A graph algorithm for linearizing simulink models

Authors:
Zhi Han;Pieter J. Mosterman;Fu Zhang
Affiliations:
MathWorks, Inc., Natick, MA;MathWorks, Inc., Natick, MA;MathWorks, Inc., Natick, MA
Venue:
Proceedings of the 2013 Summer Computer Simulation Conference
Year:
2013

Citing 6
Cited 0

The semantics and execution of a synchronous block-diagram language

Science of Computer Programming
Optimal accumulation of Jacobian matrices by elimination methods on the dual computational graph

Mathematical Programming: Series A and B
Optimal Jacobian accumulation is NP-complete

Mathematical Programming: Series A and B
Evaluating Derivatives: Principles and Techniques of Algorithmic Differentiation

Evaluating Derivatives: Principles and Techniques of Algorithmic Differentiation
Model-Based Design for Embedded Systems

Model-Based Design for Embedded Systems
Towards sensitivity analysis of hybrid systems using simulink

Proceedings of the 16th international conference on Hybrid systems: computation and control

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper presents a new efficient approach for performing linearization of Simulink® models. It improves the efficiency of existing linearization algorithms using a Jacobian graph, a graph-based data structure that captures the linear relationship between input, output and state variables. The graph-based algorithm enables the use of graph transformations to reduce the size of the Jacobian data structure, thereby improving the efficiency of subsequent computations. This paper presents a heuristic implementation of the graph-based algorithm. Experimental results on a number of Simulink models of different sizes show how the approach is able to significantly improve computational efficiency and memory use especially in models with large numbers of blocks and states.