A Comparison of Metaheuristics for the Allocation of Elevators to Calls in Buildings

Abstract

This paper deals with the car-call allocation problem in verticaltransportation in buildings. We have made a wide comparison of differentmetaheuristic optimization algorithms to identify those with a betterperformance dealing with the problem. The tested approaches are DifferentialEvolution (DE), Simulated Annealing with Random Starts (SAR), Artificial BeeColony (ABC), Bat Algorithm (BA), Bacterial Foraging Optimization Algorithm(BF), Particle Swarm Optimization (PSO), Genetic Algorithm (GA), and TabuSearch (TS). Each algorithm was tested in high-rise building simulations of 10to 24 floors, with car configurations of 2 to 6 cars. Results proved that theABC and TS algorithms generally result in better average journey times comparedto other methods. It has to be noted that we introduced a new version of theSimulated Annealing, Simulated Annealing with Restarts (SAR), which ranked asthe third best algorithm.