2015 年 80 巻 715 号 p. 2001-2010
Distance measures, such as Euclidean distance, network distance, and time distance, have been used for various urban models. In this paper, we propose a novel distance measure based on the spatiotemporal distribution of residents, which is called "Distance Measure based on Spatiotemporal Coexistence of Residents (DM_SCORE)". It is defined as a square root of Jensen-Shannon Divergence, and can be calculated using person trip survey data. We discuss the difference between DM_SCORE and ordinary distance measures, and apply it to some urban models. The results of some numerical analyses using actual data demonstrate that DM_SCORE can be of great value in being incorporated in urban models.