Limits...
A predictive machine learning approach for microstructure optimization and materials design.

Liu R, Kumar A, Chen Z, Agrawal A, Sundararaghavan V, Choudhary A - Sci Rep (2015)

Bottom Line: These challenges render traditional search-based optimization methods incompetent in terms of both searching efficiency and result optimality.A systematic framework consisting of random data generation, feature selection and classification algorithms is developed.Experiments with five design problems that involve identification of microstructures that satisfy both linear and nonlinear property constraints show that our framework outperforms traditional optimization methods with the average running time reduced by as much as 80% and with optimality that would not be achieved otherwise.

View Article: PubMed Central - PubMed

Affiliation: EECS Department, Northwestern University, Evanston IL, USA.

ABSTRACT
This paper addresses an important materials engineering question: How can one identify the complete space (or as much of it as possible) of microstructures that are theoretically predicted to yield the desired combination of properties demanded by a selected application? We present a problem involving design of magnetoelastic Fe-Ga alloy microstructure for enhanced elastic, plastic and magnetostrictive properties. While theoretical models for computing properties given the microstructure are known for this alloy, inversion of these relationships to obtain microstructures that lead to desired properties is challenging, primarily due to the high dimensionality of microstructure space, multi-objective design requirement and non-uniqueness of solutions. These challenges render traditional search-based optimization methods incompetent in terms of both searching efficiency and result optimality. In this paper, a route to address these challenges using a machine learning methodology is proposed. A systematic framework consisting of random data generation, feature selection and classification algorithms is developed. Experiments with five design problems that involve identification of microstructures that satisfy both linear and nonlinear property constraints show that our framework outperforms traditional optimization methods with the average running time reduced by as much as 80% and with optimality that would not be achieved otherwise.

No MeSH data available.


Related in: MedlinePlus

Framework of material structure optimization.The flow on top is the traditional search-based mathematical optimization method. The bottom is the machine learning based method we propose. Three additional steps are inserted to learn a refined and reduced search space.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4477370&req=5

f2: Framework of material structure optimization.The flow on top is the traditional search-based mathematical optimization method. The bottom is the machine learning based method we propose. Three additional steps are inserted to learn a refined and reduced search space.

Mentions: Herein, we propose the employment of modern machine learning (ML) techniques as a tool to explore multiple design solutions and diminished searching time in high dimensional microstructure design problems, where the number of distinct design candidates is indeed infinite. Two crucial ML steps, namely, search path refinement and search space reduction, are designed to develop heuristics that tour the search force to a much smaller preferable space. As the diagram in Fig. 2 suggests, the ML method (bottom route) has these two steps (marked as 2 and 3) executed laterally, after a data preparation step (marked as 1) that precedes. The three steps supplement a traditional direct-search method (top route) by performing a search space preprocessing, before the actual search goes into action. Such a ML-based preprocessing is designed to locate critical regions of a search space with a small overhead, so that the search force can be consciously concentrated.


A predictive machine learning approach for microstructure optimization and materials design.

Liu R, Kumar A, Chen Z, Agrawal A, Sundararaghavan V, Choudhary A - Sci Rep (2015)

Framework of material structure optimization.The flow on top is the traditional search-based mathematical optimization method. The bottom is the machine learning based method we propose. Three additional steps are inserted to learn a refined and reduced search space.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4477370&req=5

f2: Framework of material structure optimization.The flow on top is the traditional search-based mathematical optimization method. The bottom is the machine learning based method we propose. Three additional steps are inserted to learn a refined and reduced search space.
Mentions: Herein, we propose the employment of modern machine learning (ML) techniques as a tool to explore multiple design solutions and diminished searching time in high dimensional microstructure design problems, where the number of distinct design candidates is indeed infinite. Two crucial ML steps, namely, search path refinement and search space reduction, are designed to develop heuristics that tour the search force to a much smaller preferable space. As the diagram in Fig. 2 suggests, the ML method (bottom route) has these two steps (marked as 2 and 3) executed laterally, after a data preparation step (marked as 1) that precedes. The three steps supplement a traditional direct-search method (top route) by performing a search space preprocessing, before the actual search goes into action. Such a ML-based preprocessing is designed to locate critical regions of a search space with a small overhead, so that the search force can be consciously concentrated.

Bottom Line: These challenges render traditional search-based optimization methods incompetent in terms of both searching efficiency and result optimality.A systematic framework consisting of random data generation, feature selection and classification algorithms is developed.Experiments with five design problems that involve identification of microstructures that satisfy both linear and nonlinear property constraints show that our framework outperforms traditional optimization methods with the average running time reduced by as much as 80% and with optimality that would not be achieved otherwise.

View Article: PubMed Central - PubMed

Affiliation: EECS Department, Northwestern University, Evanston IL, USA.

ABSTRACT
This paper addresses an important materials engineering question: How can one identify the complete space (or as much of it as possible) of microstructures that are theoretically predicted to yield the desired combination of properties demanded by a selected application? We present a problem involving design of magnetoelastic Fe-Ga alloy microstructure for enhanced elastic, plastic and magnetostrictive properties. While theoretical models for computing properties given the microstructure are known for this alloy, inversion of these relationships to obtain microstructures that lead to desired properties is challenging, primarily due to the high dimensionality of microstructure space, multi-objective design requirement and non-uniqueness of solutions. These challenges render traditional search-based optimization methods incompetent in terms of both searching efficiency and result optimality. In this paper, a route to address these challenges using a machine learning methodology is proposed. A systematic framework consisting of random data generation, feature selection and classification algorithms is developed. Experiments with five design problems that involve identification of microstructures that satisfy both linear and nonlinear property constraints show that our framework outperforms traditional optimization methods with the average running time reduced by as much as 80% and with optimality that would not be achieved otherwise.

No MeSH data available.


Related in: MedlinePlus