Research On Intelligent Control Of Elevator Group

New types of buildings, especially the so-called intelligent buildings have caused more and more variations in building usage. Efficient elevator group control system (EGCS) is important for the operation of these kinds of large buildings. In the EGCS, there are many uncertain factors such as number of passengers, hall calls, and car calls in any time. Therefore, what the elevator group control system deals with is a stochastic, multi-object decision-making with complex, nonlinear and uncertainty. It is very difficult to obtain high quality performance by using traditional control method due to the uncertainty of traffic flow and complexity of elevator control. In this paper, a new elevator group control method is considered, which uses fuzzy neural networks to indentify the traffic pattern and then apply the improved genetic algorithm to deal with the elevator dispatching problem.Firstly, the development and current status of EGCS both in home and abroad are reviewed. By summarizing the shortcomings of the previous methods adapted in EGCS, A new mathematic model for elevator group control algorithm is proposed based on multi-target planning, and it establishes an multi-target evaluate function, considering the average waiting time, average riding time, the crowded degree of the car and the power consumption of the system.Secondly, elevator traffic flow is fundamental in elevator group control system. Accurate elevator traffic flow model is crucial to the elevator system configuration and the dispatching of elevator group control system, especially to the elevator system configuration in yet-to-be-commissioned new buildings. A new Traffic Patten identification model is proposed which is based on fuzzy neural network. This can increase the veracity of the whole traffic pattern identification model. At last, the sampled data are trained and tested Matlab software, and the simulation results indicate that the proposed identify model has very small error.Finally, an improved genetic algorithm is proposed which makes use of a novel fitness function to evaluate the individuals. When common GA technology is applied, drawbacks of slow convergent speed and sometimes just getting local optimum solution are found by test. In response to it, an improved GA is introduced, which features new real number coding the initial populations, automatic changes the size of the population according to the state of the calling information. An establishes base type adaptive evolution function, as well as performs self-adaptive cross and variance operations while ensuring selection of optimum saving strategy. The intelligent dispatching algorithm based on improved genetic algorithm can be applied in four different traffic patterns: Up-Peak Traffic Pattern, Down-Peak Traffic Pattern, Random Inter-Floor Traffic Pattern and Idle Traffic Pattern. Simulation of the allocation system based on improved GA indicates that the design is feasible and outstanding.