自複雜科學興起以來，已漸成為探討都市發展演變的典範。相較於傳統的數理模型，以複雜理論探討都市發展可彌補其不足之處，包含了可處理不同類型的個體，以及可模擬出更貼近真實世界中個體的有限理性等特性。近年來，基於個體之模擬 (Agent-Based Modeling, ABM) 在社會科學研究逐漸受到重視。其概念與Cellular Automata (CA) 相似，但是ABM 相對於CA 在探討都市發展上更貼近真實環境。雖然如此，ABM 仍有其侷限之處，如甚少研究針對於無法碰觸的抽象個體，像是決策的選擇機會。空間垃圾桶模型 (Spatial Garbage Can Model, SGCM) 正是以一種概念性的模型，來探討都市發展，因此本研究以ABM 重新詮釋SGCM，並加入了適應性與可逆性之概念，將SGCM 重新表現為一人工社會，來探討此系統是否為一自我組織的系統，以及制度與空間何者對於都市發展影響較為顯著。最後結果發現，系統呈現自我組織的現象，而且發現制度在解決問題的效率上較空間來的高，以及空間之相對自我組織性較制度為高。且制度與空間之間相互作用顯著，顯示在都市發展中，制度與空間是互相影響的，因此在都市規劃上，應同時著重兩者，而非僅以空間規劃為主。

Since the rise of Complex Science, it has emerged as the new paradigm of urban evolution. Compared with the traditional mathematical model, exploring the urban development with complexity theory can help us understand the structure of a city,such as the handling of different types of agents and the characteristics of limited rationality. Recently, Agent-Based Modeling (ABM) has been introduced in social sciences. ABM is similar to Cellular Automata (CA), but it is more effective in modeling the real world of urban development. Nevertheless, ABM has its limitation in studying intangible entities such as the strategy of decision making.Spatial Garbage Can Model (SGCM) is a conceptual model to explore urban development. This research intends to reconstruct the SGCM through ABM, and incorporate adaptability and irreversibility, by viewing the SGCM as an artificial society and testing this system in terms of self-organization, as well as finding out institutional and spatial emergence which influence urban developent processes.Results showed that the system self-organized itself, that institutional structures were more efficient in problem-solving than spatial structures, and that the characteristics of self-organization of spatial structures were more significant than institutional structures. In addition, the interaction between spatial and institutional structures was significant. The implication is that we should focus on both in making plans for urban development, rather than solely on spatial structures as traditionally practiced.