新光三越百貨公司廣場行人之自我組織現象觀察與模擬分析

OBSERVATION AND SIMULATION ON SELF -ORGANIZATION PHENOMENA OF PEDERTRIAN MOVEMENT IN SHIN-GUANG DEPARTMENT STORE PLAZA

本研究籍由錄影調查台北車站新光三越百貨之行人廣場，探討該廣場的行人移動特性，從行人移動之互動規則，建構行為者基礎模型，以探討其自我組織現象。在不同時段觀察發現，行人分佈均收斂成某一固定形態，駐足人群的群落規模與數量的關係呈現冪次趨勢，為自我組織臨界的特徵現象。本研究進而以NetLogo軟體，構建模擬模型以分析行人之移動行為，其中環境狀態變數為吸引力與進入阻力。模擬模型經以廣場現況在低、中、高三種行人密度的駐足人群分佈驗證，顯示模擬模型具有良好的詮釋能力。進一步模擬分析行人密度與延滯時間的關係發現，穿越行人延滯時間會隨著密度的增加而增加，本研究並嘗試詮釋行人移動系統於秩序、自我組織與混亂之轉變過程。本研究另分析廣場內設置咖啡座的影響，考量增加平均駐足吸引力與避免增加穿越行人延滯時間，分析得咖啡座之適當設置位置及其對行人移動之影響。由本研究之應用分析案例顯示，模擬模型可作為輔助廣場空間配置，分析廣場家具與行人移動互動影響之工具。

The study starts with observing and investigating on pedestrian movements in Shin-Guang department store plaza. An agent-based model that is based on the movement rules observed from field observation, and is built to discuss the self-organization phenomena of pedestrian movement in department store plaza. It is found that pedestrians always cluster together with some patterns in different observation periods, and the relationship of cluster scale and its frequency follows to Power Law, which is one of the evidences of self-organization phenomena. Using the characteristics of pedestrian movement behaviors and interaction rules, a pedestrian agent-based model used by NetLogo simulation software is constructed. This model uses three modules to direct pedestrian movement that pedestrian takes the way of maximum attraction and minimum walk-ability, and can simulate pedestrian movements actually. From the empirical study on movement delay according to pedestrian density, it is discussed that the phenomena of pedestrian system changes from order to self-organization, and to chaos. The second empirical study simulates and discusses the influence of allocating a coffee shop in plaza, and gets the proper position. This pedestrian agent-based model can be a good tool to analyze the movements, interactions between layout of street furniture and pedestrian movements.