OUT OF SAMPLE PREDICTION OF MULTI-CRITERIA GEOMETRIC DISPERSION THEORY WITH COMPARATIVE STUDY OF WIND ENERGY AND TRANSPORTATION
Date
2024-06-02
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Case Westren Reserve University
Abstract
This thesis presents a pivotal study focused primarily on the application of Multi-Criteria Decision Making (MCDM) methodologies to identify the most suitable locations for wind farm development across thirteen regions in Saudi Arabia, an endeavor underscored by the country's significant potential for renewable energy. This research takes into account critical factors such as average wind power density, wind speed, and terrain suitability to guide the decision-making process. In parallel, the thesis also revisits an ancillary study based on a 1987 mode choice survey involving 210 travelers making trips between Sydney, Canberra, and Melbourne, offering a comparative perspective on the utility of MCDM methods across distinct domains.
Employing an extensive array of MCDM techniques—including Linear Additive Utility functions, Geometric Dispersion Theory models, Keeney Multiplicative Utility (both with and without value functions), and a Z-Goal-programming method tailored for wind farm location selection and Transportation Mode selection—this analysis is thorough in its critical evaluation of these methodologies' efficacy in forecasting optimal choices. The precision of these models is quantitatively assessed using the Sum of Squares Error (SSE) metric, complemented by a thorough validation framework that includes both a 70% in-sample and a 30% out-of-sample comparison, alongside comparative assessments across models.
The culmination of this research underscores the unparalleled predictive prowess of Geometric Dispersion Theory (GDT) models, which demonstrated the lowest SSE, thereby affirming its superior capability in accurately determining optimal wind farm locations in Saudi Arabia. By extension, the GDT models' effectiveness was also mirrored in the mode choice study, further attesting to their robust applicability across different decision-making contexts. This thesis enriches the academic discourse on MCDM applications within the realms of renewable energy site selection and transportation.
Description
Keywords
Geometric Dispersion Theory, Simple Additive function, decision-making