Why Compare Candidate Methods?
An integrand can invite more than one reasonable approach. You might recognize a composition and try substitution, or expand an expression and integrate term by term. Both approaches may be valid, but they need not demand the same amount of algebra.
The aim is not to assign a universal score to every method. Instead, compare the work each method creates: how complicated the new integrand becomes, how many transformations are needed, and whether the resulting antiderivative is easy to check. This kind of comparison helps avoid lengthy calculations when a simpler route is available.
Three Measures of Algebraic Cost
Use three questions to compare candidate methods. They are guides rather than a rigid scoring system.
- What does the new integrand look like? Does it become a familiar power or a short sum, or does it acquire more terms and awkward factors?
- How many transformations are required? Count meaningful rewrites and substitutions, including the algebra needed to express every part of the integrand in the new variable.
- How easy is the result to check? A compact expression can be easy to differentiate; an expanded expression may be longer but can make individual terms easy to inspect.
A method with fewer visible steps is not automatically better. A substitution that leaves the original variables mixed with the new variable has not simplified the problem. Likewise, expansion can be helpful when it exposes a short polynomial, but wasteful when it produces many terms that could have been avoided.
Worked Example: Substitute or Expand?
Worked Example: Substitute or Expand?
Compare two ways to evaluate \(\int x(x^2+1)^3\,dx\).
Method 1: Use the inner expression as a substitution. Let \(u=x^2+1\). Then \(du=2x\,dx\), so \(x\,dx=\frac12\,du\). Therefore
This route requires one substitution and one power-rule integration. The new integrand is a single power.
Method 2: Expand first. Using the binomial expansion, \((x^2+1)^3=x^6+3x^4+3x^2+1\). Multiplying by \(x\) gives \(x^7+3x^5+3x^3+x\). Thus
Expansion is also valid, but it creates four terms to integrate. The two answers differ only by a constant: expanding \(\frac18(x^2+1)^4\) gives the second expression plus \(\frac18\).
Check the compact answer:
Both methods work, but substitution has lower algebraic cost here: it avoids expansion and leaves a one-term integrand. The check is also short because the expression still displays its composition.
Why Substitution Works Here
The comparison above depends on a precise relationship between the new variable and the original integrand. In this example, \(u=x^2+1\) has derivative \(2x\), and the factor \(x\,dx\) is therefore \(\frac12\,du\). More generally, the chain rule explains why substitution can simplify an integral.
This is also a useful way to assess a candidate substitution before carrying it out: look for the derivative of the proposed inner expression, including any constant factor. If that factor is absent, the proposed rewrite is not yet justified.
Worked Example: Compare Two Substitutions
Worked Example: Compare Two Substitutions
Consider \(\int x^3\sqrt{x^2+1}\,dx\). Two substitutions can turn it into a polynomial or a sum of powers.
Method 1: Set \(u=x^2+1\). Since \(du=2x\,dx\), we have \(x\,dx=\frac12\,du\), and \(x^2=u-1\). Rewrite \(x^3\,dx\) as \(x^2(x\,dx)\):
The new integrand has two power terms. The substitution requires expressing \(x^2\) as \(u-1\), as well as replacing \(x\,dx\).
Method 2: Set \(v=\sqrt{x^2+1}\). Then \(v^2=x^2+1\), so \(x^2=v^2-1\). Differentiating gives \(2v\,dv=2x\,dx\), or \(x\,dx=v\,dv\). Also \(\sqrt{x^2+1}=v\). Hence
Both substitutions produce a two-term power integral and the same antiderivative. The first keeps the substitution variable tied directly to the expression under the radical. The second removes the radical from the integrand, but requires differentiating the square root and replacing \(x\,dx\). Neither is automatically superior; the relative cost is modest, and a learner may find one rewrite clearer than the other.
Check the common answer: Put \(u=x^2+1\). Its derivative is
The final check confirms the result and shows why the two forms agree.
Transformations Should Reduce the Work
A transformation is useful when it makes the next step easier. For instance, rewriting an expression can expose a familiar integration form; earlier material on polynomial division and completing the square develops two such strategic rewrites. But an extra transformation has a cost: it takes algebra, creates opportunities for errors, and may make the final answer harder to read.
When comparing options, write down the resulting integrands before committing. Ask whether a proposed rewrite removes complexity or merely moves it. If a substitution is chosen, check that every occurrence of the original variable has been handled. If an expression is expanded, count the terms that will need integration.
Worked Example: Division Before Integration
Worked Example: Division Before Integration
Evaluate \(\int \frac{x^2+1}{x+1}\,dx\). One candidate is to simplify the rational expression by polynomial division. Since \(x^2+1=(x+1)(x-1)+2\),
Therefore, on any interval that does not cross \(x=-1\),
As a second candidate, split the numerator first:
The first fraction still needs simplification. Using \(x^2=(x+1)(x-1)+1\), it becomes \(x-1+\frac{1}{x+1}\). Adding the remaining fraction gives \(x-1+\frac{2}{x+1}\), so the same antiderivative follows. This route adds a split and then a separate simplification; division exposes the integrable terms directly.
Check:
The restriction \(x\ne-1\) matters because the original integrand is undefined there. A domain condition is part of checking that a proposed antiderivative applies to the integrand under consideration.
A Concrete Algebra Pitfall
A method that is efficient in principle can still fail if a transformation is carried out incorrectly. For example, after expanding the first worked integrand, it would be wrong to write
The final term is incorrect: multiplying the constant \(1\) by \(x\) gives \(x\), not \(1\). The correct expansion is
The wrong expansion would lead to the proposed antiderivative \(\frac{x^8}{8}+\frac{x^6}{2}+\frac{3x^4}{4}+x+C\). Differentiating gives \(x^7+3x^5+3x^3+1\), which is not the original integrand. Checking exposes the mistake. In practice, a compact route can reduce the number of expansion steps where such errors occur, but every route still needs accurate algebra.
Make the Comparison Explicit
A short comparison table can help when two routes seem plausible. Record what each route does to the integrand rather than choosing solely by habit.
| Question | What to look for |
|---|---|
| Resulting integrand | Fewer terms, familiar forms, or less complicated factors |
| Transformations | Necessary rewrites and substitutions, including all variable replacements |
| Checking | Whether differentiating the proposed answer recovers the original integrand clearly |
This comparison is especially useful when different methods appear equally plausible. A method can win by producing fewer terms, by avoiding awkward algebra, or by making the final check more transparent. The best choice is the one that reduces the total work while preserving a clear path back to the original integrand.
Check Your Understanding
For each question, compare the algebra required by the candidate methods, not just the names of the methods.
- For \(\int x(x^2+4)^2\,dx\), what new integrand results from setting \(u=x^2+4\)? How many terms result from expansion instead?
- Why must a substitution account for every occurrence of the original variable in the integrand?
- In the radical example, what additional replacement is needed after setting \(u=x^2+1\)?
- Why can an algebraically correct antiderivative still require a domain qualification?
- What does differentiating a proposed antiderivative tell you about a calculation?
