Parallel Computing in Global Optimization Toolbox

R2026b

Parallel processing can speed optimization algorithms by distributing function evaluations across multiple processors. To use parallel processing, you must have a Parallel Computing Toolbox™ license.

To set up parallel computing, open a parallel pool:

parpool

Note

Depending on your preferences, MATLAB^® can start a parallel pool automatically. To enable this feature, select Parallel > Parallel Settings... in the Environment group on the Home tab, and then select Automatically create a parallel pool.

Then set the UseParallel option to "auto" or "on" to enable parallel computing for your solver.

Solver	Command-Line Settings
ga	options = optimoptions("ga",UseParallel="auto", UseVectorized=false);
gamultiobj	options = optimoptions("gamultiobj",UseParallel="auto", UseVectorized=false);
MultiStart	ms = MultiStart(UseParallel="auto"); or ms.UseParallel = "auto"
paretosearch	options = optimoptions("paretosearch",UseParallel="auto");
particleswarm	options = optimoptions("particleswarm",UseParallel="auto", UseVectorized=false);
patternsearch	options = optimoptions("patternsearch",UseParallel="auto", UseVectorized=false, UseCompletePoll=true);
surrogateopt	options = optimoptions("surrogateopt",UseParallel="auto");

Global Optimization Toolbox solvers use parallel computing in different ways depending on the solver. For details, see the solver-specific pages:

• MultiStart — MultiStart distributes start points to multiple processors. GlobalSearch does not compute in parallel directly, but the fmincon local solver can estimate gradients in parallel.

• Parallel Computing with ga, gamultiobj, particleswarm, and paretosearch — These solvers evaluate populations or poll points in parallel.

• Parallel Computing with patternsearch — patternsearch evaluates poll points and search functions in parallel, with specific option requirements.

• Parallel Computing with surrogateopt — surrogateopt evaluates search points on workers asynchronously using parfeval.

• simulannealbnd does not run in parallel directly, but can call hybrid functions that take advantage of parallel computing. See Parallel Hybrid Functions.

Note

The documentation recommends not to use parfor or parfeval when calling Simulink^®; see Using sim Function Within parfor (Simulink). Therefore, you might encounter issues when optimizing a Simulink simulation in parallel using a solver's built-in parallel functionality. For an example showing how to optimize a Simulink model with several Global Optimization Toolbox solvers, see Optimize Simulink Model in Parallel.