Model Development for Forecasting Regional Population - Using Cohort-Component Method and Markov Chain Model -

Population of a city is a leading indicator in urban planning. Because sizes of various infrastructure including housings, streets and facilities related to urban services are mostly determined according to the population of a target year. Hence, an inaccurate forecasting of population will result in inefficient resource allocation within a city. However, there is usually a considerable difference between a forecasted population and the actual population. The study attempts to develop a more accurate population forecasting model. specifically, In the model the cohort-component method is used for natural growth and the Markov chain model for migration. Also, To enhance the accuracy of the stationary Markov chain model, the study adopts the non-stationary model by adding migration factors including the housing developments and the distance between the move-in and the move-out city. The study specifically shows that the model developed reduces the difference between forecasting and actual population. Also, using the model the policy scenarios on balanced development are examined