**About this Abstract** |

**Meeting** |
**2018 TMS Annual Meeting & Exhibition
** |

**Symposium
** |
**Hume-Rothery Award Symposium: Computational Thermodynamics and Its Implications to Kinetics, Properties, and Materials Design
** |

**Presentation Title** |
Alloy Design as the Solution to a Continuous Constraint Satisfaction Problem |

**Author(s)** |
Raymundo Arroyave, Anas Abu-Odeh, Tann Kirk, Edgar Galvan, Richard Malak |

**On-Site Speaker (Planned)** |
Raymundo Arroyave |

**Abstract Scope** |
In general, the forward phase stability problem consists of the mapping of thermodynamic conditions to their
corresponding equilibrium states. This ‘forward’ framework is resolved through the minimization of the Gibbs energy of the system subject to constraints. In this talk, we instead focus on the generalized inverse phase stability problem (GIPSP), defined as the mapping of sets of phase constitution definitions in multi-dimensional phase constitution search space to the corresponding ranges of values of thermodynamic conditions. Mathematically, the solution to the GIPSP corresponds to all the solutions to a Continuous Constraint Satisfaction Problem (CCSP). In this talk, we present novel algorithms combining computational thermodynamics, evolutionary computation and machine learning to approximate the solution sets to the GIPSP as a CCSP. Moreover, we show how conventional alloy design prescriptions can be mapped to (non-linear) constraints that can in turn be solved through the algorithms developed in this work. |

**Proceedings Inclusion?** |
Planned: Supplemental Proceedings volume |