MATLAB and Simulink See Easier Big Data Crunching and Design Space Exploration
Shawn Wasserman posted on September 29, 2016 |
Tall array visualization of Big Data. Engineers can now crunch data in MATLAB and Simulink that is too big to fit into memory. (Image courtesy of MathWorks.)

Tall array visualization of Big Data. Engineers can now crunch data in MATLAB and Simulink that is too big to fit into memory. (Image courtesy of MathWorks.)

MathWorks has released its 2016b release of MATLAB and Simulink. Since the last release in March, the team at MathWorks has been focusing on design space exploration and simplifying how engineers crunch big data.

Namely, the release will give engineers the ability to crunch big data even if it is too big to fit into memory.

“Companies are awash in data but struggle to take advantage of it to build better predictive models and gain deeper insights,” said David Rich, MATLAB marketing director at MathWorks. “With R2016b, we’ve lowered the bar to allow domain experts to work with more data, more easily. This leads to improved system design, performance and reliability.”

Using a tall array, engineers will be able to work with out-of-memory data using functions and syntax familiar to MATLAB users. Therefore, engineers will no longer need to learn how to program around the big data. These tall arrays will also be compatible with math, statistics and machine learning algorithms within MATLAB.

The code created by the engineer will be able to run on Hadoop clusters or can be integrated into Spark applications.

Other data improvements to MATLAB include:

  • Timetable data indexing and synchronizing
  • String arrays
  • Top chart shows how the cost function changes based on the model parameters. Bottom chart depicts a tornado plot to show which parameters have the most influence on a cost function. (Image courtesy of MathWorks.)

    Top chart shows how the cost function changes based on the model parameters. Bottom chart depicts a tornado plot to show which parameters have the most influence on a cost function. (Image courtesy of MathWorks.)

    New data preprocessing functions

As for Simulink, MathWorks has released tools to improve design space exploration.

The new feature will allow engineers to perform design of experiments (DoE) and Monte Carlo simulations to help optimize a product's design.

“Growing design complexity is creating increasingly large models. To maintain model accuracy, engineers are challenged with identifying which model parameters impact behavior the most,” said Paul Barnard, design automation director atMathWorks. “Now, engineers can use Simulink Design Optimization to determine model sensitivity, fit the model to test data and tune it to meet requirements.”

Using Simulink Design Optimization, engineers can determine how the cost function will be affected by the changing parameters and then plot that information into a tornado diagram to predict which parameters have the most effect on the design.

For more on big data and MATLAB, read “Creating Predictive Algorithms for HVAC Optimization in MATLAB.”


Recommended For You