MATH266 Group Project
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
1. This question is concerned with solutions of the equation C (z) = 0, where
C (z) = z3 + pz + q,
for real parameters p and q . The discriminant of C is given by
∆ = q2 + p3
There are three real roots if ∆ ≤ 0 (including repeated roots if ∆ = 0). Otherwise there is one real root and a complex conjugate pair. In your analysis, you can assume that p 0 and q 0. When taking a complex cube root, keep in mind that there are three possible values (unless z = 0).
Thus, if z = reiθ with r > 0 and −π < θ ≤ π, then
z1/3 = ^3rei(2kπ+θ)/3, k = 0, 1, 2.
You need not prove that taking the cube root of a complex number is a safe operation, and you
can assume that ∆ itself is safely calculated.
(a) Verify that the three roots of C are given by
zj = S+ + S− where S± = (− ± ∆ 1/2 )1/3 ,
(*)
roots can be taken in the‘obvious’way, because using the alternative values just interchanges S+ and S− .)
Avoid messy algebra: you may find that it helps to calculate S + S and S+ S− first.
(b) Explain how both S+ and S− can always be safely calculated (i.e. without any danger of
catastrophic cancellation).
(c) Show that if ∆ ≤ 0 then S− = S+ (the complex conjugate). Hence explain how all three roots can be safely calculated in this case.
(d) Show that if p < 0 but ∆ ≥ 0 then the real root can be safely calculated. The complex roots can then be calculated through their relationships to the coefficients of C . Can this be achieved without any danger of catastrophic cancellation? Why (or why not)?
(e) Show that the imaginary parts of the complex roots can always be safely calculated if p > 0. Can the other information about the roots (the real root and the real part of the complex roots) be safely obtained in this case? Why (or why not)?
Hint: find and simplify an expression for the product of the complex roots.
(f) (i) Write a Maple procedure that takes as its arguments real numbers p and q and returns accurately calculated roots for C (z) as its results. The procedure should work by first calculating S+ and S− , and then using one or both of these to compute the roots, taking advantage of the above analysis. Your procedure should handle p = 0 and/or q = 0 as special cases.
Be careful taking cube roots. In Maple, z^(1/3) gives the principle value (i.e. with argument divided by 3) . In contrast, surd( x , 3 ) returns a real cube root for any real x .
(ii) Test your procedure with some different values for p and q, and check that the results are accurate (no more than one or two significant figures should be lost). Make sure all three cases (p > 0, p < 0 with ∆ > 0 and ∆ < 0) are tested, and include some cases where one might expect to see cancellation were Cardano’s formula to be applied naively.
2. Let f (x) be a quartic polynomial with distinct, real roots α 1 , . . . , α4 .
Show that
S (x) = = + + + ,
and find a similar expression for T (x) = − .
(b) Let x0 and λ be real numbers such that x0 αj for j = 1, 2, 3, 4 and λ x0 , and consider
the quadratic
Q(x) = K (x − x0 )2 − (x − λ)2 ,
where
(i) What can be said about roots of Q(x) between x0 and αj ?
(ii) By writing αj − λ = (αj − x0 ) + (x0 − λ), prove that
K = 4 − 2(x0 − λ)S0 + (x0 − λ)2T0 ,
where S0 = S (x0 ) and T0 = T (x0 ).
(c) (i) Make the change of variables y = x − x0 and µ = x0 − λ to obtain an expression for
roots of Q as close as possible to the roots of f — occurs at the point where ∂Q/∂µ = 0.
Denote this value by µ0 ; find its value, and then show that Q(y, µ0 ) = 0 if
y2 [3T0 − S0(2)] − 2S0y − 4 = 0.
In your answer, you may assume that T0y2 1.
(ii) Is it possible to have T0y2 = 1 and ∂Q/∂µ = 0 at the root? Why (or why not)?
(d) By solving the quadratic from part (c) and writing f0 = f (x0 ), f = f\ (x0 ) etc. show that
x1 = ϕ(x0 ) = x0 − 4f0
Hint: move the square root into the denominator before eliminating S0 and T0 .
2023-04-04