The paper presents the advantages of using the product form of inverses of sparse matrices (PFI) in the optimization processes of the working states in power systems. It also presents ways, tested by the author, of determining the PFI. Two ways of pivoting are compared, both from the point of view of conserving the sparsity of the represented matrix - usually the Jacobi matrix of the equality type restrictions - and from the point of view of the calculus duration. In its latter part the paper presents a few methods of performing the calculations with the product form matrices.