0^j+ ddlmZmZmZmZmZmZddlmZddl m Z ddl m Z ddl mZdgZe d dd d d d d d d ddZy ))innervdot iscomplexobjzerosinffinfo)norm)sqrt)xp_capabilities) make_systemminresT)np_onlyNgh㈵>gF)rtolshiftmaxiterMcallbackshowcheckc t||||\}}} }t|rtnt} |j} |j} d}d}|j d}|d|z}gd}|rDt |dzt |d|dd |d zt |d |dd |d zt d}d}d}d}d}d}| j}t|j}||j}n||| zz }| |}| ||}|dkr td|dk(r| dfSt|}|dk(r|} | dfSt|j}| r| |}| |}| ||} | ||}!t| |!z }"| |z|dzz}#|"|#kDr td| |}| ||} | ||}!t| |!z }"| |z|dzz}#|"|#kDr tdd}$|}%d}&d}'|}(|})|}*d}+d},d}-t|j }.d}/d}0t#||}t#||}1|}|rt t t d||kr|dz }d|%z } | |z}2| |2}|||2zz }|dk\r ||%|$z |zz }| |2|}3||3|%z |zz }|}|}| |}|%}$| ||}%|%dkr tdt|%j}%|,|3dz|$dzz|%dzzz },|dk(r |%|z d|zkrd}|'}4|/|&z|0|3zz}5|0|&z|/|3zz }6|0|%z}'|/ |%z}&t|6|&g}7|)|7z}8t|6|%gj%|}9t!|9|}9|6|9z }/|%|9z }0|/|)z}:|0|)z})d|9z };|1}<|}1|2|4||6}?|?dk(r|#}?|)}(|(}|dk(s|dk(rt(}@n|||zz }@|dk(rt(}An|7|z }A|-|.z }|dk(r>d@z}BdAz}C|Cdkrd}Bdkrd}||k\rd}|d|z k\rd}|=|k\rd}A|krd}@|krd}d}D|dkrd}D|dkrd}D||dz k\rd}D|dzdk(rd}D|(d|=zkrd}D|(d|>zkrd}D|d |z krd}D|dk7rd}D|rLDrJ|d!d"| dd#d"@d$}Ed"Ad$}Fd"|d%d"|d%d"|6|z d%}Gt |E|Fz|Gz|dzdk(r t ||| |dk7rn||kr|rrt t |d&|dd'|d(zt |d)|d*d+|d*zt |d,|d*d-|d*zt |d.8d*zt |||dzz|dk(r|}H| |HfSd}H| HfS)/ak Solve ``Ax = b`` with the MINimum RESidual method, for a real symmetric or complex hermitian matrix `A`. MINRES minimizes ``norm(Ax - b)`` for a real symmetric or complex hermitian matrix `A`. Unlike the Conjugate Gradient method, `A` can be indefinite or singular. If ``shift != 0`` then the method solves ``(A - shift*I)x = b``. Parameters ---------- A : {sparse array, ndarray, LinearOperator} The real symmetric or complex hermitian N-by-N matrix of the linear system. Alternatively, `A` can be a linear operator which can produce ``Ax`` using, e.g., ``scipy.sparse.linalg.LinearOperator``. b : ndarray Right hand side of the linear system. Has shape (N,) or (N,1). Returns ------- x : ndarray The converged solution. info : int Provides convergence information: 0 : successful exit >0 : convergence to tolerance not achieved, number of iterations <0 : illegal input or breakdown Other Parameters ---------------- x0 : ndarray Starting guess for the solution. rtol : float Tolerance to achieve. The algorithm terminates when the relative residual is below ``rtol``. shift : float Value to apply to the system ``(A - shift * I)x = b``. Default is 0. maxiter : int Maximum number of iterations. Iteration will stop after maxiter steps even if the specified tolerance has not been achieved. M : {sparse array, ndarray, LinearOperator} Preconditioner for `A`. The preconditioner should approximate the inverse of `A`. Effective preconditioning dramatically improves the rate of convergence, which implies that fewer iterations are needed to reach a given error tolerance. callback : function User-supplied function to call after each iteration. It is called as ``callback(xk)``, where ``xk`` is the current solution vector. show : bool If ``True``, print out a summary and metrics related to the solution during iterations. Default is ``False``. check : bool If ``True``, run additional input validation to check that `A` and `M` (if specified) are symmetric. Default is ``False``. Notes ----- This file is a translation of the MATLAB implementation [2]_. References ---------- .. [1] Solution of sparse indefinite systems of linear equations, C. C. Paige and M. A. Saunders (1975), SIAM J. Numer. Anal. 12(4), pp. 617-629. https://web.stanford.edu/group/SOL/software/minres/ .. [2] https://web.stanford.edu/group/SOL/software/minres/minres-matlab.zip Examples -------- >>> import numpy as np >>> from scipy.sparse import csc_array >>> from scipy.sparse.linalg import minres >>> A = csc_array([[3, 2, 0], [1, -1, 0], [0, 5, 1]], dtype=float) >>> A = A + A.T >>> b = np.array([2, 4, -1], dtype=float) >>> x, exitCode = minres(A, b) >>> print(exitCode) # 0 indicates successful convergence 0 >>> np.allclose(A.dot(x), b) True zEnter minres. zExit minres. r) z3 beta2 = 0. If M = I, b and x are eigenvectors z/ beta1 = 0. The exact solution is x0 z3 A solution to Ax = b was found, given rtol z3 A least-squares solution was found, given rtol z3 Reasonable accuracy achieved, given eps z3 x has converged to an eigenvector z3 acond has exceeded 0.1/eps z3 The iteration limit was reached z3 A does not define a symmetric matrix z3 M does not define a symmetric matrix z3 M does not define a pos-def preconditioner zSolution of symmetric Ax = bz n = 3gz shift = z23.14ez itnlim = z rtol = z11.2ezindefinite preconditionergUUUUUU?znon-symmetric matrixznon-symmetric preconditioner)dtypezD Itn x(1) Compatible LS norm(A) cond(A) gbar/|A|r g? g?F(Tg{Gz?6g z12.5ez10.3ez8.1ez istop = z itn =5gz Anorm = z12.4ez Acond = z rnorm = z ynorm = z Arnorm = )r rrrmatvecshapeprintrrepscopy ValueErrorr r realabsmaxrastypeminr)IAbx0rrrrrrrxdotprodr%psolvefirstlastnmsgistopitnAnormAcondrnormynormxtyper(r1ybeta1bnormwr2stzepsaoldbbetadbarepslnqrnormphibarrhs1rhs2tnorm2gmaxgmincssnw2valfaoldepsdeltagbarrootArnormgammaphidenomw1epsxepsrdiagtest1test2t1t2prntstr1str2str3infosI m/var/www/ramen.bs-engineer-server.com/venv/lib/python3.12/site-packages/scipy/sparse/linalg/_isolve/minres.pyrr slQ2q)JAq!Q"1od5G XXF XXF E D  Aa% CC  e445 e 1R&f~FFG e 72,od5\JJK  E C E E E E GGE ,  C  z VVX 1Wr A BNE qy455 !1v GE z 1v  E  1I AY AaL AbM AJC3>) t834 4AY AaL BrN AJC3>) t8;< < D D D E F F D D F D <  D B B auA q B B    TU - q H aC 1I  M !8T$YN"Aq| dB    2Jr!} !834 4DII$'D!G#dAg-- !8EzRV# T BI%Dy29$T td{T4L!$dD\"))%0E3 E\ E\6kfE    ]U2X % . AI44 5LeAg~wqyV[[!Qs{u}s"u}t# 19D A:!EU5[)E A:E5LET  A:UBUBQwQwg~Cu}}} 7D "9D '"* D 8q=D RW D RW D DH D A:D D"XQqtEl!E%=9DuUm$DuTl!E$223 dSq\!" z d8O d8O)N)numpyrrrrrr numpy.linalgr mathr scipy._lib._array_apir utilsr __all__rrqrprysN>>1 *n$c4 DuEnnrq