arXiv:2611.14549v1 Announce Type: new Abstract: A numerical algorithm is described for computing the inverse of a non-singular multivector across dimensions. The method wraps a reversion-like product around a Faddeev--LeVerrier--Souriau (FLS) polynomial core. While FLS provides dimensional extensibility, the wrapper reduces the polynomial order of the FLS method, and introduces a sparse multivector which can be further exploited. The construction and essential properties of the initial involution are explained. For dimensions 4--7, the optimized inverse requires fewer real multiplications than a same-dimension multivector product. Dimensions 8--11 follow easily on common hardware with similar efficiency. For dimensions 12--15, higher-precision arithmetic demonstrates the generality of the algorithm when the arithmetic can support it. Performance on common hardware is reported. Directions are included for obtaining a Python package with examples for numerical computation in Clifford algebra, supporting arbitrary signatures including degenerate cases.
Continue to arXiv Mathematics →You are being forwarded in 6s — stay on this page