Login

Concepts

Read through concept notes with worked explanations and diagrams, chapter by chapter.

Concept 06 / 07Binomial Theorem

Properties of Binomial Coefficients and Series Summation

Sums of coefficients by substitution (x = 1, -1, roots of unity), even and odd coefficient sums, sums weighted by r or 1/(r + 1) via differentiation and integration, Vandermonde's identity for sums of products, and choosing the right technique.

Write Cr=(nr)C_{r} = \binom{n}{r} and start from (1+x)n=C0+C1x+C2x2+⋯+Cnxn(1 + x)^{n} = C_{0} + C_{1}x + C_{2}x^{2} + \cdots + C_{n}x^{n}. Almost every identity comes from one of four operations on this equation: substitute a value, differentiate, integrate, or multiply two expansions and compare coefficients. Substitution. x=1x = 1: C0+C1+⋯+Cn=2nC_{0} + C_{1} + \cdots + C_{n} = 2^{n}. x=−1x = -1: C0−C1+C2−⋯+(−1)nCn=0C_{0} - C_{1} + C_{2} - \cdots + (-1)^{n}C_{n} = 0 (for n≥1n \ge 1). Adding and subtracting: C0+C2+C4+⋯=C1+C3+C5+⋯=2n−1C_{0} + C_{2} + C_{4} + \cdots = C_{1} + C_{3} + C_{5} + \cdots = 2^{n-1}. x=ωx = \omega (a cube root of unity) extracts sums over every third coefficient; x=ix = i separates the sums over r≡0,1,2,3(mod4)r \equiv 0, 1, 2, 3 \pmod 4. For example, C0+C3+C6+⋯=13(2n+(1+ω)n+(1+ω2)n)=13(2n+2cos⁡nπ3)C_{0} + C_{3} + C_{6} + \cdots = \frac{1}{3}(2^{n} + (1 + \omega)^{n} + (1 + \omega^{2})^{n}) = \frac{1}{3}(2^{n} + 2\cos\frac{n\pi}{3}). Differentiation, for sums weighted by rr. Differentiating gives n(1+x)n−1=∑rrCrxr−1n(1 + x)^{n-1} = \sum_{r} rC_{r}x^{r-1}. At x=1x = 1: ∑rCr=n2n−1\sum rC_{r} = n2^{n-1}. At x=−1x = -1: ∑(−1)r−1rCr=0\sum(-1)^{r-1}rC_{r} = 0 for n≥2n \ge 2. For ∑r2Cr\sum r^{2}C_{r}, multiply by xx before differentiating again, or use r2=r(r−1)+rr^{2} = r(r - 1) + r and ∑r(r−1)Cr=n(n−1)2n−2\sum r(r - 1)C_{r} = n(n - 1)2^{n-2}; this gives ∑r2Cr=n(n+1)2n−2\sum r^{2}C_{r} = n(n + 1)2^{n-2}. An alternative to calculus is the absorption identity rCr=n(n−1r−1)rC_{r} = n\binom{n - 1}{r - 1}, which turns ∑rCr\sum rC_{r} into n∑(n−1r−1)=n2n−1n\sum\binom{n - 1}{r - 1} = n2^{n-1} directly. Integration, for sums weighted by 1r+1\frac{1}{r + 1}. Integrate from 00 to 11: ∫01(1+x)ndx=2n+1−1n+1=∑Crr+1\int_{0}^{1}(1 + x)^{n}dx = \frac{2^{n+1} - 1}{n + 1} = \sum\frac{C_{r}}{r + 1}. Integrating from −1-1 to 00 or 00 to −1-1 gives alternating versions: ∑(−1)rCrr+1=1n+1\sum\frac{(-1)^{r}C_{r}}{r + 1} = \frac{1}{n + 1}. Again there is a non-calculus route: Crr+1=1n+1(n+1r+1)\frac{C_{r}}{r + 1} = \frac{1}{n + 1}\binom{n + 1}{r + 1}. Products of coefficients (Vandermonde). Comparing the coefficient of xkx^{k} on both sides of (1+x)m(1+x)n=(1+x)m+n(1 + x)^{m}(1 + x)^{n} = (1 + x)^{m+n} gives ∑j(mj)(nk−j)=(m+nk)\sum_{j}\binom{m}{j}\binom{n}{k - j} = \binom{m + n}{k}. Special cases: C02+C12+⋯+Cn2=(2nn)C_{0}^{2} + C_{1}^{2} + \cdots + C_{n}^{2} = \binom{2n}{n} (use (nj)=(nn−j)\binom{n}{j} = \binom{n}{n - j} and take k=nk = n). C0Cr+C1Cr+1+⋯+Cn−rCn=(2nn−r)C_{0}C_{r} + C_{1}C_{r+1} + \cdots + C_{n-r}C_{n} = \binom{2n}{n - r}. The alternating version C02−C12+C22−⋯C_{0}^{2} - C_{1}^{2} + C_{2}^{2} - \cdots comes from (1−x)n(1+x)n=(1−x2)n(1 - x)^{n}(1 + x)^{n} = (1 - x^{2})^{n}: it equals 00 for odd nn and (−1)n/2(nn/2)(-1)^{n/2}\binom{n}{n/2} for even nn. Choosing a technique. Look at the general term of the sum. A factor rr (or a polynomial in rr) suggests differentiation or absorption. A factor 1r+1\frac{1}{r + 1} (or 1r+k\frac{1}{r + k}) suggests integration. A product of two binomial coefficients whose lower indices add to a constant suggests Vandermonde. A pattern such as every other or every third coefficient suggests substituting roots of unity. Key Takeaways: - ∑Cr=2n\sum C_{r} = 2^{n}; even-index sum == odd-index sum =2n−1= 2^{n-1}. - ∑rCr=n2n−1\sum rC_{r} = n2^{n-1}; ∑r(r−1)Cr=n(n−1)2n−2\sum r(r - 1)C_{r} = n(n - 1)2^{n-2}; ∑r2Cr=n(n+1)2n−2\sum r^{2}C_{r} = n(n + 1)2^{n-2}. - ∑Crr+1=2n+1−1n+1\sum\frac{C_{r}}{r + 1} = \frac{2^{n+1} - 1}{n + 1}; ∑(−1)rCrr+1=1n+1\sum\frac{(-1)^{r}C_{r}}{r + 1} = \frac{1}{n + 1}. - Vandermonde: ∑j(mj)(nk−j)=(m+nk)\sum_{j}\binom{m}{j}\binom{n}{k - j} = \binom{m + n}{k}; ∑Cr2=(2nn)\sum C_{r}^{2} = \binom{2n}{n}. - rCr=n(n−1r−1)rC_{r} = n\binom{n - 1}{r - 1} and Crr+1=1n+1(n+1r+1)\frac{C_{r}}{r + 1} = \frac{1}{n + 1}\binom{n + 1}{r + 1} avoid calculus. - Roots of unity filter coefficients by residue class of rr. - Pitfall: in ∑CrCr+k\sum C_{r}C_{r+k}, the lower indices must be arranged to add to a constant before applying Vandermonde. Brain-Teaser 1: Evaluate S=∑r=0n(r+1)CrS = \sum_{r=0}^{n}(r + 1)C_{r} and T=∑r=0nCr(r+1)(r+2)T = \sum_{r=0}^{n}\frac{C_{r}}{(r + 1)(r + 2)}. Solution: S=∑rCr+∑Cr=n2n−1+2n=(n+2)2n−1S = \sum rC_{r} + \sum C_{r} = n2^{n-1} + 2^{n} = (n + 2)2^{n-1}. For TT, apply the absorption step twice: Crr+1=1n+1(n+1r+1)\frac{C_{r}}{r + 1} = \frac{1}{n + 1}\binom{n + 1}{r + 1} and 1r+2(n+1r+1)=1n+2(n+2r+2)\frac{1}{r + 2}\binom{n + 1}{r + 1} = \frac{1}{n + 2}\binom{n + 2}{r + 2}. So T=1(n+1)(n+2)∑r=0n(n+2r+2)=2n+2−1−(n+2)(n+1)(n+2)=2n+2−n−3(n+1)(n+2)T = \frac{1}{(n + 1)(n + 2)}\sum_{r=0}^{n}\binom{n + 2}{r + 2} = \frac{2^{n+2} - 1 - (n + 2)}{(n + 1)(n + 2)} = \frac{2^{n+2} - n - 3}{(n + 1)(n + 2)}. The missing terms (n+20)\binom{n + 2}{0} and (n+21)\binom{n + 2}{1} must be subtracted from 2n+22^{n+2}, which is the step most often forgotten. Brain-Teaser 2: Find C0C2+C1C3+C2C4+⋯+Cn−2CnC_{0}C_{2} + C_{1}C_{3} + C_{2}C_{4} + \cdots + C_{n-2}C_{n}. Solution: Rewrite the second factor using symmetry: Cr+2=(nn−r−2)C_{r+2} = \binom{n}{n - r - 2}. Then each term is (nr)(nn−2−r)\binom{n}{r}\binom{n}{n - 2 - r}, whose lower indices add to n−2n - 2. By Vandermonde, the sum is (2nn−2)\binom{2n}{n - 2}. The whole trick is to flip one coefficient so that the lower indices sum to a constant, which is the pattern Vandermonde needs. Brain-Teaser 3: Find C0−C2+C4−C6+⋯C_{0} - C_{2} + C_{4} - C_{6} + \cdots for n=10n = 10. Solution: Substitute x=ix = i: (1+i)10=∑Crir(1 + i)^{10} = \sum C_{r}i^{r}. The real part of the right side is C0−C2+C4−C6+C8−C10C_{0} - C_{2} + C_{4} - C_{6} + C_{8} - C_{10}. The left side is (2eiπ/4)10=32ei⋅10π/4=32eiπ/2=32i(\sqrt{2}e^{i\pi/4})^{10} = 32e^{i \cdot 10\pi/4} = 32e^{i\pi/2} = 32i, whose real part is 00. So the sum is 00. (Check: 1−45+210−210+45−1=01 - 45 + 210 - 210 + 45 - 1 = 0.) The imaginary part gives C1−C3+C5−C7+C9=32C_{1} - C_{3} + C_{5} - C_{7} + C_{9} = 32. Complex substitutions filter coefficients by the remainder of rr modulo 44.