Apply the Naïve Bayes Classifier to Elevator-Group Control System in a Multi-Purposed Building

• Main Authors: Cuo-Chan Peng, 彭國昌
• Other Authors: Jin-Ling Lin
• Format: Others
• Language: zh-TW
• Published: 2002
• Online Access: http://ndltd.ncl.edu.tw/handle/37998184389084049750

碩士 === 華梵大學 === 工業管理學系碩士班 === 90 === In this paper, here derive a new method, based on the Naive Bayesian Network, to decide the efficient stopping floors for an elevators’ group. The variable parameters, in our model, are the number of elevators, simulation time, the number of floors, and several categories of persons. In the initial of learning stage, we randomly and dynamically generate these related parameters. Then, we followed by training the Naive Bayesian Network in order to modify the stopping floors and change the directions for the whole set of elevators. Finally, another set of testing data was used to judge the proposed model. Simulation results strongly encourage the proposed algorithm. In general, it is more efficient than previous works. The results indicate that elevator numbers factor have significant influence to the elevators performance. The more the elevator numbers the more reduce the people net waiting time. And Bayes elevators performance is better than tradition elevators. Key words: Naive Bayesian Classifier, Concept Learning, Elevators Group