Pa2 - Equations with transformed roots

Pa2 - Equations with transformed roots

You can move or rescale every root of a cubic or quartic without finding any root individually. Pa2 uses a linear relation between an old root and a new one to build the new equation by substitution. You will handle shifts, scale factors and their combinations, keeping all roots and simplifying the final coefficients.

Moving every root by the same amount

The cubic x3−3x2−x+3=0x^3-3x^2-x+3=0 has roots α\alpha, β\beta, γ\gamma. What equation has roots α+2\alpha+2, β+2\beta+2, γ+2\gamma+2? You do not need to solve the original cubic first.

Call an old root xx and its corresponding new root yy. The requested change is y=x+2y=x+2, so the old value is x=y−2x=y-2. Substitute that old value into the original equation. Whenever yy is a requested new root, y−2y-2 is an old root and makes the original polynomial zero.

You already know

If roots have sum S1S_1, the coefficient of the next-highest power in a monic polynomial is −S1-S_1. Multiplying an equation by a non-zero constant does not change its roots. A repeated root is counted with its multiplicity.

The picture makes the direction clear. Moving roots to the right by 22 requires evaluating the old polynomial at an input 22 smaller.

[DIAGRAM: asset_name: Shifting all three roots; asset_slug: transformed_roots_shift; caption: Replacing the input by y minus 2 moves every root two units to the right; description: Two equal-scale graphs show f(x) equals x cubed minus 3x squared minus x plus 3 with roots minus 1, 1 and 3, and g(y) equals f(y minus 2) with roots 1, 3 and 5. Matching root labels and horizontal dimension arrows identify the shift of 2. The second graph has the same shape translated two units right.]
Replacing the input by y minus 2 moves every root two units to the right

Use an inverse substitution: describe the new root, then rearrange for the old one.

Shifting roots

If f(x)=0f(x)=0 is a cubic or quartic and the required roots are each old root plus hh, set

y=x+h⟹x=y−hy=x+h\quad\Longrightarrow\quad x=y-h

The new equation is

f(y−h)=0f(y-h)=0

The substitution is reversible, so each old root gives a new root and each new root comes from an old root, with multiplicities preserved.

Not in the formula booklet: learn it.

Keep the substituted expression bracketed until every power has been expanded. The shift affects the entire input, not only the leading term.

Worked example 1: shifting roots to the right

Question. The roots of x3−3x2−x+3=0x^3-3x^2-x+3=0 are α\alpha, β\beta, γ\gamma. Determine a cubic equation whose roots are α+2\alpha+2, β+2\beta+2, γ+2\gamma+2.

Plan. Write y=x+2y=x+2, rearrange to x=y−2x=y-2, substitute, expand and collect powers of yy.

x=y−2old root in terms of new(y−2)3−3(y−2)2−(y−2)+3=0substitute everywherey3−6y2+12y−8expand the cubic−3y2+12y−12−y+2+3=0expand the other termsy3−9y2+23y−15=0collect like powers\begin{aligned} x&=y-2 && \text{old root in terms of new} \\ (y-2)^3-3(y-2)^2-(y-2)+3&=0 && \text{substitute everywhere} \\ y^3-6y^2+12y-8&\quad && \text{expand the cubic} \\ {}-3y^2+12y-12-y+2+3&=0 && \text{expand the other terms} \\ y^3-9y^2+23y-15&=0 && \text{collect like powers} \end{aligned}

Why this step? Substituting y−2y-2, rather than y+2y+2, ensures that the new root α+2\alpha+2 sends the original polynomial back to its root α\alpha.

Answer. y3−9y2+23y−15=0y^3-9y^2+23y-15=0.

Check. The old root sum is 33, so the new sum is 3+3(2)=93+3(2)=9, matching the coefficient −9-9. Also, the factorisation (x+1)(x−1)(x−3)(x+1)(x-1)(x-3) shifts to (y−1)(y−3)(y−5)(y-1)(y-3)(y-5), which expands to the same result.

Worked example 2 changes only the direction of the shift: subtract 22 from each old root. The sign inside the substitution reverses too.

Worked example 2: shifting roots to the left

Question. Find an equation whose roots are α−2\alpha-2, β−2\beta-2, γ−2\gamma-2, for the same original cubic.

Plan. Start with y=x−2y=x-2, so x=y+2x=y+2, then substitute into the original equation.

x=y+2invert the new relation(y+2)3−3(y+2)2−(y+2)+3=0substitutey3+6y2+12y+8expand the cubic−3y2−12y−12−y−2+3=0expand the other termsy3+3y2−y−3=0collect terms\begin{aligned} x&=y+2 && \text{invert the new relation} \\ (y+2)^3-3(y+2)^2-(y+2)+3&=0 && \text{substitute} \\ y^3+6y^2+12y+8&\quad && \text{expand the cubic} \\ {}-3y^2-12y-12-y-2+3&=0 && \text{expand the other terms} \\ y^3+3y^2-y-3&=0 && \text{collect terms} \end{aligned}

Why this step? Moving the roots left uses the old value y+2y+2, because adding 22 to a new root recovers its old position.

Answer. y3+3y2−y−3=0y^3+3y^2-y-3=0.

Check. The new root sum is 3−3(2)=−33-3(2)=-3, matching the coefficient +3+3. The shifted factors are (y+3)(y+1)(y−1)(y+3)(y+1)(y-1).

The sign error is easy to detect before expanding anything: put one requested new root into your substituted input.

Watch out

Wrong: for new roots α+2\alpha+2, substitute x=y+2x=y+2. Then y=α+2y=\alpha+2 gives x=α+4x=\alpha+4, which need not be an old root. Right: x=y−2x=y-2 gives x=αx=\alpha, exactly as required.

The next cubic and shift are different. Show the inverse relation before expanding.

Multiplying every root by a scale factor

If every new root is kk times an old one, divide the new variable by kk before putting it into the original polynomial.

For example, a desired new root 2α2\alpha must become α\alpha inside the original polynomial. Its input must therefore be y/2y/2. Multiplying out afterwards may introduce fractions, but these can be cleared by multiplying the whole equation by a suitable non-zero constant.

Scaling roots

For a cubic or quartic f(x)=0f(x)=0 and a non-zero scale factor kk,

y=kx⟹x=yky=kx\quad\Longrightarrow\quad x=\frac yk

The equation with scaled roots is

f(yk)=0f\left(\frac yk\right)=0

For degree nn, where n=3n=3 or 44, multiplying by knk^n clears powers of kk from the denominators. This changes no root because kn≠0k^n\ne0.

Not in the formula booklet: learn it.

A factor such as 12\frac12 in the input is raised to the power of each term. It does not multiply the whole polynomial uniformly.

Worked example 3: doubling all roots

Question. The roots of x3−3x2−x+3=0x^3-3x^2-x+3=0 are α\alpha, β\beta, γ\gamma. Determine a cubic equation with integer coefficients whose roots are 2α2\alpha, 2β2\beta, 2γ2\gamma.

Plan. Substitute x=y/2x=y/2, then multiply the entire equation by 88.

x=y2invert y=2xy38−3y24−y2+3=0substitute with powersy3−6y2−4y+24=0multiply every term by 8\begin{aligned} x&=\frac y2 && \text{invert }y=2x \\ \frac{y^3}{8}-\frac{3y^2}{4}-\frac y2+3&=0 && \text{substitute with powers} \\ y^3-6y^2-4y+24&=0 && \text{multiply every term by }8 \end{aligned}

Why this step? The cubic term has denominator 23=82^3=8, so multiplying the whole equation by 88 clears every denominator at once.

Answer. y3−6y2−4y+24=0y^3-6y^2-4y+24=0.

Check. The original sum 33 doubles to 66. The original product is −3-3, and three doubled roots have product 23(−3)=−242^3(-3)=-24, matching the new constant +24+24.

Worked example 4 changes only the scale factor from 22 to 12\frac12. The inverse substitution now multiplies the input by 22.

Worked example 4: halving all roots

Question. Find a cubic equation with integer coefficients whose roots are α2\frac\alpha2, β2\frac\beta2, γ2\frac\gamma2, for the same original cubic.

Plan. Use y=x/2y=x/2, so x=2yx=2y, then substitute and simplify.

x=2yinvert the halving relation(2y)3−3(2y)2−2y+3=0substitute everywhere8y3−12y2−2y+3=0raise each factor to its power\begin{aligned} x&=2y && \text{invert the halving relation} \\ (2y)^3-3(2y)^2-2y+3&=0 && \text{substitute everywhere} \\ 8y^3-12y^2-2y+3&=0 && \text{raise each factor to its power} \end{aligned}

Why this step? Halving the roots means that twice a new root is an old root. Substituting y/2y/2 would double the roots instead.

Answer. 8y3−12y2−2y+3=08y^3-12y^2-2y+3=0.

Check. The coefficient ratios give root sum 12/8=3/212/8=3/2 and product −3/8-3/8, half the old sum and one eighth of the old product.

For the boundary case k=0k=0, every requested new root is zero, so the required equation is y3=0y^3=0 or y4=0y^4=0, according to the number of roots. The inverse substitution is unavailable because division by zero is undefined; form that special equation directly.

Watch out

Wrong: (y/2)3=y3/2(y/2)^3=y^3/2. Right: (y/2)3=y3/8(y/2)^3=y^3/8. Each power acts on the whole substituted fraction, including its denominator.

The next polynomial is not monic, and its scale factor is 33. Preserve its leading coefficient when clearing denominators.

Combining a scale factor and a shift

A relation such as y=(x+1)/2y=(x+1)/2 changes both the scale and the position of each root, so rearrange the complete relation before substituting.

The order matters. Halving x+1x+1 gives x/2+1/2x/2+1/2, whereas halving xx and then adding 11 gives x/2+1x/2+1. Read the brackets in the requested roots exactly as printed.

A general linear transformation of roots

For a cubic or quartic f(x)=0f(x)=0, let every new root be y=mx+cy=mx+c, where m≠0m\ne0. Rearranging gives

x=y−cmx=\frac{y-c}{m}

so the required equation is

f(y−cm)=0f\left(\frac{y-c}{m}\right)=0

Clear denominators with a non-zero constant and simplify. If m=0m=0, all new roots equal cc; form (y−c)3=0(y-c)^3=0 or (y−c)4=0(y-c)^4=0 directly.

Not in the formula booklet: learn it.

You may also construct the new root sums and products, but substitution keeps the whole relationship in one place. It is particularly helpful when fractions and translations occur together.

Worked example 5: halving after adding one

Question. The cubic 2x3−3x−1=02x^3-3x-1=0 has roots α\alpha, β\beta, γ\gamma. Determine a cubic equation with integer coefficients whose roots are α+12\frac{\alpha+1}{2}, β+12\frac{\beta+1}{2}, γ+12\frac{\gamma+1}{2}.

Plan. Rearrange y=(x+1)/2y=(x+1)/2, substitute x=2y−1x=2y-1, then collect and remove a common constant factor.

x=2y−1multiply by 2 then subtract 12(2y−1)3−3(2y−1)−1=0substitute in the cubic2(8y3−12y2+6y−1)−6y+3−1=0expand the cube16y3−24y2+6y=0collect all terms8y3−12y2+3y=0divide by non-zero constant 2\begin{aligned} x&=2y-1 && \text{multiply by }2\text{ then subtract }1 \\ 2(2y-1)^3-3(2y-1)-1&=0 && \text{substitute in the cubic} \\ 2(8y^3-12y^2+6y-1)-6y+3-1&=0 && \text{expand the cube} \\ 16y^3-24y^2+6y&=0 && \text{collect all terms} \\ 8y^3-12y^2+3y&=0 && \text{divide by non-zero constant }2 \end{aligned}

Why this step? Dividing by 22 preserves every root; dividing by yy would lose the root y=0y=0.

Answer. 8y3−12y2+3y=08y^3-12y^2+3y=0.

Check. The original equation has root −1-1, which transforms to 00; the new constant is correctly zero. Its root sum is (0+3)/2=3/2(0+3)/2=3/2, matching 12/812/8.

Worked example 6 changes only the +1+1 in the requested roots to −1-1. The rearranged substitution changes from 2y−12y-1 to 2y+12y+1.

Worked example 6: halving after subtracting one

Question. For the same original cubic, determine an equation with integer coefficients whose roots are α−12\frac{\alpha-1}{2}, β−12\frac{\beta-1}{2}, γ−12\frac{\gamma-1}{2}.

Plan. Rearrange y=(x−1)/2y=(x-1)/2, substitute x=2y+1x=2y+1, then simplify the resulting equation.

x=2y+1invert the new relation2(2y+1)3−3(2y+1)−1=0substitute2(8y3+12y2+6y+1)−6y−3−1=0expand the cube16y3+24y2+6y−2=0collect terms8y3+12y2+3y−1=0divide by 2\begin{aligned} x&=2y+1 && \text{invert the new relation} \\ 2(2y+1)^3-3(2y+1)-1&=0 && \text{substitute} \\ 2(8y^3+12y^2+6y+1)-6y-3-1&=0 && \text{expand the cube} \\ 16y^3+24y^2+6y-2&=0 && \text{collect terms} \\ 8y^3+12y^2+3y-1&=0 && \text{divide by }2 \end{aligned}

Why this step? The final −1-1 in the original polynomial stays outside the substituted expression. Only occurrences of xx are replaced.

Answer. 8y3+12y2+3y−1=08y^3+12y^2+3y-1=0.

Check. The old root −1-1 becomes (−1−1)/2=−1(-1-1)/2=-1. Substitution gives −8+12−3−1=0-8+12-3-1=0. The new root sum is −3/2-3/2, matching the coefficient ratio −12/8-12/8.

A requested equation must remain an equation after simplification.

Exam tip

End with the polynomial equal to zero, using integer coefficients when requested. You may change its variable name, or multiply it by a non-zero constant, but a polynomial expression alone does not state the required equation.

The next root relation adds 22 and divides by 33. Work backwards through those operations before expanding.

Watch out

Wrong: from y=(x−1)/2y=(x-1)/2, substitute x=(y−1)/2x=(y-1)/2. That applies the change again instead of undoing it. The correct rearrangement is 2y=x−12y=x-1, so x=2y+1x=2y+1.

Rearrange the next complete root relation before replacing any occurrence of the old variable.

Transforming quartics and checking coefficients

A quartic follows exactly the same inverse-substitution rule; the extra power makes organised expansion and a coefficient check more valuable.

For a cubic, a shift by cc changes the root sum by 3c3c. For a quartic it changes it by 4c4c. In general, for these degrees, a transformation y=mx+cy=mx+c changes a root sum S1S_1 to mS1+ncmS_1+nc, where nn is the number of roots counting multiplicity. This gives an independent check on the next-highest coefficient after expansion.

Checking the transformed root sum

For a cubic or quartic with nn roots, n=3n=3 or 44, and transformed roots yj=mxj+cy_j=mx_j+c,

∑j=1nyj=m∑j=1nxj+nc\sum_{j=1}^{n}y_j=m\sum_{j=1}^{n}x_j+nc

The transformed sum must equal minus the ratio of the next-highest coefficient to the leading coefficient in the new equation.

Not in the formula booklet: learn it.

For expansion, first clear denominators while retaining brackets. Then work from the highest power down, grouping coefficients of the same power together.

Worked example 7: a quartic with scaled and shifted roots

Question. The roots of x4−2x3+3x2−4x+2=0x^4-2x^3+3x^2-4x+2=0 are α\alpha, β\beta, γ\gamma, δ\delta. Determine an equation with integer coefficients whose roots are 2α+12\alpha+1, 2β+12\beta+1, 2γ+12\gamma+1, 2δ+12\delta+1.

Plan. Substitute x=(y−1)/2x=(y-1)/2, multiply the equation by 1616, expand and check the root sum.

x=y−12invert y=2x+1(y−1)416−(y−1)34+3(y−1)24−2(y−1)+2=0substitute(y−1)4−4(y−1)3+12(y−1)2−32(y−1)+32=0multiply by 16\begin{aligned} x&=\frac{y-1}{2} && \text{invert }y=2x+1 \\ \frac{(y-1)^4}{16}-\frac{(y-1)^3}{4} +\frac{3(y-1)^2}{4}-2(y-1)+2&=0 && \text{substitute} \\ (y-1)^4-4(y-1)^3+12(y-1)^2-32(y-1)+32&=0 && \text{multiply by }16 \end{aligned}

The needed expansions are

(y−1)4=y4−4y3+6y2−4y+1(y−1)3=y3−3y2+3y−1(y−1)2=y2−2y+1\begin{aligned} (y-1)^4&=y^4-4y^3+6y^2-4y+1 \\ (y-1)^3&=y^3-3y^2+3y-1 \\ (y-1)^2&=y^2-2y+1 \end{aligned}

Collecting gives

y4+(−4−4)y3+(6+12+12)y2+(−4−12−24−32)y+(1+4+12+32+32)=0y4−8y3+30y2−72y+81=0\begin{aligned} y^4+(-4-4)y^3+(6+12+12)y^2&\quad \\ {}+(-4-12-24-32)y+(1+4+12+32+32)&=0 \\ y^4-8y^3+30y^2-72y+81&=0 \end{aligned}

Why this step? Multiplying before expanding leaves integer coefficients outside each bracket, which makes the powers and signs easier to track.

Answer. y4−8y3+30y2−72y+81=0y^4-8y^3+30y^2-72y+81=0.

Check. The old root sum is 22, so the new sum is 2(2)+4(1)=82(2)+4(1)=8, matching the coefficient −8-8. At y=1y=1, the substituted input is 00, and both sides of the polynomial identity give 16f(0)=3216f(0)=32.

Worked example 8 changes only the translation from +1+1 to −1-1. This time the new root sum is zero, so the cubic term should disappear.

Worked example 8: making the cubic term disappear

Question. For the same original quartic, determine an equation whose roots are 2α−12\alpha-1, 2β−12\beta-1, 2γ−12\gamma-1, 2δ−12\delta-1.

Plan. Substitute x=(y+1)/2x=(y+1)/2, clear denominators and check for a zero root sum.

x=y+12invert y=2x−1(y+1)4−4(y+1)3+12(y+1)2−32(y+1)+32=0substitute and multiply by 16\begin{aligned} x&=\frac{y+1}{2} && \text{invert }y=2x-1 \\ (y+1)^4-4(y+1)^3+12(y+1)^2-32(y+1)+32&=0 && \text{substitute and multiply by }16 \end{aligned}

Using the positive-sign binomial expansions,

y4+(4−4)y3+(6−12+12)y2+(4−12+24−32)y+(1−4+12−32+32)=0y4+6y2−16y+9=0the cubic terms cancel\begin{aligned} y^4+(4-4)y^3+(6-12+12)y^2&\quad \\ {}+(4-12+24-32)y+(1-4+12-32+32)&=0 \\ y^4+6y^2-16y+9&=0 && \text{the cubic terms cancel} \end{aligned}

Why this step? The missing cubic term is predicted by the transformed sum 2(2)−4=02(2)-4=0. It is a mathematical feature, not an omitted term.

Answer. y4+6y2−16y+9=0y^4+6y^2-16y+9=0.

Check. At y=−1y=-1, the old input is 00. The new polynomial gives 1+6+16+9=32=16f(0)1+6+16+9=32=16f(0), agreeing with the substituted identity.

Here is a useful surprise: you can always remove the next-highest term of a cubic or quartic by subtracting the mean of its roots from every root. The new sum is zero even when you do not know the individual roots. For a cubic use the old sum divided by 33; for a quartic divide by 44. This follows directly from the root-sum check above and is not in the formula booklet: derive it when needed.

Watch out

Wrong: cancel a factor yy from the transformed equation because every term contains it. That removes the root y=0y=0. Right: divide only by a known non-zero constant when simplifying an equation intended to retain the complete root list.

The opening question needed no root solving: substituting x=y−2x=y-2 produced y3−9y2+23y−15=0y^3-9y^2+23y-15=0. The quartic method is the same operation with one additional power to expand.

In the final check, the original quartic is not monic. Keep its leading coefficient until the final common constant is removed.

Quick Recap

  • Name a new root yy, write its relation to an old root xx, and rearrange for xx before substituting.
  • A shift by hh uses x=y−hx=y-h; scaling by non-zero kk uses x=y/kx=y/k.
  • A combined change y=mx+cy=mx+c uses x=(y−c)/mx=(y-c)/m. The method works for cubics and quartics without finding their individual roots.
  • Clear denominators and simplify using non-zero constants. Preserve zero roots and repeated roots, and finish with an equation equal to zero.
  • Check the new root sum against the next-highest coefficient: it is mS1+ncmS_1+nc.

For the opening cubic, the equation with roots α+2\alpha+2, β+2\beta+2, γ+2\gamma+2 is y3−9y2+23y−15=0y^3-9y^2+23y-15=0.