One of the most pervasive trends in the global airline industry over the past few three decades has been the rapid development of low cost carriers (LCCs). Australia has not been immune to this trend. Following deregulation of Australia’s domestic air travel market in the 1990s, a number of LCCs have entered the market, and these carriers have now captured around 31 per cent of the market. Australia’s LCCs require reliable and accurate passenger demand forecasts as part of their fleet, network, and commercial planning and for scaling investments in fleet and their associated infrastructure. Historically, the multiple linear regression (MLR) approach has been the most popular and recommended method for forecasting airline passenger demand. In more recent times, however, new advanced artificial intelligence-based forecasting approaches – artificial neural networks (ANNs), genetic algorithm (GA), and adaptive neuro-fuzzy inference system (ANFIS) - have been applied in a broad range of disciplines. In light of the critical importance of passenger demand forecasts for airline management, as well as the recent developments in artificial intelligence-based forecasting methods, the key aim of this thesis was to specify and empirically examine three artificial intelligence-based approaches (ANNs, GA and ANFIS) as well as the MLR approach, in order to identify the optimum model for forecasting Australia’s domestic LCCs demand. This is the first time that such models – enplaned passengers (PAX) and revenue passenger kilometres performed (RPKs) – have been proposed and tested for forecasting Australia’s domestic LCCs demand. The results show that of the four modeling approaches used in this study that the new, and novel, ANFIS approach provides the most accurate, reliable, and highest predictive capability for forecasting Australia’s LCCs demand. A second aim of the thesis was to explore the principal determinants of Australia’s domestic LCCs demand in order to achieve a greater understanding of the factors which influence air travel demand. The results show that the primary determinants of Australia’s domestic LCCs demand are real best discount airfare, population, real GDP, real GDP per capita, unemployment, world jet fuel prices, real interest rates, and tourism attractiveness. Interestingly three determinants, unemployment, tourism attractiveness, and real interest rates, which have not been empirically examined in any previously reported study of Australia’s domestic LCCs demand, proved to be important predictor variables of Australia’s domestic LCCs demand. The thesis also found that Australia’s LCCs have increasingly embraced a hybrid business model over the past decade. This strategy is similar to LCCs based in other parts of the world. The core outcome of this research, the fact that modelling based on artificial intelligence approaches is far more effective than the traditional models prescribed by the International Civil Aviation Organization (ICAO), means that future work is essential to validate this. From an academic perspective, the modelling presented in this study offers considerable promise for future air travel demand forecasting. The results of this thesis provide new insights into LCCs passenger demand forecasting methods and can assist LCCs executives, airports, aviation consultants, and government agencies with a variety of future planning considerations.
展开▼
