Bowl championship series vulnerability analysis

Authors:
Caio Soares;George Page;Jonathan MacDonald;Sanjeev Baskiyar;John Hamilton, Jr.;Gerry Dozier
Affiliations:
Auburn University, AL;Auburn University, AL;Auburn University, AL;Auburn University, AL;Auburn University, AL;North Carolina A&T State University, Greensboro, NC
Venue:
TAPIA '07 Proceedings of the 2007 conference on Diversity in computing
Year:
2007

Citing 1
Cited 0

Evolutionary computation: comments on the history and current state

IEEE Transactions on Evolutionary Computation

Quantified Score

Hi-index	0.00

Visualization

Abstract

The Bowl Championship Series (BCS) is the system used to select the top two college football teams to play in a championship game at the end of a season. During the nine seasons played under the BCS, few have not caused considerable debate and controversy. We believe that an analysis of the BCS and its components is therefore required. This work uses a GRNN to model the eight polls that compose the ranking system of the BCS. Then, a genetic algorithm (GA) is used to evolve season schedules that would perform poorly under the current BCS. These difficult seasons are then analyzed to highlight weaknesses in the current model.