Evolution is a highly contingent process, where the quality of the solutions produced is affected by many factors. I explore and describe the contributions of three such aspects that influence overall evolutionary potential: the prior history of a population, the type and frequency of mutations that the organisms are subject to, and the composition of the underlying genetic hardware. I have systematically tested changes to a digital evolution system, Avida, measuring evolutionary potential in seven different computational environments ranging in complexity of the underlying fitness landscapes. I have examined trends and general principles that these measurements demonstrate and used my results to optimize the evolutionary potential of the system, broadly enhancing performance. The results of this work show that history and mutation rate play significant roles in evolutionary potential, but the final fitness levels of populations are remarkably stable to substantial changes in the genetic hardware and a broad range of mutation types.
Read
