CMAEvolutionStrategyOptimizer.UseDiagonalCovarianceMatrix Property

public bool UseDiagonalCovarianceMatrix { get; set; }

Public Property UseDiagonalCovarianceMatrix As Boolean
	Get
	Set

public:
property bool UseDiagonalCovarianceMatrix {
	bool get ();
	void set (bool value);
}

member UseDiagonalCovarianceMatrix : bool with get, set

Property Value

The UseDiagonalCovarianceMatrix property determines whether the optimizer uses a diagonal covariance matrix instead of a full covariance matrix. When set to true, the optimizer will only consider the variances of the individual dimensions and ignore the covariances between different dimensions. This can significantly reduce the computational complexity and memory usage, especially for high-dimensional optimization problems.

Using a diagonal covariance matrix can be beneficial when the problem is known to have independent dimensions or when computational resources are limited. However, it may lead to suboptimal performance if the dimensions of the problem are correlated, as the optimizer will not be able to capture and exploit these correlations.

It is important to carefully consider the nature of the optimization problem and the available computational resources when deciding whether to use a diagonal covariance matrix.

Reference