Doctoral Dissertations
Date of Award
5-2025
Degree Type
Dissertation
Degree Name
Doctor of Philosophy
Major
Geography
Major Professor
Hyun Kim
Committee Members
Shih-Lung Shaw, Nicholas Nagle, Hugh Medal
Abstract
This dissertation offers a new perspective on Geographic Information Science (GIScience), particularly in location problems and modeling, by creatively integrating two quantitative geospatial methods – spatial statistics and spatial optimization problems. This dissertation consists of three essays that demonstrate the effectiveness of spatially informed solving approaches, which is a unique solving approach to solve spatial optimization by leveraging additional spatial information provided by spatial statistical methods, in different classes of spatial optimization problems: p-dispersion, hub location, and p-obnoxious location problems. The first chapter introduces the spatially informed p-dispersion model, which leverages the unique characteristics of the distance-based spatial statistical measure, Ripley’s local K-function, as an indicator of essential decision variables. The spatially informed dispersion model using distance-based spatial properties outperforms standard counterparts in most controlled instances. The second chapter proposes a spatially informed solution approach for hub location problems. By employing unique hub inclusion convex hull detection methods, the relationship between location-allocation behavior and innate network autocorrelation in the hub location model is identified. The network autocorrelation structure of the flow network is integrated into the model-building process to develop the spatially informed hub location model, which improves computational efficiency compared to the traditional hub location model. The third chapter addresses the spatially informed approach to building the obnoxious location model, overcoming the unrealistic location-allocation behaviors observed in traditional obnoxious location problems. By uniquely integrating spatial statistical measurement into the model specification, the spatially informed obnoxious location problem provides a better spatial arrangement of the optimal location of facilities applicable to practical planning situations. In essence, this dissertation provides a new foundation for developing more efficient solution approaches by leveraging innate spatial characteristics and structures, at the same time it offers insights into the general location-allocation behaviors of the models underlying geographic patterns.
Recommended Citation
Oh, Changwha, "Spatially Informed Solving Approaches for Spatial Optimization Problems. " PhD diss., University of Tennessee, 2025.
https://trace.tennessee.edu/utk_graddiss/12401