Expand description
The asymptotic numerical method (ANM): the Taylor series z(h) = sum_k z_k h^k of a solution branch of F(z, p0 + h) = 0 in the problem’s parameter, from a solution z_0 at p0 (Cochelin, Comput. Struct. 53 (1994) 1181; Cochelin, Damil and Potier-Ferry, Int. J. Numer. Methods Eng. 37 (1994) 1187).
Order k of F(z(h), p0 + h) is J z_k plus what the lower orders give, with J the Jacobian at z_0, so each
coefficient solves J z_k = -F_k, where F_k is order k of the residual’s series with z_k left out, taken one order
at a time from the problem’s ResidualJet::jet_orders. One factorization of J serves every order. With refine > 0 each solve is refined by the exact
products J x (Jvp): the residual b - J x, solved with the same factors, corrects x.
The series is valid up to about Branch::radius, Cochelin’s estimate (delta |z_1| / |z_N|)^(1 / (N - 1)), where
the last term would be delta times the first.
Structs§
- Branch
- The Taylor coefficients z_0..z_N of a branch, and for each order k >= 1 the relative residual |J z_k + F_k| / |F_k| of its solve after refinement (max norms, NaN where either holds a NaN, and NaN without refinement)
Functions§
- series
- The branch’s coefficients z_1..z_order at the solution z0 (P::PARTS part arrays), the module’s documentation,
each order’s solve refined
refinetimes.