Numerical Optimization

SCILAB training course for optimization

You might not be familiar with the numerical concept of optimization, but I am pretty sure your daily work requires you to:

  • Build something that is more and more efficient / Increase your profits
  • While limiting the production & operation costs / Limiting the associated risk

Managers
Operational assistance
Maximize efficiency, lower costs		 Traders
Decision support system
Maximize revenue, lower risks		 Engineers
Multidisciplinary & robust design optimization / Model Predictive Control

If you are still reading, I can say: I told you! And what you are currently facing is called an optimization problem: It is all about trying to maximize one or many objectives under given constraints.

Training objective

Provide you with the basics to be able to:

  1. Recognize what in your daily work is an optimization problem
  2. Understand some basics on optimization theory and its numerical implementation
  3. Qualify your porblem to choose the most efficient method to solve it
  4. Discover different methods implemented in Scilab and make efficient use of it

Training content

What is optimization used for
Introduction to optimization problem in real life
  • Optimal control
  • Optimal design
  • Inverse problem

Numerical setting
Introduction to optimization problem in mathematics & numeric
  • Optimization problem formulation
  • Equivalent formulation (e.g. Dual)
  • Classic solving process
  • Different solving methods (direct search vs differential calculus)
Problem qualification
Introduction to problem segmentation and problematic of solver efficiency
  • Discrete/Continuous
  • Simple or multi-objective
  • (Non)Smooth
  • (Un)Constrained
  • (Non)Linear
Algorithms in SCILAB
Detailed description of each kind of problem with classic use-cases, and dedicated SCILAB function to solve it. Exercises are provide to learn how to handle those functions
  • Linear
  • Quadratic
  • Semi definite
  • Non-linear least squares
  • Non-linear
  • Multi-objective
  • Discrete
  • Non-smooth
Robust optimization
Introduction to the reliability issue and need for robust optimization
  • Sensitivity analysis
  • Reliability analysis
  • Robust design optimization

Training duration

Depending on your mathematical level:

  • Beginner (3 days)
    We will take 2 days for the theoritical content to be sure you master the basis and one day to explore the exercises
  • Intermediate (2 to 3 days)
    We can take only 1 day for the theoritical content, focus on the exercise on another day. If you are interested in more dedicated Scilab experience, you can also provide us with one of your use case that we will investigate together on one more day.

FOR MORE INFORMATION ON SESSION AND PRICING