Matrix and Vector Operators

The ROOT::Math::SVector and ROOT::Math::SMatrix classes defines the following operators described below. The m1,m2,m3 are vectors or matrices of the same type (and size) and a is a scalar value:

m1 == m2           _// returns whether m1 is equal to m2 (element by element comparison)_ 
m1 != m2           _// returns whether m1 is NOT equal to m2 (element by element comparison)_ 
m1 < m2            _// returns whether m1 is less than m2 (element wise comparison)_
m1 > m2            _// returns whether m1 is greater than m2 (element wise comparison)_
_// in the following m1 and m3 can be general and m2 symmetric, but not vice-versa_
m1 += m2           _// add m2 to m1_
m1 -= m2           _// subtract m2 to m1_ 
m3 = m1 + m2       _// addition_ 
m1 - m2            _// subtraction_

_// Multiplication and division via a scalar value a_
m3 = a*m1; m3 = m1*a; m3 = m1/a;

Vector-Vector multiplication

The operator * defines an element by element multiplication between vectors. For the standard vector-vector multiplication, \( a = v^T v \), (dot product) one must use the ROOT::Math::Dot function. In addition, the Cross (only for vector sizes of 3), ROOT::Math::Cross, and the Tensor product, ROOT::Math::TensorProd, are defined.

Matrix - Vector multiplication

The operator * defines the matrix-vector multiplication, \( y_i = \sum_{j} M_{ij} x_j\):

_// M is a  N1xN2 matrix, x is a N2 size vector, y is a N1 size vector_
y = M * x

It compiles only if the matrix and the vectors have the right sizes. Matrix - Matrix multiplication The operator * defines the matrix-matrix multiplication, \( C_{ij} = \sum_{k} A_{ik} B_{kj}\):

_// A is a N1xN2 matrix, B is a N2xN3 matrix and C is a N1xN3 matrix_
C = A * B

The operation compiles only if the matrices have the right size. In the case that A and B are symmetric matrices, C is a general one, since their product is not guaranteed to be symmetric.

Matrix and Vector Functions

The most used matrix functions are:

ROOT::Math::Transpose(M) : return the transpose matrix, \( M^T \)
ROOT::Math::Similarity( v, M) : returns the scalar value resulting from the matrix- vector product \( v^T M v \)
ROOT::Math::Similarity( U, M) : returns the matrix resulting from the product \( U M U^T \). If M is symmetric, the returned resulting matrix is also symmetric
ROOT::Math::SimilarityT( U, M) : returns the matrix resulting from the product \( U^T M U \). If M is symmetric, the returned resulting matrix is also symmetric

See Matrix Template Functions for the documentation of all existing matrix functions in the package. The major Vector functions are:

ROOT::Math::Dot( v1, v2) : returns the scalar value resulting from the vector dot product
ROOT::Math::Cross( v1, v2) : returns the vector cross product for two vectors of size 3. Note that the Cross product is not defined for other vector sizes
ROOT::Math::Unit( v) : returns unit vector. One can use also the v.Unit() method.
ROOT::Math::TensorProd(v1,v2) : returns a general matrix M of size N1xN2 resulting from the Tensor Product between the vector v1 of size N1) and v2 of size N2

See Vector Template Functions for the list and documentation of all of them.

Matrix and Vector I/O

One can print (or write in an output stream) Vectors (and also Matrices) using the Print method or the << operator, like:

v.Print(std::cout);
std::cout << v << std::endl;

In the ROOT distribution, the CINT dictionary is generated for SMatrix and SVector for double types and sizes up to 5. This allows the storage of them in a ROOT file.

*/