Use Optimization Explorer App

Video Walkthrough

For a walkthrough of the example, play the video.

Create Problem

To use the Optimization Explorer app, you must have an existing problem, either an OptimizationProblem object or the inputs for an optimization solver, such as fmincon or patternsearch. You can get a problem into your workspace by running the Optimize Live Editor task or by creating a problem at the command line. For this example, run the following script to get the prob problem into your workspace.

x = optimvar("x",LowerBound=-100,UpperBound=99);
y = optimvar("y",LowerBound=-98,UpperBound=102);
rng default % For reproducibility
offset = 30*randn(1,2); % Move optimal solution from [0,0]
f2 = @(x,y)sawtoothxy(x-offset(1),y-offset(2));
f = fcn2optimexpr(f2,x,y);
prob = optimproblem(Objective=f);
x0.x = 0;
x0.y = 0;

function f = sawtoothxy(x,y)
[t,r] = cart2pol(x,y); % Change to polar coordinates
h = cos(2*t - 1/2)/2 + cos(t) + 2;
g = (sin(r) - sin(2*r)/2 + sin(3*r)/3 - sin(4*r)/4 + 4) ...
    .*r.^2./(r+1);
f = g.*h;
end

To ensure that a solver can run the problem without error, run a solver on the problem at the command line.

[sol,fval] = solve(prob,x0)

Solving problem using fmincon.

Local minimum found that satisfies the constraints.

Optimization completed because the objective function is non-decreasing in 
feasible directions, to within the value of the optimality tolerance,
and constraints are satisfied to within the value of the constraint tolerance.

sol = struct with fields:
    x: 9.7505
    y: 62.7420

fval = 23.5804

Open Optimization Explorer

Open the app by clicking its icon in the Apps gallery or by entering optimizationExplorer at the command line.

Choose Problem and Attributes

In this case, the problem is an OptimizationProblem object. Import the problem object.

Select prob.

Optimization Explorer displays the problem characteristics.

Specify the Starting point as x0.

Click Next. Specify the problem attributes if you know them. These attributes help the app to choose appropriate solvers. For this problem, select No for each attribute.

Click Next.

Run Optimization Explorer

Specify to run the app in auto mode, which means the app runs all recommended solvers, start points, and options automatically, up to the time limit you set. In this case, set Time limit to 2 minutes. If you have Parallel Computing Toolbox™, select Use parallel so the app runs in parallel, which is often faster than running in serial. In parallel, the app generally runs several solvers simultaneously.

Click Start Session.

Examine Results

Examine the results displayed in the app. First, sort the Results Table in ascending order by objective function value by clicking on the Objective value column header in the Results Table.

Many of the table rows have the objective function value of approximately 0.

You can apply filters to focus on specific results. To do so, click the Show Filters button in the Review Results section of the toolstrip. The Results Filters panel opens, displaying options you can use to filter the results that appear in the plots and the Results Table.

To view only the top 10 results, first show all results by clicking the reset button at the top of the Results Filters pane. Then click Show leaderboard at the top of the Results Table.

To view the solution for a specific configuration, double-click the entry in the Solution column for that configuration.

You can examine the details of the configuration that is currently selected in the Results Table. Click the three vertical dots on the right side of the table for the configuration, and select Show details.

A tab opens and displays details for the configuration, the iterative display, and the reason the solver stopped. The tab also provides links to related topics to help you interpret the results.

Remove the leaderboard filtering by clicking the reset button.

You can choose which columns of data appear in the table.

Add or delete table columns.

Rerun Optimization

Select some solvers to rerun manually. You can reschedule the configuration that is currently selected in the Results Table. Click the three vertical dots on the right side of the table for the configuration, and select Reschedule.

Choose other solvers to run by selecting them from the Select Solvers section of the toolstrip. The app must be in manual mode for you to select solvers. To switch to manual mode, select Manual in the Select Solvers section. As you select solvers to run, they appear as tabs in the main window.

The Solver Configurations panel shows the same scheduled configurations.

After selecting the solver configurations you want to run, click Run in the toolstrip.

Export Results

You can export the entire Results Table or a function that creates the currently selected result in the table.

The exported table contains only the data in the current filter, and has all possible columns and additional results details.

An exported function does not include your optimization problem or solver inputs. To use an exported function, save the problem separately or save the solver inputs.

For example, if you started Optimization Explorer from an OptimizationProblem object, the exported function code begins:

function [solution,objectiveValue] = solveProblem(problem)

This exported function requires you to give the original OptimizationProblem object as the problem argument.

If you started Optimization Explorer from solver inputs, the exported function code begins with the following or equivalent:

function [solution,objectiveValue] = solveProblem(fun,A,b,Aeq,beq,lb,ub,nonlcon)
% This function accepts the problem inputs imported into the Optimization
% Explorer app and reproduces the selected result.

% Auto-generated by MATLAB on 24-Oct-2025 15:57:35

% Load data
data = load("explorerConfiguration1_1FunctionData1.mat");
x0 = data.x0;

This exported function requires you to give the original objective function fun, the linear constraint matrices A, b, Aeq, and beq, the bounds lb and ub, and the nonlinear constraint function nonlcon. The original start point x0 is saved in the Optimization Explorer session data, which you also need to have in your file system in order to run the exported function.