Energy-saving elevator dispatching has been recognized as a challenging issue in building transportation, and we develop a novel energy-saving dispatching strategy for regenerative group-elevator system. Group-elevator dispatching is a typical combinatorial optimization problem, and three keys of the dispatching optimization are optimization method, objective, and model. The three keys of energy-saving-oriented elevator dispatching are studied in this paper. First, robust optimization method is introduced to handle dispatching optimization under uncertain elevator traffic flows; uncertain flows influence energy-saving dispatching seriously. Second, dispatching energy-objective function for regenerative group-elevator system is derived both schedule energy for four traffic patterns (up-peak, down-peak, up/down-mixed, and night) and return energy for two peak patterns (up-peak and down-peak) are considered. Third, four robust optimization-dispatching models for four traffic patterns are built, and optimization objectives of four models are minimizing the energy-objective function. Moreover, because we cannot solve robust optimization models with uncertain parameters directly, model counterpart transformation is studied. Finally, we solve the four transformed models by Linear Interactive and General Optimizer software and obtain robust optimization-dispatching solutions. In practice, four energy-saving-dispatching robust optimization models are switched according to real-time traffic patterns, and elevators are dispatched based on the dispatching solutions. We reduce elevator system-energy consumption effectively and keep average waiting time of the passengers acceptable under multi-traffic patterns. Simulation results demonstrate the validity of our strategy. Practical application: Group-elevator system spends much unnecessary energy because of the uncertainty of elevator passenger-traffic flows. This paper develops an energy-saving elevator-dispatching optimization strategy, which is immune to the uncertainty of four typical traffic flows. In practice, we update the group-elevator controller by our algorithm to realize energy-saving dispatching of regenerative group-elevator system under multi-traffic patterns.
