12.07.2015 Views

Numerical Mathematics - A Collection of Solved Problems

Numerical Mathematics - A Collection of Solved Problems

Numerical Mathematics - A Collection of Solved Problems

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

252 INTERPOLACIJA I APROKSIMACIJANije teško zaključiti da je tražena aproksimaciona funkcija data sagde suΦ ∗ (x) = 1 2 (y s + y t ) = ā 0 + ā 1 x ,ā 1 =f(d) − f(c)d − c, ā 0 = 1 2 (f(c) + f(x 2)) − 1 2 (c + x 2)f(d) − f(c)d − c,pri čemu tačku x 2 nalazimo iz (1).Drugi način: Na osnovu teoreme o Čebiševljevoj alternansi (videti [2, str. 118–119]), polinom Φ ∗ (x) = ā 0 +ā 1 x je najbolja mini-max aproksimacija za f ∈ C[c, d],ako i samo ako na [c, d] postoje bar tri tačke x 1 , x 2 , x 3 (x 1 < x 2 < x 3 ), takve daje(2) δ ∗ (x 1 ) = −δ ∗ (x 2 ) = δ ∗ (x 3 ) = ±‖δ ∗ (x)‖ ∞ .S druge strane, s obzirom da jezaključujemo da jeδ ′′ (x) = f ′′ (x) > 0 (< 0)δ ′ (x) = f ′ (x) − a 1monotona funkcija, pa kao takva može imati najviše jednu realnu nulu.Dakle, na osnovu prethodnog, zaključujemo da jea tačka x 2 je koren jednačinex 1 = c, x 3 = d,(3) δ ′ (x 2 ) = f ′ (x 2 ) − a 1 = 0 .Sada, na osnovu (2) imamoodakle dobijam<strong>of</strong>(c) − ā 0 − ā 1 c = −(f(x 2 ) − ā 0 − ā 1 x 2 ) = f(d) − ā 0 − ā 1 d ,ā 1 =f(d) − f(c)d − c, ā 0 = 1 2 (f(c) + f(x 2)) − 1 2 (c + x 2)f(d) − f(c)d − c,pri čemu je x 2 koren jednačine (3), tj.f ′ (x 2 ) =f(d) − f(c)d − c(x 2 ∈ (c, d)).

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!