LEARNING SYSTEM BASED ON SYNTHETIC BRAIN TO PREDICT CONGESTION PEAK GENERATING PERIOD FOR DISTRIBUTION OF ELEVATOR CAGE

摘要

PURPOSE: To provide a simple learning method and a complicated learning method. CONSTITUTION: This learning system can predict times for start and end of a peak period. According to a simple method, the time when a loading capacity of a car reaches a predetermined level is recorded everyday, and the time for start and end of the peak period on the following day is measured by approximate calculation by means of exponential functions based on these time information. On the other hand, according to a complicated method, the number of passengers boarding at lobbys and the frequency of arrival and departure of cars are collected per a fixed time interval everyday. The number of passengers boarding at lobbys and the frequency of arrival and departure of cars on the day can be predicted based on the data collected until the previous day. The optimum results of prediction can be obtained by combining real time prediction by the use of the data on the day, time series prediction by the use of the data until the previous day, and time series prediction by the use of the data until the previous day.