Research On Fuzzy Neural Network Based Recognition Of Traffic Patterns Of Elevator Group Control System And Multi-target Dispatching Algorithm

The elevator is an indispensable vertical traffic means of transportation in the skysraper, with the emerging in a large amount of the high-story intelligent building, passengers have put forward higher and higher demand to the systematic function of the elevator. Elevator Group Control System(EGCS) is a control system that manages three or more elevators in a building in order to optimize elevator assignment to transport passengers efficiently and improve quality of service.The thesis regards the EGCS as the research object. On the basis of analyzing in depth t o the control technology of the elevator, deep research has been carried on to the group control intelligence system of the elevators regarding improving the systematic intelligent level of EGCS, reducing waiting time, improving the service quality and meeting optimize the needs of modem work as the goal. A kind of new multi-objective intelligent dispatching method has been putting forward by setting up an EGCS controlled on the basis of the traffic pattern-recognition.The thesis analyses the systematic characteristic, control method, performance evaluation index and traffic flow situation of EGCS at first. In traffic pattern-recognition, fuzzy neural network is adopted to distinguish the traffic pattern, this method has very good recognition capability to the complicated traffic flow and can play very good guidance function to adopt the corresponding group control policy to EGCS according to different traffic.In order to be adaptive to different traffic patterns, a multi-objective control method that can adjust the weight of every objective according to the changing traffic patterns is presented in this paper. An evaluation function is established to realize the control objects for reducing average waiting time, riding time and run times of elevators.In order to test the performance of group control algorithms, a set of software of virtual environment of elevator group control system is developed. Make programs of the algorithm and then it can be tested in the simulation environment. The simulation results prove that the multi-objective control method is efficient.