Examples

These examples illustrate the main features and applications of the NEXTorch.

Specifically these tutorials fall into three categories:

• Basic API Usage

• Applications in Reaction Engineering

• Mixed Type Parameters

• Multi-Objective Optimization(MOO)

Basic API Usage

Basic usage of the NEXTorch API

• Example 1 and 2: Implementation of NEXTorch for single-objective optimization

• Example 12 and 13: different DOE methods and comparing NEXTorch to Ax syntax

Example 1 - Simple 1d nonlinear function

Example 2 - Sin(x) 1d function

Example 12 - Using different DOE methods for PFR yield optimization

Example 13 - Using AX for plug flow reactor yield optimization

Applications in Reaction Engineering

Applications of NEXTorch in the real-world problems, such as catalyst synthesis, reaction condition optimizations, and reactor design

• Example 3, 5, and 8: Automated optimization loop for computational simulations

• Example 4: Human-in-the-loop optimization for laboratory experiments

Example 3 - Langmuir Hinshelwood mechanism

Example 4 - Nitrogen-doped carbon catalysts

Example 5 - Plug flow reactor yield

Example 8 - Stub tuner of the microwave cavity

Mixed Type Parameters

Demo for systems with parameters of mixed types

Example 10 - Plug flow reactor yield with mixed type inputs

Multi-Objective Optimization(MOO)

Demo for MOO systems

• Example 6, 7 and 9: MOO with the weighted sum method

• Example 11: MOO with EHVI

Example 6 - Multi-objective optimization for an ellipse function

Example 7 - Multi-objective optimization for plug flow reactor

Example 9 - Multi-objective optimization for Microwave operating conditions

Example 11 - Multi-objective optimization for plug flow reactor