A quadratic-based modified projection method

Abstract

COMMON N,K,ANRMF REAL*8 IP REAL*8 X(3),F(3).81 31,C(3>.A(3. 3),J(3),G(3,3) N=3 READ. (X(I) >1=1 ,N) REAO,(Bl I).I=1 ,N) READ , (C(I) .1 = 1.N) DO 3 1=1,N READ , (A(I.L).L=l,N) 3 CONTINUE c c function for the root of x**3+a*x**2+b*x+c=0 giving c the minimizer of (x**4)/4+a/3*x**3+8/2*x**2+c*x+d c c double precision function soln(a.b.c) double precision a.b.c.d.s»l.pi.cbrt pi=3«141592653589793d0 s=(3.d0*b-a*a)/9.d0 l=(a*b-3.00*c)/6.do-a*a*a/2 7.00 0=l+l+s*s*s if(0«ge.o.oo)then do sol n=cbrt (i)sqmt (d) +l) +cbrt(l-dscr t(dj)-a/3.d0 else dc if{l.gt.oado .or.(l.eq.0.00 .and. a.ge.o.do))then dc soln=2.do*dsqrt(-sj•dcos(darcos(-l/s/dsqrt(-s))/