Adaptation and Speciation in Genetic Modeling of Physical Systems
Abstract
A genetic algorithm has been used to allow a population of first principles lithium-ion(Li-ion) battery cell models to adapt to a database obtained from life-testing 50-Ah Li-ion cells. This adaptive process allows the fundamental processes that control the cell performance to be extracted from the life test data. During the process of genetic adaptation, the population of battery cell models has been observed to separate into two distinct species of models. The emergence of separate species is recognized by the inability to produce viable offspring models from mixtures of the genetic codes for the two species. In this situation, the two species are found to co-exist for up to many hundreds of generations. However, one of the two species eventually suffered extinction because it did not retain adequate genetic diversity and because its fitness did not continue to improve as did that of the other species. The dynamics involved in genetic adaptation of battery cell models, including the roles of speciation and extinction, are described.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 15, 2009
- Accession Number
- ADA514694
Entities
People
- Albert H. Zimmerman
Organizations
- The Aerospace Corporation