Font Size: a A A

Research Of Intelligent Elevator Group Control Dispatching Methods In Virtual Simulation Environment

Posted on:2004-12-30Degree:DoctorType:Dissertation
Country:ChinaCandidate:Q ZongFull Text:PDF
GTID:1102360122482134Subject:Power electronics and electric drive
Abstract/Summary:PDF Full Text Request
The intelligent elevator group control dispatching methods are summarized in this paper. And several key problems on elevators such as traffic flow forecasting, traffic pattern recognition, elevator group control methods in up-peak, idle and stochastic interfloor, intelligent multi-pattern elevator group control dispatching methods and elevator virtual simulation environment are studied, and these have directive significance in theory and practical value to intelligent elevator group control system research.This paper, based on the characteristics of traffic flow, constructs a traffic flow forecasting model which is based on time series neural network (NN), and put forwards a method of adjusting structure of NN prediction to improve the precision of forecasting. A fuzzy neural networks model applied to traffic pattern recognition is established. This model might involve two steps and in each step a fuzzy neural network is used. Such a model simplifies the structure of networks and samples and improves real-time of traffic pattern recognition. The samples of networks training are obtained by elevator expert knowledge, and a three-step hybrid study method is adopted to adjust the two networks. Testing results prove that each traffic pattern can be distinguished exactly by two fuzzy neural networks. Monte Carlo Method is used to handle real traffic data with which the elevator traffic probability simulation model is established. Based on this model, elevators in idle traffic model can be dispatched. An objective function named uniform UPPINT is presented, and the dynamic zoning method in peak is treated as an optimization problem, and uses dynamic programming method to solve dynamic zoning problem. A multi-objective function in elevator system for interfloor traffic mode is put forward, and Genetic Algorithm based on the multi-objective function is adopted for dynamic optimization dispatching strategy in order to improve service of special floor and optimize performance of the whole system at the same time. Simulation results show efficiency of the method above.A batch service queuing model consisting of a single queue served by multiple finite-capacity bulk servers of elevator in up-peak is proposed. An Markov Decision Processes (MDP) model for the up-peak elevator dispatching problem, state transition probabilities, one-step cost, discounted criteria and object function are described in detail. Dynamic programming equation of optimal dispatching in up-peak is established. In order to simplify the process of solving dynamic programming equation, the lemmas and corollaries concerning properties of the optimal value function in up-peak of elevator are presented and proved; the structure of optimal dispatching policy in up-peak is constructed, and conclusion that the optimal dispatching policy in up-peak is a threshold policy is obtained.The structure of intelligent multi-pattern elevator group control system is presented, and this structure is made up of handling information unit and elevator optimal dispatching unit. Genetic-based fuzzy neural network is used to fuse a lot of data of information handling unit, and the method of using Genetic Algorithm to optimize general network parameter and using BP Algorithm to modulate and optimize local parameter is presented. Niche Genetic Algorithm is applied to optimize general parameter in elevator optimal dispatching unit to adapt to the changing of traffic flow of intelligent multi-pattern. Simulation results show that the method is effective. General structure and realization of the virtual simulation environment of intelligent elevator group control system are described in detail.
Keywords/Search Tags:traffic flow forecasting, traffic pattern recognition, multi-objective function, markov decision processes, dynamic programming equation, intelligent elevator group control dispatching method, virtual simulation environment
PDF Full Text Request
Related items