Research On Forecast And Early Warning Method About Flight Delay Based On Data Warehouse

Flight delay is a worldwide problem which troubled the staffs of civil aviation and the general passengers, the normal rate of flights is generally about 80% in worldwide. There are many factors causing flight delay, the factors are generally divided into controllable factors and uncontrollable factors. Controllable factors mainly related to flight deployment, aircraft maintenance, ensure sailing materials, the crew and so on; while uncontrollable factors are mainly related to the carrier(airline) can not control factors, such as the weather, air traffic control, flow control, emergencies, reasons of security check.Flight delay related to airlines, air traffic control, airport, security check, aviation fuel,passengers, etc., the delay not only directly induce the increase of operating costs for airlines,difficulty of airport to run the command and management, but also lead to passengers complaints and legal disputes what impact the air service quality and even a series of aviation safety issues. For a large-scale hub airports, how to effectively forecast the development trend of flight delay, under the large-scale flight delay conditions, provide the early warning information timely for the airport management personnel and the passengers, as much as possible to reduce the flight delay distress and the losses, at the same time, will be play a very important role in improving management and service level of civil aviation.This paper around the problem of flight delay management in domestic large hub airports,take the Capital International Airport as the research object, through analysis the actual running data structure in the airport, introduce the data warehouse technology to analysis the characteristics of flight delay information of large-scale hub airport, given the design and implementation process of flight delay data warehouse of airport; focus on the research of forecast methods of flight delay in airports, advance a combined model which combine the KNN with support vector regression based on data warehouse, and adopt the Capital International Airport's actual data of flight delay to making experimental analysis, the assessment methods of experimental results has been given; combine the status of airport about flight command and management, achieve the development of simulation platform about flight delay forecasting and early warning which oriented towards hub airport.