optimization_gradient_descent

optimization_gradient_descent, looks at the use of the gradient descent method to minimize a function of one or several variables.

The notes:

• optimization_gradient_descent.pdf
• optimization_gradient_descent.tex

Scripts and functions:

• gradient_descent1.m, gradient descent for a function of one variable.
• gradient_descent2.m, gradient descent for a function of multiple variables.
• hex2.m, a test function of two variables.
• hex2_df.m, the gradient of hex2().
• hex2_gradient2.m, seeks a minimizer of hex2() using gradient_descent2().
• quartic.m, a test function of one variable.
• quartic_df.m, the derivative of quartic().
• quartic_gradient1.m, seeks a minimizer of quartic() using gradient_descent1().
• rosenbrock.m, a test function of two variables.
• rosenbrock_df.m, the gradient of rosenbrock().
• rosenbrock_gradient2.m, seeks a minimizer of rosenbrock() using gradient_descent2().
• sag.m, a test function of one variable.
• sag_df.m, the derivative of sag().
• sag_gradient1.m, seeks a minimizer of sag() using gradient_descent1().

Text files:

• ford_data.txt
• gradient_descent1_pseudocode.txt
• gradient_descent2_pseudocode.txt
• quartic_gradient1.txt
• sag_gradient1.txt

Images:

• ford_formula.png
• hex2.png
• quartic.png
• quartic_minimizer.png
• rosenbrock.png
• sag.png
• sag_minimizer.png
• skater.png