Huang, Arthur and David Levinson (2011) Why retailers cluster: an agent model of location choice on supply chains volume 38(1) pages 82 - 94. [download from Environment and Planning b website]
Abstract. This paper investigates the emergence of retail clusters on supply chains comprised of suppliers, retailers, and consumers. An agent-based model is employed to study retail location choice in a market of homogeneous goods and a market of complementary goods. On a circle comprised of discrete locales, retailers play a noncooperative game by choosing locales to maximize profits which are impacted by their distance to consumers and to suppliers. Our findings disclose that in a market of homogeneous products symmetric distributions of retail clusters arise out of competition between individual retailers; average cluster density and cluster size change dynamically as retailers enter the market. In a market of two complementary goods, multiple equilibria of retail distributions are found to be common; a single cluster of retailers has the highest probability to emerge. Overall, our results show that retail clusters emerge from the balance between retailers' proximity to their customers, their competitors, their complements, and their suppliers.
The software underlying this paper, CLUSTER, has just been made available on the STREET website, so you are free to test and reproduce the results yourself. The software is free and open source, so feel free to modify, please let us know if you do anything.