Abstract An irreducible algebraic decomposition ∪ i = 0 d X i = ∪ i = 0 d ( ∪ j = 1 d i X ij ) $ \cup _{i = 0}^d X_i = \cup _{i = 0}^d ({\cup _{j = 1}^{d_i } X_{ij} } )$ of an affine algebraic variety X can be represented as a union of finite disjoint sets ∪ i = 0 d W i = ∪ i = 0 d ( ∪ j = 1 d i W ij ) $\cup _{i = 0}^d W_i = \cup _{i = 0}^d ({\cup _{j = 1}^{d_i } W_{ij} } )$ called numerical irreducible decomposition (cf. [14],[15],[18],[19],[20],[22],[23],[24]). The Wi correspond to the pure i-dimensional components Xi , and the Wij present the i-dimensional irreducible components Xij . The numerical irreducible decomposition is implemented in Bertini (cf. [3]). The algorithms use homotopy continuation methods. We modify this concept using partially Gröbner bases, triangular sets, local dimension, and the so-called zero sum relation. We present in this paper the corresponding algorithms and their implementations in Singular (cf. [8]). We give some examples and timings, which show that the modified algorithms are more efficient if the number of variables is not too large. For a large number of variables Bertini is more efficient*.