At the MSU Digital Evolution Laboratory (Devolab), we perform experimental studies on digital organisms with the twin goals of improving our understanding of how natural evolution works, and applying this knowledge to solving computational, engineering, and biological problems.
Several new Devolab publications have been accepted to appear at Artificial Life 13.
“Digital Evolution Exhibits Surprising Robustness to Poor Design Decisions” by David M. Bryson and Charles Ofria
When designing an evolving software system, a researcher must set many aspects of the representation and inevitably make arbitrary decisions. Here we explore the consequences of poor design decisions in the development of a virtual instruction set in digital evolution systems. We evaluate the introduction of three different severities of poor choices. (1) functionally neutral instructions that water down mutational options, (2) actively deleterious instructions, and (3) a lethal die instruction. We further examine the impact of a high level of neutral bloat on the short term evolutionary potential of genotypes experiencing environmental change. We observed surprising robustness to these poor design decisions across all seven environments designed to analyze a wide range challenges. Analysis of the short term evolutionary potential of genotypes from the principal line of descent of case study populations demonstrated that the negative effects of neutral bloat in a static environment are compensated by retention of evolutionary potential during environmental change.
“Finger-painting Fitness Landscapes: An Interactive Tool for Exploring Complex Evolutionary Dynamics.” by Luis Zaman, Charles Ofria, and Richard E. Lenski
Evolution involves only a few simple processes, yet the resulting dynamics are surprisingly rich and complex. Sewall Wright developed the metaphor of fitness landscapes to provide deeper insight into the complex workings of evolution. Here we extend that metaphor by visualizing in real time the dynamic processes that drive evolution. We allow viewers to construct fitness landscapes interactively while also varying key parameters including population size, mutation effect size, mode of reproduction (asexual or sexual), and densitydependent selection. This application is both mechanistic and visual, and it thereby allows the active exploration of evolutionary processes. We walk the reader through several exercises including both simple activities potentially suitable for education and examples of deeply conceptual topics that remain the focus of current research in evolutionary biology.
“Evolutionary Potential is Maximized at Intermediate Diversity Levels” by Bess L. Walker and Charles Ofria
Diversity in a population is often cited as a major facilitator for the evolution of new complex features. The intuition behind this dynamic is that if a population is exploring multiple regions of a fitness landscape, more opportunities exist to find new functionality. We use the digital evolution software platform Avida to explore the effect of multiple limited resources on phenotypic Shannon diversity and, in turn, on evolvability of populations. We show that Shannon diversity peaks at intermediate levels of resource availability to the population, and we map the evolvability of a complex computational task on this availability-diversity gradient. While the evolvability of the complex task is highest at intermediate availabilities, it does not peak at the same resource inflow level as Shannon diversity, and it is more robust than diversity in its response to inflow level. These results indicate that while phenotypic Shannon diversity may play into the evolution of complex features, the selective pressures caused by diversity cannot be the only — or indeed even the main — pressures behind such evolution.