Theory & Evolutionary Algorithms
Optimization of the Gecko Adhesive System
My research focuses on the biomechanics, evolution, and ecology of the gecko adhesive system. To consider how the gecko adhesive system may be adapted to particular environments, I use field, lab, and museum based experiments, observations, and simulation. I am currently using computed tomography (CT) to image the gecko adhesive system and generating idealized finite element models. Finite element analyses will allow us to understand the causal relationship between morphology and performance. We will then use genetic algorithm simulations to identify optimum morphologies, maximizing performance on each species’ native perch texture.
Evolutionary computation and machine learning for bioinformatics
James A. Foster Ph.D.
We are applying evolutionary computation and other machine learning algorithms interpret large biological datasets. For example, we have developed and algorithms that use genetic programming and genetic algorithms for multiple (DNA or protein) sequence alignment, for phylogenetic reconstruction, and for disease-microbial ecology association studies.
Algorithmic efficiency for computational biology
James A. Foster Ph.D.
We help colleagues design efficient algorithms and implement them as effective programs. This uses our expertise in computational complexity theory, the theory and practice of algorithm analysis, and algorithmic design.
Teaching Evolution (K-12)
Terence Soule Ph.D.
Understanding evolution is critical to understanding the natural world and evolutionary theory has become a fundamental feature of scientific fields, including medicine, economics, sociology, political science. However, acceptance of evolution is surprisingly low, particularly in the United States, but also in many European countries and, even among individuals who accept evolution, understanding of the basic tenants of evolution are often incorrect. The goal of this project is develop educational resources, such as The Ladybug Game to teach evolution in grades K-12.
COTSbots: Inexpensive Powerful Robots for Evolutionary Research
Robert Heckendorn Ph.D. and Terence Soule Ph.D.
Previously, robots available for researchers have been very expensive, underpowered, custom designed units ill-suited for research in massive evolutionary and swarm robotics. We are designing robots out of Commodity Off The Shelf (COTS) parts including smart phones, RC cars, and custom built software libraries that create an affordable powerful alternative. We provide publicly available quick assembly instructions that require no mechanical or electrical expertise. The result is affordable swarms of robots for our robotics research projects which emphasize populations of evolving and adapting robots. Visit project website here.
Nature Inspired Traffic Optimization for Disaster Scenarios
Juan Marulanda, Ahmed Abdel-Rahim Ph.D., and Robert Heckendorn Ph.D.
During disasters such as a hurricanes, tsunamis, or urban attack by gigafauna such as Godzilla, optimizing the safety of the population is a critical problem. Features of these disasters are the large volume of traffic, changing parameters of zones of safety and traffic volume, and the changing nature of the disruption of available traffic routes. This is a very complex and important problem in realtime adaptive optimization. A similar but simpler problem is network packet routing; and one of the best solutions for this problem involves optimization techniques taken from nature. In collaboration with the Civil Engineering Department, we will explore several nature based optimization techniques including those based on evolution to create realtime evacuation control algorithms.
Theoretical Limits of Epistasis in Natural Evolution
Max McKinnon, Dan Weinreich Ph.D., Ian Dworkin Ph.D., and Robert Heckendorn Ph.D.
Of great interest to evolutionary biologists is the trajectory of populations on fitness landscapes. Any nontrivial landscape requires some form of epistasis and yet if the landscape is too complex evolution would become ineffective. This suggests patterns of simplicity in the landscapes that would be critical to understanding the nature of evolution as an optimization algorithm. Using data from more than 20 previously published research works in various areas of biology, we are examining the statistical nature of higher order epistasis in the natural world. Furthermore, we are generating tools to support biologists in higher order epistatic analysis as well is providing mathematical insight into the nature of naturally occurring epistasis.