| Bio | 
| Poster |
| Paper |
Bryan Culbertson,
Mark Kokoska
Lafayette College
Subject Listing - Computer Science
Advisor: Dr. Chun Wai Liew
Friday, Poster Session 5, Presentation Kiosk 28 A, Health & Fitness Center
AN IMPROVED METHOD FOR VARYING EXPLORATION AND CONVERGENCE IN ITERATED GENETIC ALGORITHMS
Rooted in evolutionary biology, genetic algorithms are valuable optimization techniques for solving complex biological problems. Current genetic algorithms, however, cannot solve these increasingly complex problems in an acceptable amount of time. Our approach applies an iterative genetic algorithm that uses new patterns of convergence and exploration to quickly converge on a satisficing answer and then explore more possible solutions. These patterns of exploration and convergance coupled with other mechanics for moving between one iteration and the next allow the iterated genetic algorithm to be more effective. Using our algorithm we optimized two biological models each with complex search spaces. The first such model involves optimizing the angles of various bones within a snake's jaw to find the maximum cross sectional area of the jaw's gap.
The second model involves determining the parameter values that minimize the difference between the calculated swimming motion of a sunfish generated by the model and the actual motion deduced from video of the modeled sunfish swimming in a flume. The difficulties in optimizing both of these models for accuracy include the substantial size of the search space, the minuscule number of usable points in that space, and the complex interdependence between each of the model's parameters. These obstacles manifest themselves in an irregularly shaped landscape that standard optimization algorithms cannot easily navigate. Results indicate that our algorithm generates more accurate solutions and does so sooner than previous algorithms. In addition to biological problems, our approach of converging before exploration could potentially be applied to any such sufficently complex optimization problem. The method has the potential to be used to optimize a growing field of practical problems in real world domains and applications.
Advisor: Dr. Chun Wai Liew, Associate Professor, Computer Science, Lafayette College, Easton, PA