passenger pick-ups is the average waiting time of all passengers in the system, i.e., the time period from the moment when the passenger arrives until the moment when this passenger boards some elevator. Alternative criteria are sometimes used as well, such as the ride time, defined from the time of boarding until a passenger arrives at the desired floor, the percentage of waiting times exceeding a certain limit or the number of elevator stops, which is related to energy consumption. Considering two or more of these criteria simultaneously results into a multiobjective optimization problem, where the group control algorithm should be tuned in order to find a good compromise between the different criteria. We study the relations between the criteria and some tuning parameters of to our Estimated Time of Arrival (ETA) algorithm by simulating different traffic patterns. Then we define a linear utility function form combining the criteria. Based on reasonable weights for the criteria we determine find optimal values for the tuning parameters. H. Hakonen, A. Rong and R. Lahdelma 2