Simulink Control Design™ - MathWorks

Here, x, u, and y are vectors of mapped states, inputs, and outputs, respectively. These

vectors contain the mapped values specified in indx, indu, and indy, respectively.

Add the updated custom functions to the operating point specification.

opspec.CustomConstrFcn = @myConstraintsMap;

opspec.CustomObjFcn = @myObjectiveMap;

Trim the model using the custom mapping, and view the trimmed states, which match the

previous results in op2.

[op3,rpt3] = findop(mdl,opspec,opt);

op3.States

(1.) scdTanks/Inertia

x: 0

(2.) scdTanks/Tank1

x: 15

(3.) scdTanks/Tank2

x: 15

(4.) scdTanks/Tank3

x: 16

Add Analytic Gradients to Custom Functions

For faster or more reliable computations, you can add analytic gradients to your custom

constraint and objective functions. Adding gradients can reduce the number of function

calls during optimization and potentially improve the accuracy of the optimization result.

If you specify gradients, you must specify them for both the custom constraint and

objective functions. (Gradients for custom trimming are not supported for Simscape™

models.)

To define the gradient of a given constraint or objective function, take the derivative of

the function with respect to a given state, input, or output. For example, if the objective

function is

F = (u(1)+3)^2 + y(1)^2

then the gradient of F with respect to u(1) is

G = 2*(u(1)+3)

To add gradients to your custom constraint function, specify the following additional

output arguments:

1 Steady-State Operating Points