A refinement of the gravity model for competitive facility location

Most competitive location models assume that as the distance increases, the patronage of a facility declines at the same rate regardless of the facility attractiveness. We observed that the rate at which patronage declines is slower for more attractive facilities. Customers are willing to drive long distances to patronize an attractive facility. Less attractive facilities hardly attract customers from long distances. We propose to modify the effect of attractiveness on the appeal of the facility to customers. Many methods for estimating the market share captured by a facility can be modified to incorporate such property. We implemented the new modification on the gravity model and tested it on a real data set of shopping malls in Orange County, California. The approach was statistically validated and is computationally straightforward to implement with existing software such as R. Further, new facility location decisions are expected to be more accurate due to the improved market share estimates by the proposed model.