| Summary: | Thesis: S.B., Massachusetts Institute of Technology, Department of Materials Science and Engineering, June 2014. === Cataloged from PDF version of thesis. "May 2013." === Includes bibliographical references (pages 86-87). === Connecting structure to properties, and optimizing properties by controlling structure is one of the fundamental goals of materials science and engineering. No where is this connection more apparent than with biomaterials, whose unparalleled properties are the result of the evolution via cumulative selection of highly specialized structures. Beyond biomaterials, cumulative selection offers a generalizable model for materials optimization via accumulative of beneficial mutations in a material's genome that improve the properties for a given function. A genetic algorithm is one method for applying the principals of cumulative selection to material's optimization. One of unique property that cumulative selection generated was the ability of trabecular bone to optimize and adjust its structure in vivo in response to changes in its loading conditions. This work presents a model for trabecular microstructure evolution using a genetic algorithm, the same mechanism through which that ability evolved. The algorithm begins by translating a trabecular genome into a developed structure. It then simulates the structure's response under an applied load and selects for the genome which translates into the best structure. The selected genome is then replicated and mutated. Simulations of microstructure evolution consist of iterating through this process across multiple generations. A series of simulations was conducted demonstrating the ability of the algorithm to improve trabecular architecture. The systems tended to converge to a uniform stress distribution, after which additional generations of evolution had no effect on performance. During the simulations it was found that the length of the computation was most sensitive to the number of offspring per generation. Although focused on trabecular microstructure, this work establishes the use of a genetic algorithm as a general tool for material's optimization. === by Samuel W. L. Shames. === S.B.
|
|---|
