The Principia Machina/Tools/ModelingToolkit: Difference between revisions
m 1 revision imported |
DeltaS Zero (talk | contribs) Formulas written for the wiki's math renderer: 0 $$..$$ and 3 $..$ as <math>; no other text changed (tpm_tools 0.1.0) |
||
| Line 6: | Line 6: | ||
== Mathematical Evaluation for TPM & SVD == | == Mathematical Evaluation for TPM & SVD == | ||
* '''Euler-Lagrange Derivation & DAE Compilation (9.5/10):''' MTK accepts symbolic Lagrangian actions | * '''Euler-Lagrange Derivation & DAE Compilation (9.5/10):''' MTK accepts symbolic Lagrangian actions <math>L(q, \dot{q}, t)</math> and natively computes symbolic partial derivatives (<math>\frac{\partial L}{\partial \dot{q}}</math>, <math>\frac{\partial L}{\partial q}</math>) to generate full Euler-Lagrange DAEs automatically before performing Pantelides index reduction. | ||
* '''Structural Simplification:''' Automatically eliminates redundant algebraic constraints and simplifies complex multi-body systems prior to numerical code generation. | * '''Structural Simplification:''' Automatically eliminates redundant algebraic constraints and simplifies complex multi-body systems prior to numerical code generation. | ||
* '''Composable Scientific Machine Learning (SciML):''' Integrates natively with neural ODEs and parameter estimation pipelines. | * '''Composable Scientific Machine Learning (SciML):''' Integrates natively with neural ODEs and parameter estimation pipelines. | ||
Revision as of 00:59, 6 October 2026
ModelingToolkit.jl: Symbolic-Numeric Acausal DAE Optimization
Overview & Architectural Fit
ModelingToolkit.jl (MTK) is a symbolic-numeric composable modeling framework written in Julia. It acts as an advanced, next-generation acausal compiler that directly addresses the symbolic limitations of traditional Modelica.
Mathematical Evaluation for TPM & SVD
- Euler-Lagrange Derivation & DAE Compilation (9.5/10): MTK accepts symbolic Lagrangian actions and natively computes symbolic partial derivatives (, ) to generate full Euler-Lagrange DAEs automatically before performing Pantelides index reduction.
- Structural Simplification: Automatically eliminates redundant algebraic constraints and simplifies complex multi-body systems prior to numerical code generation.
- Composable Scientific Machine Learning (SciML): Integrates natively with neural ODEs and parameter estimation pipelines.