Consider the problem of finding a minimum of the function:

Plot the function to get an idea of where it is minimized

The plot shows that the minimum is near the point (-1/2,0).

Usually you define the objective function as a MATLAB file. For now, this function is simple enough to define as an anonymous function:

Take a guess at the solution:

Set optimization options to not use fminunc's default large-scale algorithm, since that algorithm requires the objective function gradient to be provided:

View the iterations as the solver calculates:

Call fminunc, an unconstrained nonlinear minimizer:

The solver found a solution at:

The function value at the solution is:

We will use the number of function evaluations as a measure of efficiency in this example. The total number of function evaluations is:

We will now pass extra parameters as additional arguments to the objective function. We show two different ways of doing this - using a MATLAB file, or using a nested function.

Consider the objective function from the previous section:

We parameterize the function with (a,b,c) in the following way:

This function is a shifted and scaled version of the original objective function.

Method 1: MATLAB file Function

Suppose we have a MATLAB file objective function called bowlpeakfun defined as:

Define the parameters:

Create an anonymous function handle to the MATLAB file:

Call fminunc to find the minimum:

Method 2: Nested Function

Consider the following function that implements the objective as a nested function

In this method, the parameters (a,b,c) are visible to the nested objective function called nestedfun. The outer function, nestedbowlpeak, calls fminunc and passes the objective function, nestedfun.

Define the parameters, initial guess, and options:

Run the optimization:

You can see both methods produced identical answers, so use whichever one you find most convenient.

Consider the above problem with a constraint:

The constraint set is the interior of a tilted ellipse. Look at the contours of the objective function plotted together with the tilted ellipse

The plot shows that the lowest value of the objective function within the ellipse occurs near the lower right part of the ellipse. We are about to calculate the minimum that was just plotted. Take a guess at the solution:

Set optimization options: use the interior-point algorithm, and turn on the display of results at each iteration:

Solvers require that nonlinear constraint functions give two outputs: one for nonlinear inequalities, the second for nonlinear equalities. So we write the constraint using the dealfunction to give both outputs:

Call the nonlinear constrained solver. There are no linear equalities or inequalities or bounds, so pass [ ] for those arguments:

A solution to this problem has been found at:

The function value at the solution is:

The total number of function evaluations was:

The inequality constraint is satisfied at the solution.

Since c(x) is close to 0, the constraint is "active," meaning the constraint affects the solution. Recall the unconstrained solution was found at

and the unconstrained objective function was found to be

The constraint moved the solution, and increased the objective by

Optimization problems can be solved more efficiently and accurately if gradients are supplied by the user. This example shows how this may be performed. We again solve the inequality-constrained problem

To provide the gradient of f(x) to fmincon, we write the objective function in the form of a MATLAB file:

The constraint and its gradient are contained in the MATLAB file tiltellipse:

Make a guess at the solution:

Set optimization options: we continue to use the same algorithm for comparison purposes.

We also set options to use the gradient information in the objective and constraint functions. Note: these options MUST be turned on or the gradient information will be ignored.

There should be fewer function counts this time, since fmincon does not need to estimate gradients using finite differences.

Call the solver:

fmincon estimated gradients well in the previous example, so the iterations in the current example are similar.

The solution to this problem has been found at:

The function value at the solution is:

The total number of function evaluations was:

Compare this to the number of function evaluations without gradients:

This time we solve the same constrained problem

more accurately by overriding the default termination criteria (options.StepTolerance and options.OptimalityTolerance). We continue to use gradients. The default values for fmincon's interior-point algorithm are options.StepTolerance = 1e-10, options.OptimalityTolerance = 1e-6.

Override two default termination criteria: termination tolerances on X and fval.

Call the solver:

We now choose to see more decimals in the solution, in order to see more accurately the difference that the new tolerances make.

The optimizer found a solution at:

Compare this to the previous value:

The change is

The function value at the solution is:

The solution improved by

(this is negative since the new solution is smaller)

The total number of function evaluations was:

Compare this to the number of function evaluations when the problem is solved with user-provided gradients but with the default tolerances:

If you give not only a gradient, but also a Hessian, solvers are even more accurate and efficient.

fmincon's interior-point solver takes a Hessian matrix as a separate function (not part of the objective function). The Hessian function H(x,lambda) should evaluate the Hessian of the Lagrangian; see the User's Guide for the definition of this term.

Solvers calculate the values lambda.ineqnonlin and lambda.eqlin; your Hessian function tells solvers how to use these values.

In this problem we have but one inequality constraint, so the Hessian is:

In order to use the Hessian, you need to set options appropriately:

The tolerances have been set back to the defaults. There should be fewer function counts this time.

Call the solver:

There were fewer, and different iterations this time.

The solution to this problem has been found at:

The function value at the solution is:

The total number of function evaluations was:

Compare this to the number using only gradient evaluations, with the same default tolerances: