9522882 Francis This research is concerned with the location of a facility to serve a large number of customers. A common measure of the quality of a location decision is the distances that customers will have to travel to receive service from the facility. However, when the customer pool is large, the sources of customers may be aggregated to reduce problem size and improve computational efficiency. The objective of this research is to employ aggregation methods to develop and solve facility location problems and to use the model to calibrate the quality or error associated with such aggregation since aggregation involves using approximate customer locations instead of actual locations. Approximate locations obviously lead to an approximate solution for the distances customers will actually travel to reach the facility. The research will develop and test aggregation methods which are designed to control error and exploit special problem structures to improve computational efficiency. The testing process will involve both computer implementation of the methods and testing using real-world data. The use of aggregation/approximation is very common in applied facility location when there are many facility users/customers. Despite the common acceptance of the approach, little analytical work exists to evaluate the quality or error of aggregation schemes. The developed aggregation and testing algorithms in this research are designed to keep approximation error small. Thus, the algorithm should be useful to urban and regional planners, transportation/logistics specialists, and geographers who must choose aggregation schemes for large -scale location problems.
