Mathematics Cheat Sheet

Definitions Series
iff ∃ positive 𝑐, 𝑛0 such that 0 ≤ 𝑛
𝑛(𝑛 + 1) 𝑛 2 𝑛(𝑛 + 1)(2𝑛 + 1) 𝑛 3 𝑛2 (𝑛 + 1)2
𝑓(𝑛) = 𝑂(𝑔(𝑛))
𝑓(𝑛) ≤ 𝑐𝑔(𝑛) ∀𝑛 ≥ 𝑛0 . ∑𝑖 = 2
,∑𝑖 =
6
,∑𝑖 =
4
.
𝑖=1 𝑖=1 𝑖=1
iff ∃ positive 𝑐, 𝑛0 such that 𝑓(𝑛) ≥
𝑓(𝑛) = Ω(𝑔(𝑛)) In general:
𝑐𝑔(𝑛) ≥ 0 ∀𝑛 ≥ 𝑛0 .
𝑛 𝑛
iff 𝑓(𝑛) = 𝑂(𝑔(𝑛)) and 𝑓(𝑛) = 1
𝑓(𝑛) = Θ(𝑔(𝑛)) ∑ 𝑖 𝑚 = 𝑚 + 1 [(𝑛 + 1)𝑚+1 − 1 − ∑ ((𝑖 + 1)𝑚+1 − 𝑖 𝑚+1 − (𝑚 + 1)𝑖 𝑚 )]
Ω(𝑔(𝑛)). 𝑖=1 𝑖=1
𝑛−1 𝑚
1 𝑚+1
𝑓(𝑛) = 𝑜(𝑔(𝑛)) iff lim𝑛→∞ 𝑓(𝑛)/𝑔(𝑛) = 0. ∑ 𝑖 𝑚 = 𝑚 + 1 ∑ ( 𝑘 ) 𝐵𝑘 𝑛𝑚+1−𝑘 .
𝑖=1 𝑘=0
lim 𝑎 =𝑎 iff ∀𝜖 > 0, ∃𝑛0 such that |𝑎 𝑛 −
𝑛→∞ 𝑛 𝑎| < 𝜖, ∀𝑛 ≥ 𝑛0 . Geometric series:
𝑛 ∞ ∞
least 𝑏 ∈ ℝ such that 𝑏 ≥ 𝑠, ∀𝑠 ∈ 𝑐 𝑛+1 − 1 1 𝑐
sup 𝑆 ∑ 𝑐𝑖 = 𝑐−1
, 𝑐 ≠ 1, ∑ 𝑐𝑖 = 1 − 𝑐 , ∑ 𝑐𝑖 = 1 − 𝑐 , |𝑐| < 1,
𝑆. 𝑖=0 𝑖=0 𝑖=1
𝑛 ∞
greatest 𝑏 ∈ ℝ such that 𝑏 ≤ 𝑠, 𝑛𝑐 𝑛+2 − (𝑛 + 1)𝑐 𝑛+1 + 𝑐 𝑐
inf 𝑆 ∑ 𝑖𝑐 𝑖 = , 𝑐 ≠ 1, ∑ 𝑖𝑐 𝑖 = , |𝑐| < 1.
∀𝑠 ∈ 𝑆.
𝑖=0
(𝑐 − 1)2 𝑖=0
(1 − 𝑐)2
lim inf 𝑎 𝑛 lim inf{𝑎 𝑖 ∣ 𝑖 ≥ 𝑛, 𝑖 ∈ ℕ}. Harmonic series:
𝑛→∞ 𝑛→∞

lim sup 𝑎 𝑛 𝑛 𝑛
lim sup{𝑎 𝑖 ∣ 𝑖 ≥ 𝑛, 𝑖 ∈ ℕ}. 1 𝑛(𝑛 + 1) 𝑛(𝑛 − 1)
𝑛→∞ 𝑛→∞ 𝐻𝑛 = ∑ , ∑ 𝑖𝐻𝑖 = 𝐻𝑛 − .
𝑖=1
𝑖 𝑖=1
2 4
𝑛 𝑛
𝑛 Combinations: Size 𝑘 subsets of a 𝑖 𝑛+1 1
(𝑘) size 𝑛 set. ∑ 𝐻𝑖 = (𝑛 + 1)𝐻𝑛 − 𝑛, ∑ ( 𝑚 ) 𝐻𝑖 = ( 𝑚 + 1 ) (𝐻𝑛+1 − 𝑚 + 1 ) .
𝑖=1 𝑖=1

Stirling numbers (st kind): Ar- 𝑛

𝑛 𝑛 𝑛! 𝑛 𝑛 𝑛
[𝑘] rangements of an 𝑛 element set 1. ( ) = 2. ∑ ( ) = 2𝑛 , 3. ( ) = (
𝑛 − 𝑘)
, ,
𝑘 (𝑛 − 𝑘)!𝑘! 𝑘 𝑘
into 𝑘 cycles. 𝑘=0
𝑛 𝑛 𝑛−1 𝑛 𝑛−1 𝑛−1
4. ( ) = ( 5. ( ) = ( +
𝑘 𝑘 − 1) 𝑘 ) (𝑘 − 1)
Stirling numbers (nd kind): Par- , ,
𝑛 𝑘 𝑘
[𝑘] titions of an 𝑛 element set into 𝑘 𝑛
𝑛 𝑚 𝑛 𝑛−𝑘 𝑟+𝑘 𝑟+𝑛+1
non-empty sets. 6. ( ) ( ) = ( ) ( )=(
𝑚 − 𝑘)
, 7. ∑ ( ),
𝑚 𝑘 𝑘 𝑘=0
𝑘 𝑛
st order Eulerian numbers: 𝑛
𝑘 𝑛+1 𝑛
𝑟 𝑠 𝑟+𝑠
𝑛 8. ∑ ( ) = ( 9. ∑ ( ) ( ) = ( 𝑛 ),
𝑚 + 1)
⟨𝑘⟩ Permutations 𝜋1 𝜋2 … 𝜋𝑛 on ,
𝑚
𝑘=0 𝑘=0
𝑘 𝑛 − 𝑘
{1, 2, … , 𝑛} with 𝑘 ascents.
𝑛 𝑘−𝑛−1 𝑛 𝑛
10. ( ) = (−1)𝑘 ( ), 11. [ ] = [ ] = 1,
𝑘 𝑘 1 𝑛
𝑛
⟪𝑘⟫ nd order Eulerian numbers. 𝑛 𝑛 𝑛−1 𝑛−1
12. [ ] = 2𝑛−1 − 1, 13. [ ] = 𝑘 [ +
𝑘 ] [𝑘 − 1]
,
2 𝑘
Catalan Numbers: Binary trees
𝐶𝑛
with 𝑛 + 1 vertices.

𝑛 𝑛 𝑛 𝑛 𝑛 𝑛 𝑛−1 𝑛−1
14. [ ] = (𝑛 − 1)!, 15. [ ] = (𝑛 − 1)!𝐻𝑛−1 , 16. [ ] = 1, 17. [ ] ≥ [ ], 18. [ ] = (𝑛 − 1) [ +
𝑘 ] [𝑘 − 1]
,
1 2 𝑛 𝑘 𝑘 𝑘
𝑛
𝑛 𝑛 𝑛 𝑛 1 2𝑛 𝑛 𝑛
19. [ = = 20. ∑ [ ] = 𝑛!, 21. 𝐶𝑛 = 22. ⟨ ⟩ = ⟨ = 1,
𝑛 − 1] [𝑛 − 1] (2) 𝑛+1( 𝑛 ) 𝑛 − 1⟩
, ,
𝑘=0
𝑘 0
𝑛 𝑛 𝑛 𝑛−1 𝑛−1 0 1 if k=,
23. ⟨ ⟩ = ⟨ 24. ⟨ ⟩ = (𝑘 + 1) ⟨ + (𝑛 − 𝑘) ⟨ 25. ⟨ ⟩ = {
𝑛 − 1 − 𝑘⟩ 𝑘 ⟩ 𝑘 − 1⟩
, , ,
𝑘 𝑘 𝑘 0 otherwise
𝑛
𝑛 𝑛 𝑛+1 𝑛 𝑥+𝑘
26. ⟨ ⟩ = 2𝑛 − 𝑛 − 1, 27. ⟨ ⟩ = 3𝑛 − (𝑛 + 1)2𝑛 + ( 28. 𝑥𝑛 = ∑ ⟨ ⟩ (
2 ) 𝑛 )
, ,
1 2 𝑘=0
𝑘
𝑚 𝑛 𝑛
𝑛 𝑛+1 𝑛 𝑛 𝑘 𝑛 𝑛 𝑛−𝑘
29. ⟨ ⟩ = ∑ (
𝑚 𝑘 ) (𝑚 + 1 − 𝑘)𝑛 (−1)𝑘 , 30. 𝑚! [ ] = ∑ ⟨ ⟩ (
𝑚 𝑘 𝑛 − 𝑚)
, 31. ⟨ ⟩ = ∑ [ ] (
𝑚 𝑘 𝑚 ) (−1)𝑛−𝑘−𝑚 𝑘!,
𝑘=0 𝑘=0 𝑘=0
𝑛
𝑛 𝑛 𝑛 𝑛−1 𝑛−1 𝑛 (2𝑛)𝑛
32. ⟪ ⟫ = 1, 33. ⟪ ⟫ = 0 for 𝑛 ≠ 0, 34. ⟪ ⟫ = (𝑘 + 1) ⟪ + (2𝑛 − 1 − 𝑘) ⟪ 35. ∑ ⟪ ⟫ =
𝑘 ⟫ 𝑘 − 1⟫
, ,
0 𝑛 𝑘 𝑘=0
𝑘 2𝑛
𝑛 𝑛
𝑥 𝑛 𝑥+𝑛−1−𝑘 𝑛+1 𝑛 𝑘 𝑘
36. [ ] = ∑ ⟪ ⟫( ), 37. [ ] = ∑ ( ) [ ] = ∑ [ ] (𝑚 + 1)𝑛−𝑘 ,
𝑥−𝑛 𝑘=0
𝑘 2𝑛 𝑚+1 𝑘
𝑘 𝑚 𝑘=0
𝑚


 Mathematics Cheat Sheet
Identities Cont. Trees
𝑛 𝑛 𝑛 Every tree with 𝑛 vertices has
𝑛+1 𝑛 𝑘 𝑘 1 𝑘 𝑥 𝑛 𝑥+𝑘
38. [ ] = ∑ [ ] ( ) = ∑ [ ] 𝑛𝑛−𝑘 = 𝑛! ∑ [ ], ] = ∑ ⟪ ⟫( 39. [
), 𝑛 − 1 edges.
𝑚+1 𝑘
𝑘 𝑚 𝑘=0
𝑚 𝑘=0
𝑘! 𝑚 𝑥−𝑛 𝑘=0
𝑘 2𝑛
𝑛 𝑛 𝑘+1 𝑛 𝑛 + 1 𝑘 Kraft inequality: If the depths
40. [ ] = ∑ ( ) [ ] (−1)𝑛−𝑘 , 41. [ ] = ∑ [ ] ( ) (−1)𝑚−𝑘 , of the leaves of a binary tree
𝑚 𝑘
𝑘 𝑚+1 𝑚 𝑘
𝑘+1 𝑚
are 𝑑1 , … , 𝑑𝑛 :
𝑚 𝑚
𝑚+𝑛+1 𝑛+𝑘 𝑚+𝑛+1 𝑛+𝑘
42. [
𝑚 ] = ∑ 𝑘 [ 𝑘 ], 43. [
𝑚 ] = ∑ 𝑘(𝑛 + 𝑘) [ 𝑘 ], 𝑛
𝑘=0 𝑘=0 ∑ 2−𝑑𝑖 ≤ 1,
𝑛 𝑛+1 𝑘 𝑛 𝑛+1 𝑘 𝑖=1
44. ( ) = ∑ [
𝑚 𝑘 + 1 ] [ 𝑚 ] (−1)𝑚−𝑘 , 45. (𝑛 − 𝑚)! ( 𝑚 ) = ∑ [ 𝑘 + 1 ] [ 𝑚 ] (−1)𝑚−𝑘 , for 𝑛 ≥ 𝑚,
𝑘 𝑘 and equality holds only if every
𝑛 𝑚−𝑛 𝑚+𝑛 𝑚+𝑘 𝑛 𝑚−𝑛 𝑚+𝑛 𝑚+𝑘 internal node has  sons.
46. [ = ] = ∑ ( 𝑚 + 𝑘 ) ( 𝑛 + 𝑘 ) [ 𝑘 ],
𝑛 − 𝑚] ∑ ( 𝑚 + 𝑘 ) ( 𝑛 + 𝑘 ) [ 𝑘 ], 47. [
𝑘
𝑛 − 𝑚 𝑘
𝑛 ℓ+𝑚 𝑘 𝑛−𝑘 𝑛 𝑛 ℓ+𝑚 𝑘 𝑛−𝑘 𝑛
48. [ = = [ℓ][ 𝑚 ](𝑘).
ℓ + 𝑚] ( ℓ ) ∑
49. [
𝑘
[ ℓ ] [ 𝑚 ] ( 𝑘 ), ℓ + 𝑚] ( ℓ ) ∑𝑘

Recurrences
Master method: Generating functions:
1(𝑇(𝑛) − 3𝑇(𝑛/2) = 𝑛) . Multiply both sides of the equation by
𝑇(𝑛) = 𝑎𝑇(𝑛/𝑏)+𝑓(𝑛), 𝑎 ≥ 1, 𝑏 > 1
3(𝑇(𝑛/2) − 3𝑇(𝑛/4) = 𝑛/2) 𝑥𝑖 .
If ∃𝜖 > 0 such that 𝑓(𝑛) = 𝑂(𝑛log𝑏 𝑎−𝜖 ) . Sum both sides over all 𝑖 for which the
⋮ ⋮ ⋮
then equation is valid.
log2 𝑛−1
3 (𝑇(2) − 3𝑇(1) = 2) . Choose a generating function 𝐺(𝑥). Usu-
𝑇(𝑛) = Θ(𝑛log𝑏 𝑎 ). ∞
Let 𝑚 = log2 𝑛. Summing the left side we get ally 𝐺(𝑥) = ∑𝑖=0 𝑥𝑖 𝑔𝑖 .
If 𝑓(𝑛) = Θ(𝑛log𝑏 𝑎 ) then
𝑇(𝑛) − 3𝑚 𝑇(1) = 𝑇(𝑛) − 3𝑚 = 𝑇(𝑛) − 𝑛𝑘 where . Rewrite the equation in terms of the gen-
𝑇(𝑛) = Θ(𝑛log𝑏 𝑎 log2 𝑛). 𝑘 = log2 3 ≈ 1.58496. Summing the right side erating function 𝐺(𝑥).
we get . Solve for 𝐺(𝑥).
If ∃𝜖 > 0 such that 𝑓(𝑛) = Ω(𝑛log𝑏 𝑎+𝜖 ), . The coefficient of 𝑥𝑖 in 𝐺(𝑥) is 𝑔𝑖 .
𝑚−1 𝑚−1 𝑖
and ∃𝑐 < 1 such that 𝑎𝑓(𝑛/𝑏) ≤ 𝑐𝑓(𝑛) for 𝑛 3
large 𝑛, then ∑ 2𝑖 3𝑖 = 𝑛 ∑ ( 2 ) . Example:
𝑖=0 𝑖=0 𝑔𝑖+1 = 2𝑔𝑖 + 1, 𝑔0 = 0.
𝑇(𝑛) = Θ(𝑓(𝑛)). 3
Let 𝑐 = . Then we have Multiply and sum:
Substitution (example): Consider the fol- 2
lowing recurrence 𝑚−1
𝑐𝑚 − 1 ∑ 𝑔𝑖+1 𝑥𝑖 = ∑ 2𝑔𝑖 𝑥𝑖 + ∑ 𝑥𝑖 .
𝑛∑ 𝑐𝑖 = 𝑛(
𝑐−1 )
𝑖≥0 𝑖≥0 𝑖≥0
𝑖
𝑇𝑖+1 = 22 ⋅ 𝑇𝑖2 , 𝑇1 = 2. 𝑖=0 𝑖
We choose 𝐺(𝑥) = ∑𝑖≥0 𝑥 𝑔𝑖 .
Rewrite in terms
Note that 𝑇𝑖 is always a power of two. Let = 2𝑛(𝑐 log2 𝑛 − 1) of 𝐺(𝑥):
𝑡𝑖 = log2 𝑇𝑖 . Then we have = 2𝑛(𝑐 (𝑘−1) log𝑐 𝑛 − 1) 𝐺(𝑥) − 𝑔0
= 2𝐺(𝑥) + ∑ 𝑥𝑖 .
𝑡𝑖+1 = 2𝑖 + 2𝑡𝑖 , 𝑡1 = 1. = 2𝑛𝑘 − 2𝑛, 𝑥
𝑖≥0
Let 𝑢𝑖 = 𝑡𝑖 /2𝑖 . Dividing both sides of the and so 𝑇(𝑛) = 3𝑛𝑘 − 2𝑛.
Full history recurrences Simplify:
previous equation by 2𝑖+1 we get can often be changed to limited history ones (ex- 𝐺(𝑥) 1
= 2𝐺(𝑥) + .
𝑡𝑖+1 2𝑖 𝑡 ample): Consider 𝑥 1−𝑥
𝑖+1 = 𝑖+1 + 𝑖𝑖 .
2 2 2 𝑖−1 Solve for 𝐺(𝑥):
Substituting we find 𝑇𝑖 = 1 + ∑ 𝑇𝑗 , 𝑇0 = 1. 𝑥
𝑗=0 𝐺(𝑥) = .
1 1 (1 − 𝑥)(1 − 2𝑥)
𝑢𝑖+1 = + 𝑢𝑖 , 𝑢1 = , Note that
2 2 Expand this using partial fractions:
which is simply 𝑢𝑖 = 𝑖/2. So we find that 𝑖
𝑖−1 𝑇𝑖+1 = 1 + ∑ 𝑇𝑗 . 2 1
𝐺(𝑥) = 𝑥 ( −
𝑇𝑖 has the closed form 𝑇𝑖 = 2𝑖2 . 𝑗=0 1 − 2𝑥 1 − 𝑥 )
Summing factors (example): Consider the
Subtracting we find
following recurrence = 𝑥 ( 2 ∑ 2𝑖 𝑥 𝑖 − ∑ 𝑥 𝑖 )
𝑖 𝑖−1 𝑖≥0 𝑖≥0
𝑇(𝑛) = 3𝑇(𝑛/2) + 𝑛, 𝑇(1) = 1. 𝑇𝑖+1 − 𝑇𝑖 = 1 + ∑ 𝑇𝑗 − 1 − ∑ 𝑇𝑗
=∑ 𝑖+1 𝑖+1
(2 − 1)𝑥 .
Rewrite so that all terms involving 𝑇 are on 𝑗=0 𝑗=0
𝑖≥0
the left side = 𝑇𝑖 .
So 𝑔𝑖 = 2𝑖 − 1.
𝑇(𝑛) − 3𝑇(𝑛/2) = 𝑛. And so 𝑇𝑖+1 = 2𝑇𝑖 = 2𝑖+1 .
Now expand the recurrence, and choose a
factor which makes the left side “telescope”


 Mathematics Cheat Sheet
1+√5 1−√5
𝜋 ≈ 3,14159, 𝑒 ≈ 2,71828, 𝛾 ≈ 0,57721, 𝜙= ≈ 1,61803, 𝜙̂ = ≈ −0,61803
2 2
𝑖 2𝑖 𝑝𝑖 General Probability
   Bernoulli Numbers (𝐵𝑖 = 0, odd 𝑖 ≠ 1): Continuous distributions:
   1 1 1 If
𝐵0 = 1, 𝐵1 = − , 𝐵2 = , 𝐵4 = − , 𝑏
2 6 30
   1 1 5 ℙ[𝑎 < 𝑋 < 𝑏] = ∫ 𝑝(𝑥) 𝑑𝑥,
𝐵6 = , 𝐵8 = − , 𝐵10 = . 𝑎
   42 30 66 then 𝑝 is the probability density function of 𝑋. If
   Change of base, quadratic formula: ℙ[𝑋 < 𝑎] = 𝑃(𝑎),
log𝑎 𝑥 then 𝑃 is the distribution function of 𝑋. If 𝑃 and 𝑝
   −𝑏 ± √𝑏2 − 4𝑎𝑐
log𝑏 𝑥 = , . both exist then 𝑎
   log𝑎 𝑏 2𝑎
𝑃(𝑎) = ∫ 𝑝(𝑥) 𝑑𝑥.
Euler’s number 𝑒: −∞
  
1 1 1 1 Expectation:
   𝑒=1+ + + + +⋯
2 6 24 120
𝑥 𝑛 If 𝑋 is discrete
    lim (1 + ) = 𝑒𝑥 . 𝔼[𝑔(𝑋)] = ∑ 𝑔(𝑥)ℙ[𝑋 = 𝑥].
𝑛→∞ 𝑛
    1 𝑛 1 𝑛+1 𝑥
(1 + 𝑛 ) < 𝑒 < (1 + 𝑛 ) . If 𝑋 continuous then
∞ ∞
   
1 𝑛 𝑒 11𝑒 1 𝔼[𝑔(𝑋)] = ∫ 𝑔(𝑥)𝑝(𝑥) 𝑑𝑥 = ∫ 𝑔(𝑥) 𝑑𝑃(𝑥).
(1 + 𝑛 ) = 𝑒 − 2𝑛 + −𝑂( 3). −∞ −∞
    24𝑛2 𝑛
Variance, standard deviation:
    Harmonic numbers:
3 11 25 137 49 363 761 7129 Var[𝑋] = 𝔼[𝑋 2 ] − 𝔼[𝑋]2 ,
    1, , , , , , , , ,…
2 6 12 60 20 140 280 2520 𝜎 = √Var[𝑋].
   
ln 𝑛 < 𝐻𝑛 < ln 𝑛 + 1, For events 𝐴 and 𝐵:
   
1 ℙ[𝐴 ∨ 𝐵] = ℙ[𝐴] + ℙ[𝐵] − ℙ[𝐴 & 𝐵]
    𝐻𝑛 = ln 𝑛 + 𝛾 + 𝑂( ).
𝑛
ℙ[𝐴 & 𝐵] = ℙ[𝐴] ⋅ ℙ[𝐵],
    Factorial, Stirling’s approximation:
iff 𝐴 and 𝐵 are independent.
     , , , , , , , , , …
𝑛 𝑛 1 ℙ[𝐴 & 𝐵]
     𝑛! = √2𝜋𝑛( ) (1 + Θ( )). ℙ[𝐴|𝐵] =
𝑒 𝑛 ℙ[𝐵]
    
Ackermann’s function and inverse: For random variables 𝑋 and 𝑌:
     𝑗 𝔼[𝑋 ⋅ 𝑌] = 𝔼[𝑋] ⋅ 𝔼[𝑌],
{ 2
⎧ i=
     𝑎(𝑖, 𝑗) = ⎨ 𝑎(𝑖 − 1, 2) j= if 𝑋 and 𝑌 are independent.
     { 𝑎(𝑖 − 1, 𝑎(𝑖, 𝑗 − 1))
⎩ i,j ≥ 𝔼[𝑋 + 𝑌] = 𝔼[𝑋] + 𝔼[𝑌],
     𝛼(𝑖) = min{𝑗 ∣ 𝑎(𝑗, 𝑗) ≥ 𝑖}. 𝔼[𝑐𝑋] = 𝑐𝔼[𝑋].
     Bayes’ theorem:
     Binomial distribution: ℙ[𝐵|𝐴𝑖 ]ℙ[𝐴𝑖 ]
ℙ[𝐴𝑖 |𝐵] = 𝑛 .
𝑛 ∑𝑗=1 ℙ[𝐴𝑗 ]ℙ[𝐵|𝐴𝑗 ]
     ℙ[𝑋 = 𝑘] = ( ) 𝑝𝑘 𝑞 𝑛−𝑘 , 𝑞 = 1 − 𝑝,
𝑘 Inclusion-exclusion:
      𝑛
𝑛 𝑘 𝑛−𝑘 𝑛 𝑛
𝔼[𝑋] = ∑ 𝑘( )𝑝 𝑞 = 𝑛𝑝.
     
𝑘=1
𝑘 ℙ[ ⋁𝑋𝑖 ] = ∑ ℙ[𝑋𝑖 ] +
      𝑖=1 𝑖=1
Poisson distribution:
𝑛 𝑘
Pascal’s Triangle 𝑒 −𝜆 𝜆𝑘
ℙ[𝑋 = 𝑘] = , 𝔼[𝑋] = 𝜆. ∑ (−1)𝑘+1 ∑ ℙ[ ⋀ 𝑋𝑖𝑗 ].
 𝑘! 𝑘=2 𝑖𝑖 <⋯<𝑖𝑘 𝑗=1
 Normal (Gaussian) distribution: Moment inequalities:
 1 2 2 1
𝑝(𝑥) = 𝑒 −(𝑥−𝜇) /2𝜎 , 𝔼[𝑋] = 𝜇. ℙ[|𝑋| ≥ 𝜆𝔼[𝑋]] ≤ ,
 √2𝜋𝜎 𝜆
The “coupon collector”: We are given a random 1
ℙ[∣𝑋 − 𝔼[𝑋]∣ ≥ 𝜆 ⋅ 𝜎] ≤ 2 .
 coupon each day, and there are 𝑛 different types of 𝜆
      coupons. The distribution of coupons is uniform. Geometric distribution:
The expected number of days to pass before we to ℙ[𝑋 = 𝑘] = 𝑝𝑞 𝑘−1 , 𝑞 = 1 − 𝑝,
       collect all 𝑛 types is ∞
𝑛𝐻𝑛 . 1
        𝔼[𝑋] = ∑ 𝑘𝑝𝑞 𝑘−1 = .
𝑘=1
𝑝
        
         
          


 Mathematics Cheat Sheet
Trigonometry Matrices More Trig.

Multiplication: 𝐶

→
(0, 1) 𝑛
𝐶 = 𝐴 ⋅ 𝐵, 𝑐𝑖,𝑗 = ∑ 𝑎 𝑖,𝑘 𝑏𝑘,𝑗 . 𝑎
(cos 𝜃, sin 𝜃) 𝑏 ℎ
𝑏 𝐶 (−1, 0) 𝑘=1
𝐴 ← 𝜃 → Determinants:
(1, 0) 𝐴 𝑐 𝐵
𝑐 det 𝐴 ≠ 0 iff 𝐴 is non-singular.
𝑎 Law of cosines:
det 𝐴 ⋅ 𝐵 = det 𝐴 ⋅ det 𝐵,
𝐵 (0, −1) 𝑛
←
det 𝐴 = ∑ ∏ sign(𝜋)𝑎 𝑖,𝜋(𝑖) . 𝑐 2 = 𝑎2 + 𝑏2 − 2𝑎𝑏 cos 𝐶.
𝜋 𝑖=1 Area:
Pythagorean theorem: 2 × 2 and 3 × 3 determinant: 1
𝐴 = ℎ𝑐,
𝐶 2 = 𝐴2 + 𝐵 2 . 𝑎 𝑏
2
[𝑐 = 𝑎𝑑 − 𝑏𝑐, 1
Definitions: 𝑑] = 𝑎𝑏 sin 𝐶,
2
sin 𝑎 = 𝐴/𝐶, cos 𝑎 =𝐵/𝐶, 2
𝑐 sin 𝐴 sin 𝐵
⎡𝑎 𝑏 𝑐⎤
𝑏 𝑐 𝑎 𝑐 𝑎 𝑏 = .
csc 𝑎 = 𝐶/𝐴, sec 𝑎 =𝐶/𝐵, ⎢𝑑 𝑒 𝑓⎥ = 𝑔 [ − ℎ[ + 𝑖[ 2 sin 𝐶
𝑒 𝑓] 𝑑 𝑓] 𝑑 𝑒]
⎣𝑔 ℎ 𝑖⎦
Area, radius of inscribed circle: = 𝑎𝑒𝑖 + 𝑏𝑓𝑔 + 𝑐𝑑ℎ − 𝑐𝑒𝑔 − 𝑓ℎ𝑎 − 𝑖𝑏𝑑. Heron’s formula:
1 𝐴𝐵 Permanents: 𝐴 = √𝑠 ⋅ 𝑠 𝑎 ⋅ 𝑠 𝑏 ⋅ 𝑠 𝑐 ,
𝐴𝐵, .
2 𝐴+𝐵+𝐶 𝑛 1
𝑠 = (𝑎 + 𝑏 + 𝑐),
perm 𝐴 = ∑ ∏ 𝑎 𝑖,𝜋(𝑖) . 2
Identities:
𝜋 𝑖=1 𝑠𝑎 = 𝑠 − 𝑎,
1 1
sin 𝑥 = , cos 𝑥 = , 𝑠𝑏 = 𝑠 − 𝑏,
csc 𝑥 sec 𝑥 Hyperbolic Functions
1 𝑠𝑐 = 𝑠 − 𝑐.
tan 𝑥 = , sin2 𝑥 + cos2 𝑥 = 1, Definitions:
cot 𝑥
𝑒 𝑥 −𝑒−𝑥 𝑒𝑥 +𝑒 −𝑥
1 + tan2 𝑥 = sec2 𝑥, 1 + cot2 𝑥 = csc2 𝑥, sinh 𝑥 = , cosh 𝑥 = , More identities:
2 2
𝑒 𝑥 −𝑒−𝑥 1
𝜋 tanh 𝑥 = , csch 𝑥 = , 𝑥 1 − cos 𝑥
sin 𝑥 = cos ( − 𝑥), sin 𝑥 = sin(𝜋 − 𝑥), 𝑒 𝑥 +𝑒 −𝑥 sinh 𝑥 sin =√ ,
2 1 1 2 2
𝜋 sech 𝑥 = , coth 𝑥 = .
cos 𝑥 = − cos(𝜋 − 𝑥), tan 𝑥 = cot ( − 𝑥), cosh 𝑥 tanh 𝑥 𝑥 1 + cos 𝑥
2 Identities: cos = √ ,
2 2
𝑥
cot 𝑥 = − cot(𝜋 − 𝑥), csc 𝑥 = cot − cot 𝑥, cosh2 𝑥 − sinh2 𝑥 = 1, tanh2 𝑥 + sech2 𝑥 = 1, 𝑥 1 − cos 𝑥
2
tan = √ ,
2 1 + cos 𝑥
sin(𝑥 ± 𝑦) = sin 𝑥 cos 𝑦 ± cos 𝑥 sin 𝑦, coth2 𝑥 − csch2 𝑥 = 1, sinh(−𝑥) = − sinh 𝑥, 1 − cos 𝑥
= ,
sin 𝑥
cos(𝑥 ± 𝑦) = cos 𝑥 cos 𝑦 ∓ sin 𝑥 sin 𝑦, cosh(−𝑥) = cosh 𝑥, tanh(−𝑥) = − tanh 𝑥, sin 𝑥
tan 𝑥 ± tan 𝑦 = ,
1 + cos 𝑥
tan(𝑥 ± 𝑦) = , sinh(𝑥 + 𝑦) = sinh 𝑥 cosh 𝑦 + cosh 𝑥 sinh 𝑦,
1 ∓ tan 𝑥 tan 𝑦 𝑥 1 + cos 𝑥
cot 𝑥 cot 𝑦 ∓ 1 cot = √ ,
cot(𝑥 ± 𝑦) = , 2 1 − cos 𝑥
cot 𝑥 ± cot 𝑦 cosh(𝑥 + 𝑦) = cosh 𝑥 cosh 𝑦 + sinh 𝑥 sinh 𝑦,
1 + cos 𝑥
2 tan 𝑥 = ,
sin 2𝑥 = 2 sin 𝑥 cos 𝑥, sin 2𝑥 = , sinh 2𝑥 = 2 sinh 𝑥 cosh 𝑥, sin 𝑥
1 + tan2 𝑥 sin 𝑥
= ,
cos 2𝑥 = cos2 𝑥 − sin2 𝑥, cos 2𝑥 = 2 cos2 𝑥 − 1, cosh 2𝑥 = cosh2 𝑥 + sinh2 𝑥, cosh 𝑥 + sinh 𝑥 = 𝑒 𝑥 , 1 − cos 𝑥
1 − tan2 𝑥 𝑒 𝑖𝑥 − 𝑒 −𝑖𝑥
cos 2𝑥 = 1 − 2 sin2 𝑥,, cos 2𝑥 = sin 𝑥 = ,
1 + tan2 𝑥
, cosh 𝑥 − sinh 𝑥 = 𝑒 −𝑥 , 2𝑖
2 tan 𝑥 2
cot 𝑥 − 1 𝑒 𝑖𝑥 + 𝑒 −𝑖𝑥
tan 2𝑥 = cot 2𝑥 = cos 𝑥 = ,
1 − tan2 𝑥
,
2 cot 𝑥
, (cosh 𝑥 + sinh 𝑥)𝑛 = cosh 𝑛𝑥 + sinh 𝑛𝑥, 𝑛 ∈ ℤ, 2
𝑒 𝑖𝑥 − 𝑒 −𝑖𝑥
sin(𝑥 + 𝑦) sin(𝑥 − 𝑦) = sin2 𝑥 − sin2 𝑦, 2 sinh2
𝑥
= cosh 𝑥 − 1, 2 cosh2
𝑥
= cosh 𝑥 + 1 tan 𝑥 = −𝑖 𝑖𝑥
𝑒 + 𝑒 −𝑖𝑥
,
2 2
cos(𝑥 + 𝑦) cos(𝑥 − 𝑦) = cos2 𝑥 − sin2 𝑦. 𝑒 2𝑖𝑥 − 1
= −𝑖 2𝑖𝑥 ,
𝑒 +1
𝜃 sin 𝜃 cos 𝜃 tan 𝜃 sinh 𝑖𝑥
Euler’s equation: sin 𝑥 = ,
0 0 1 0 … in mathematics 𝑖
𝑒 𝑖𝑥 = cos 𝑥 + 𝑖 sin 𝑥, 𝑒 𝑖𝜋 + 1 = 0. 𝜋 1 √3 √3 you don’t understand cos 𝑥 = cosh 𝑖𝑥,
6 2 2 3 things, you just get
√2 √2
tanh 𝑖𝑥
𝜋
1 used to them. tan 𝑥 = .
4 2 2
𝑖
v. ©– by Steve Seiden – J. von Neumann
𝜋 √3 1
[email protected] √3
3 2 2
v. ©, port to LaTEX by Alain Aubord 𝜋
[email protected] 1 0 ∞
2


 Mathematics Cheat Sheet
Number Theory Graph Theory
The Chinese remainder theorem: There exists a num- Definitions: Notation:
ber 𝐶 such that: Loop An edge connecting a vertex to itself. 𝐸(𝐺) Edge set
𝐶 ≡ 𝑟1 mod 𝑚 1 Directed Each edge has a direction. 𝑉(𝐺) Vertex set
⋮ ⋮ ⋮ Simple Graph with no loops or multi-edges. 𝑐(𝐺) Number of components
𝐶 ≡ 𝑟𝑛 mod 𝑚 𝑛 Walk A sequence 𝑣0 𝑒1 𝑣1 … 𝑒ℓ 𝑣ℓ . 𝐺[𝑆] Induced subgraph
Trail A walk with distinct edges. deg(𝑣) Degree of 𝑣
Path A trail with distinct vertices. Δ(𝐺) Maximum degree
if 𝑚 𝑖 and 𝑚 𝑗 are relatively prime for 𝑖 ≠ 𝑗. 𝛿(𝐺) Minimum degree
Connected A graph where there exists a path be-
Euler’s function: 𝜙(𝑥) is the number of positive inte- tween any two vertices. 𝜒(𝐺) Chromatic number
𝑛 𝑒
gers less than 𝑥 relatively prime to 𝑥. If ∏𝑖=1 𝑝𝑖 𝑖 is Component A maximal connected subgraph. 𝜒𝐸 (𝐺) Edge chromatic number
the prime factorization of 𝑥 then Tree A connected acyclic graph. 𝐺𝑐 Complement graph
𝑛
𝑒 −1 Free tree A tree with no root. 𝐾𝑛 Complete graph
𝜙(𝑥) = ∏ 𝑝𝑖 𝑖 (𝑝𝑖 − 1).
𝑖=1 DAG Directed acyclic graph. 𝐾𝑛1 ,𝑛2 Complete bipartite graph
Euler’s theorem: If 𝑎 and 𝑏 are relatively prime then Eulerian Graph with a trail visiting each edge ex- r(𝑘, ℓ) Ramsey number
1 ≡ 𝑎𝜙(𝑏) mod 𝑏. actly once.
Hamiltonian Graph with a cycle visiting each vertex Geometry
Fermat’s theorem:
exactly once.
1 ≡ 𝑎𝑝−1 mod 𝑝.
Cut A set of edges whose removal increases the Projective coordinates: triples (𝑥, 𝑦, 𝑧),
The Euclidean algorithm: if 𝑎 > 𝑏 are integers then number of components. not all 𝑥, 𝑦 and 𝑧 zero.
gcd(𝑎, 𝑏) = gcd(𝑎 mod 𝑏, 𝑏). Cut-set A minimal cut. (𝑥, 𝑦, 𝑧) = (𝑐𝑥, 𝑐𝑦, 𝑐𝑧) ∀𝑐 ≠ 0.
𝑛 𝑒
If ∏𝑖=1 𝑝𝑖 𝑖 is the prime factorization of 𝑥 then Cut edge A size  cut. Cartesian Projective
𝑛 𝑒 +1
𝑝 𝑖 −1 k-Connected A graph connected with the removal (𝑥, 𝑦) (𝑥, 𝑦, 1)
𝑆(𝑥) = ∑ 𝑑 = ∏ 𝑖 . of any 𝑘 − 1 vertices.
𝑑|𝑥 𝑖=1
𝑝𝑖 − 1 𝑦 = 𝑚𝑥 + 𝑏 (𝑚, −1, 𝑏)
k-Tough ∀𝑆 ⊆ 𝑉, 𝑆 ≠ ∅ we have 𝑥=𝑐 (1, 0, −𝑐)
Perfect Numbers: 𝑥 is an even perfect number iff 𝑘 ⋅ 𝑐(𝐺 − 𝑆) ≤ |𝑆|.
𝑥 = 2𝑛−1 (2𝑛 − 1) and 2𝑛 − 1 is prime. Distance formula, 𝐿𝑝 and 𝐿∞ metric:
k-Regular A graph where all vertices have degree
2 2
Wilson’s theorem: 𝑛 is a prime iff 𝑘. √(𝑥1 − 𝑥0 ) + (𝑦1 − 𝑦0 ) ,
(𝑛 − 1)! ≡ −1 mod 𝑛. k-Factor A 𝑘-regular spanning subgraph. 1/𝑝
[|𝑥1 − 𝑥0 |𝑝 + |𝑦1 − 𝑦0 |𝑝 ] ,
Matching A set of edges, no two of which are 1/𝑝
Möbius inversion:
adjacent. lim [|𝑥1 − 𝑥0 |𝑝 + |𝑦1 − 𝑦0 |𝑝 ] .
⎧ 1 if 𝑖 = 1. 𝑝→∞
{ 0 if 𝑖 is not square-free. Clique A set of vertices, all of which are adjacent.
𝜇(𝑖) = ⎨ Area of triangle (𝑥0 , 𝑦0 ), (𝑥1 , 𝑦1 ) and
(−1)𝑟 if 𝑖 is the product of Ind. set A set of vertices, none of which are adja-
(𝑥2 , 𝑦2 ):
{ cent.
⎩ 𝑖𝑟 distinct primes. 1 𝑥 − 𝑥0 𝑦1 − 𝑦0
abs ∣ 1 .
𝑥2 − 𝑥0 𝑦2 − 𝑦0 ∣
If Vertex cover A set of vertices which cover all 2
𝐺(𝑎) = ∑ 𝐹(𝑑), edges.
𝑑|𝑎 Planar graph A graph which can be embeded in Angle formed by three points:
then the plane.
𝑎 ∘
𝐹(𝑎) = ∑ 𝜇(𝑑)𝐺( ). Plane graph An embedding of a planar graph.
𝑑|𝑎
𝑑 (𝑥2 , 𝑦2 )
Prime numbers: ℓ2
ln ln 𝑛
𝑝𝑛 = 𝑛 ln 𝑛 + 𝑛 ln ln 𝑛 − 𝑛 + 𝑛 ∑ deg(𝑣) = 2𝑚.
ln 𝑛 𝜃
𝑛 𝑣∈𝑉 ∘ ∘
+ 𝑂 ( ), (0, 0) ℓ1 (𝑥1 , 𝑦1 )
ln 𝑛 If 𝐺 is planar then 𝑛 − 𝑚 + 𝑓 = 2, so
𝑛 𝑛 2!𝑛
𝜋(𝑛) = + +
ln 𝑛 (ln 𝑛)2 (ln 𝑛)3 𝑓 ≤ 2𝑛 − 4, 𝑚 ≤ 3𝑛 − 6. (𝑥1 , 𝑦1 ) ⋅ (𝑥2 , 𝑦2 )
𝑛 cos 𝜃 = .
+ 𝑂( ). ℓ1 ℓ2
(ln 𝑛)4 Any planar graph has a vertex with degree ≤ 5.
Line through two points (𝑥0 , 𝑦0 ) and
(𝑥1 , 𝑦1 ):
𝑥 𝑦 1
∣𝑥0 𝑦0 1∣ = 0.
𝑥1 𝑦1 1
Area of circle, volume of sphere:
4
𝐴 = 𝜋𝑟 2 , 𝑉 = 𝜋𝑟 3 .
3

If I have seen farther than others, it is

because I have stood on the shoulders
of giants.
– Issac Newton


 Mathematics Cheat Sheet
𝜋 Calculus
Wallis’ identity: Derivatives:
2 ⋅ 2 ⋅ 4 ⋅ 4 ⋅ 6 ⋅ 6⋯
𝜋 =2⋅
1⋅3⋅3⋅5⋅5⋅7⋯ 𝑑(𝑐𝑢) 𝑑𝑢 𝑑(𝑢 + 𝑣) 𝑑𝑢 𝑑𝑣 𝑑(𝑢𝑣) 𝑑𝑣 𝑑𝑢
1. =𝑐 , 2. = + , 3. =𝑢 +𝑣 ,
Brouncker’s continued fraction expansion: 𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥
𝜋 12 𝑑𝑢 𝑑𝑣
𝑣 ( ) − 𝑢( )
=1+ 𝑑(𝑢𝑛 ) 𝑑𝑢 𝑑(𝑢/𝑣) 𝑐𝑢
𝑑(𝑒 ) 𝑑𝑢
4
2+
32 4. = 𝑛𝑢𝑛−1 , 5. = 𝑑𝑥 2 𝑑𝑥 , 6. = 𝑐𝑒 𝑐𝑢 ,
2 5 𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑣 𝑑𝑥 𝑑𝑥
2+
72 𝑑(𝑐 𝑢 ) 𝑑𝑢 𝑑(ln 𝑢) 1 𝑑𝑢 𝑑(sin 𝑢) 𝑑𝑢
2+
2+⋯ 7. = (ln 𝑐)𝑐 𝑢 , 8. = , 9. = cos 𝑢 ,
Gregrory’s series: 𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑢 𝑑𝑥 𝑑𝑥 𝑑𝑥
𝜋 1 1 1 1
=1− + − + −⋯ 𝑑(cos 𝑢) 𝑑𝑢 𝑑(tan 𝑢) 𝑑𝑢 𝑑(cot 𝑢) 𝑑𝑢
4 3 5 7 9 10. = − sin 𝑢 , 11. = sec2 𝑢 , 12. = csc2 𝑢 ,
𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥
Newton’s series:
𝜋 1 1 1⋅3 𝑑(sec 𝑢) 𝑑𝑢 𝑑(csc 𝑢) 𝑑𝑢
= + + +⋯ = tan 𝑢 sec 𝑢 , = − cot 𝑢 csc 𝑢 ,
6 2 2 ⋅ 3 ⋅ 2 3 2 ⋅ 4 ⋅ 5 ⋅ 25 13.
𝑑𝑥 𝑑𝑥
14.
𝑑𝑥 𝑑𝑥
Sharp’s series:
𝜋 1 1 1 1 𝑑(arcsin 𝑢) 1 𝑑𝑢 𝑑(arccos 𝑢) −1 𝑑𝑢
= (1 − 1 + 2 − 3 +⋯) 15. = , 16. = ,
6 √3 3 ⋅3 3 ⋅5 3 ⋅7 𝑑𝑥 √1 − 𝑢 𝑑𝑥
2 𝑑𝑥 √1 − 𝑢2 𝑑𝑥
Euler’s series: 𝑑(arctan 𝑢) 1 𝑑𝑢 𝑑(arccot 𝑢) −1 𝑑𝑢
𝜋2 1 1 1 1 1 17. = , 18. = ,
6
= 2 + 2 + 2 + 2 + 2 +⋯
1 2 3 4 5
𝑑𝑥 1 + 𝑢2 𝑑𝑥 𝑑𝑥 1 + 𝑢2 𝑑𝑥
𝜋2 1 1 1 1 1 𝑑(arcsec 𝑢) 1 𝑑𝑢 𝑑(arccsc 𝑢) −1 𝑑𝑢
= + + + 2 + 2 +⋯ 19. = , 20. = ,
12 32 52 𝑑𝑥
8 7 9
𝑢 1 − 𝑢 𝑑𝑥
√ 2 𝑑𝑥 𝑢 1 − 𝑢 𝑑𝑥
√ 2
𝜋2 1 1 1 1 1
= − + − 2 + 2 −⋯ 𝑑(sinh 𝑢) 𝑑𝑢 𝑑(cosh 𝑢) 𝑑𝑢
12 12 22 32 4 5 21. = cosh 𝑢 , 22. = sinh 𝑢 ,
𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥
𝑑(tanh 𝑢) 𝑑𝑢 𝑑(coth 𝑢) 𝑑𝑢
23. = sech2 𝑢 , 24. = − csch2 𝑢 ,
𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥
Partial Fractions 𝑑(sech 𝑢) 𝑑𝑢 𝑑(csch 𝑢) 𝑑𝑢
25. = − sech 𝑢 tanh 𝑢 , 26. = − csch 𝑢 coth 𝑢 ,
𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥
Let 𝑁 (𝑥) and 𝐷(𝑥) be polynomial functions of 𝑥.
We can break down 𝑁 (𝑥)/𝐷(𝑥) using partial fraction 𝑑(arcsinh 𝑢) 1 𝑑𝑢 𝑑(arccosh 𝑢) 1 𝑑𝑢
27. = , 28. = ,
expansion. First, if the degree of 𝑁 is greater than or 𝑑𝑥 √1 + 𝑢2 𝑑𝑥 𝑑𝑥 √𝑢2 − 1 𝑑𝑥
equal to the degree of 𝐷, divide 𝑁 by 𝐷, obtaining 𝑑(arctanh 𝑢) 1 𝑑𝑢 𝑑(arccoth 𝑢) 1 𝑑𝑢
𝑁 (𝑥) 𝑁 ′ (𝑥) 29. = , 30. = 2 ,
= 𝑄(𝑥) + , 𝑑𝑥 1 − 𝑢2 𝑑𝑥 𝑑𝑥 𝑢 − 1 𝑑𝑥
𝐷(𝑥) 𝐷(𝑥)
′
where the degree of 𝑁 is less than that of 𝐷. Second, 𝑑(arcsech 𝑢) −1 𝑑𝑢 𝑑(arccsch 𝑢) −1 𝑑𝑢
31. = , 32. = .
factor 𝐷(𝑥). Use the following rules: For a non- 𝑑𝑥 𝑢√1 − 𝑢 𝑑𝑥
2 𝑑𝑥 |𝑢|√1 + 𝑢 𝑑𝑥
2
repeated factor:
𝑁 (𝑥) 𝐴 𝑁 ′ (𝑥)
= + ,
(𝑥 − 𝑎)𝐷(𝑥) 𝑥−𝑎 𝐷(𝑥) Integrals:
where
𝑁 (𝑥)
𝐴=[ .
𝐷(𝑥) ] 1. ∫ 𝑐𝑢 𝑑𝑥 = 𝑐 ∫ 𝑢 𝑑𝑥, 2. ∫(𝑢 + 𝑣) 𝑑𝑥 = ∫ 𝑢 𝑑𝑥 + ∫ 𝑣 𝑑𝑥,
𝑥=𝑎
For a repeated factor:
𝑚−1 1 𝑛+1 1
𝑁 (𝑥) 𝐴𝑘 𝑁 ′ (𝑥) 3. ∫ 𝑥𝑛 𝑑𝑥 = 𝑥 , 𝑛 ≠ −1, 4. ∫ 𝑑𝑥 = ln 𝑥, 5. ∫ 𝑒𝑥 𝑑𝑥 = 𝑒 𝑥 ,
𝑚 =∑ 𝑚−𝑘 + , 𝑛+1 𝑥
(𝑥 − 𝑎) 𝐷(𝑥) 𝑘=0 (𝑥 − 𝑎) 𝐷(𝑥)
where 𝑑𝑥 𝑑𝑣 𝑑𝑢
1 𝑑 𝑘 𝑁 (𝑥) 6. ∫ = arctan 𝑥, 7. ∫ 𝑢 𝑑𝑥 = 𝑢𝑣 − ∫ 𝑣 𝑑𝑥, 8. ∫ sin 𝑥 𝑑𝑥 = − cos 𝑥,
𝐴𝑘 = . 1 + 𝑥2 𝑑𝑥 𝑑𝑥
𝑘! [ 𝑑𝑥𝑘 ( 𝐷(𝑥) )]
𝑥=𝑎
9. ∫ cos 𝑥 𝑑𝑥 = sin 𝑥, 10. ∫ tan 𝑥 𝑑𝑥 = − ln | cos 𝑥|, 11. ∫ cot 𝑥 𝑑𝑥 = ln | cos 𝑥|,
The reasonable man adapts himself to the world; the
unreasonable persists in trying to adapt the world to 12. ∫ sec 𝑥 𝑑𝑥 = ln | sec 𝑥 + tan 𝑥|, 13. ∫ csc 𝑥 𝑑𝑥 = ln | csc 𝑥 + cot 𝑥|,
himself. Therefore all progress depends on the un-
reasonable. 𝑥 𝑥
14. ∫ arcsin 𝑑𝑥 = arcsin + √𝑎2 − 𝑥2 , 𝑎 > 0,
– George Bernard Shaw 𝑎 𝑎


 Mathematics Cheat Sheet
Calculus Cont.

𝑥 𝑥 𝑥 𝑥 𝑎
15. ∫ arccos 𝑑𝑥 = arccos − √𝑎2 − 𝑥2 , 𝑎 > 0, 16. ∫ arctan 𝑑𝑥 = 𝑥 arctan − ln(𝑎2 + 𝑥2 ), 𝑎 > 0,
𝑎 𝑎 𝑎 𝑎 2

1 1
17. ∫ sin2 (𝑎𝑥)𝑑𝑥 = (𝑎𝑥 − sin(𝑎𝑥) cos(𝑎𝑥)), 18. ∫ cos2 (𝑎𝑥)𝑑𝑥 = (𝑎𝑥 + sin(𝑎𝑥) cos(𝑎𝑥)), 19. ∫ sec2 𝑥 𝑑𝑥 = tan 𝑥,
2𝑎 2𝑎

sin𝑛−1 𝑥 cos 𝑥 𝑛 − 1
20. ∫ csc2 𝑥 𝑑𝑥 = − cot 𝑥, 21. ∫ sin𝑛 𝑥 𝑑𝑥 = − + ∫ sin𝑛−2 𝑥 𝑑𝑥,
𝑛 𝑛
cos𝑛−1 𝑥 sin 𝑥 𝑛 − 1 tan𝑛−1 𝑥
22. ∫ cos𝑛 𝑥 𝑑𝑥 = + ∫ cos𝑛−2 𝑥 𝑑𝑥, 23. ∫ tan𝑛 𝑥 𝑑𝑥 = − ∫ tan𝑛−2 𝑥 𝑑𝑥, 𝑛 ≠ 1,
𝑛 𝑛 𝑛−1
cot𝑛−1 𝑥 tan 𝑥 sec𝑛−1 𝑥 𝑛 − 2
24. ∫ cot𝑛 𝑥 𝑑𝑥 = − − ∫ cot𝑛−2 𝑥 𝑑𝑥, 𝑛 ≠ 1, 25. ∫ sec𝑛 𝑥 𝑑𝑥 = + ∫ sec𝑛−2 𝑥 𝑑𝑥, 𝑛 ≠ 1,
𝑛−1 𝑛−1 𝑛−1
cot 𝑥 csc𝑛−1 𝑥 𝑛 − 2
26. ∫ csc𝑛 𝑥 𝑑𝑥 = − + ∫ csc𝑛−2 𝑥 𝑑𝑥, 𝑛 ≠ 1, 27. ∫ sinh 𝑥 𝑑𝑥 = cosh 𝑥, 28. ∫ cosh 𝑥 𝑑𝑥 = sinh 𝑥,
𝑛−1 𝑛−1
𝑥
29. ∫ tanh 𝑥 𝑑𝑥 = ln | cosh 𝑥|, 30. ∫ coth 𝑥 𝑑𝑥 = ln | sinh 𝑥|, 31. ∫ sech 𝑥 𝑑𝑥 = arctan sinh 𝑥, 32. ∫ csch 𝑥 𝑑𝑥 = ln ∣tanh ∣,
2

1 1 1 1
33. ∫ sinh2 𝑥 𝑑𝑥 = sinh(2𝑥) − 𝑥, 34. ∫ cosh2 𝑥 𝑑𝑥 = sinh(2𝑥) + 𝑥, 35. ∫ sech2 𝑥 𝑑𝑥 = tanh 𝑥,
4 2 4 2

𝑥 𝑥 𝑥 𝑥 𝑎
36. ∫ arcsinh 𝑑𝑥 = 𝑥 arcsinh − √𝑥2 + 𝑎2 , 𝑎 > 0, 37. ∫ arctanh 𝑑𝑥 = 𝑥 arctanh + ln |𝑎2 − 𝑥2 |,
𝑎 𝑎 𝑎 𝑎 2
𝑥 2 2 𝑥
⎧
{ 𝑥 arccosh 𝑎 − √𝑥 + 𝑎 , if arccosh 𝑎 > 0 and 𝑎 > 0,
𝑥 𝑑𝑥
38. ∫ arccosh =⎨ 𝑑𝑥 𝑥 𝑥
, 39. ∫ = ln (𝑥 + √𝑎2 + 𝑥2 ) , 𝑎 > 0,
𝑎 { 2 + 𝑎2 , if arccosh < 0 and 𝑎 > 0, √𝑎 + 𝑥2
2
⎩ 𝑥 arccosh + √𝑥
𝑎 𝑎
𝑑𝑥 1 𝑥 𝑥 𝑎2 𝑥
40. ∫ 2 2 = arctan , 𝑎 > 0, 41. ∫ √𝑎2 − 𝑥2 𝑑𝑥 = √𝑎2 − 𝑥2 + arcsin , 𝑎 > 0,
𝑎 +𝑥 𝑎 𝑎 2 2 𝑎

𝑥 3𝑎4 𝑥 𝑑𝑥 𝑥
42. ∫(𝑎2 − 𝑥2 )3/2 𝑑𝑥 = (5𝑎2 − 2𝑥2 )√𝑎2 − 𝑥2 + arcsin , 𝑎 > 0, 43. ∫ = arcsin , 𝑎 > 0,
8 8 𝑎 √𝑎2 − 𝑥2 𝑎

𝑑𝑥 1 𝑎+𝑥 𝑑𝑥 𝑥 𝑥 𝑎2
44. ∫ = ln 45. ∫ = 46. ∫ √𝑎2 ± 𝑥2 𝑑𝑥 = √𝑎2 ± 𝑥2 ± ln ∣𝑥 + √𝑎2 ± 𝑥2 ∣,
2𝑎 ∣ 𝑎 − 𝑥 ∣
, ,
𝑎2 − 𝑥2 (𝑎2−𝑥 )2 3/2 2 2
𝑎 √𝑎 − 𝑥 2 2 2

𝑑𝑥 𝑑𝑥 1 𝑥 2(3𝑏𝑥 − 2𝑎)(𝑎 + 𝑏𝑥)3/2

47. ∫ = ln ∣𝑥 + √𝑥2 − 𝑎2 ∣ , 𝑎 > 0, 48. ∫ 2 = ln ∣ ∣, 49. ∫ 𝑥√𝑎 + 𝑏𝑥 𝑑𝑥 = ,
√𝑥2 − 𝑎2 𝑎𝑥 + 𝑏𝑥 𝑎 𝑎 + 𝑏𝑥 15𝑏2
√𝑎 + 𝑏𝑥 1 𝑥 1 √𝑎 + 𝑏𝑥 − √𝑎
50. ∫ 𝑑𝑥 = 2√𝑎 + 𝑏𝑥 + 𝑎 ∫ 𝑑𝑥, 51. ∫ 𝑑𝑥 = ln ∣ ∣, 𝑎 > 0,
𝑥 𝑥√𝑎 + 𝑏𝑥 √𝑎 + 𝑏𝑥 √2 √𝑎 + 𝑏𝑥 + √𝑎
√𝑎2 − 𝑥2 𝑎 + √𝑎2 − 𝑥2 1
52. ∫ 𝑑𝑥 = √𝑎2 − 𝑥2 − 𝑎 ln ∣ ∣, 53. ∫ 𝑥√𝑎2 − 𝑥2 𝑑𝑥 = − (𝑎2 − 𝑥2 )3/2 ,
𝑥 𝑥 3

𝑥 𝑎4 𝑥 𝑑𝑥 1 𝑎 + √𝑎2 − 𝑥2
54. ∫ 𝑥2 √𝑎2 − 𝑥2 𝑑𝑥 = (2𝑥2 − 𝑎2 )√𝑎2 − 𝑥2 + arcsin , 𝑎 > 0, 55. ∫ = − ln ∣ ∣,
8 8 𝑎 √𝑎2 − 𝑥2 𝑎 𝑥
𝑥 𝑑𝑥 𝑥2 𝑑𝑥 𝑥 𝑎2 𝑥
56. ∫ = −√𝑎2 − 𝑥2 , 57. ∫ = − √𝑎2 − 𝑥2 + arcsin , 𝑎 > 0,
√𝑎2 − 𝑥2 √𝑎2 − 𝑥2 2 2 𝑎

√𝑎2 + 𝑥2 𝑎 + √𝑎2 + 𝑥2 √𝑥2 − 𝑎2 𝑎

58. ∫ 𝑑𝑥 = √𝑎2 + 𝑥2 − 𝑎 ln ∣ ∣, 59. ∫ 𝑑𝑥 = √𝑥2 − 𝑎2 − 𝑎 arccos , 𝑎 > 0,
𝑥 𝑥 𝑥 |𝑥|

1 𝑑𝑥 1 𝑥
60. ∫ 𝑥√𝑥2 ± 𝑎2 𝑑𝑥 = (𝑥2 ± 𝑎2 )3/2 , 61. ∫ = ln ∣ ∣,
3 𝑥√𝑥2 + 𝑎2 𝑎 𝑎+ √𝑎 2 + 𝑥2


 Mathematics Cheat Sheet
Calculus cont. Finite Calculus

𝑑𝑥 2 1 𝑎 𝑑𝑥 √𝑥2 ± 𝑎2
62. ∫ √𝑥 − 𝑎2 = arccos , 𝑎 > 0, 63. ∫ =∓ , Difference, shift operators:
𝑥 𝑎 |𝑥| 𝑥2 √𝑥2 ± 𝑎2 𝑎2 𝑥
Δ𝑓(𝑥) = 𝑓(𝑥 + 1) − 𝑓(𝑥),
𝑥 𝑑𝑥 2
√𝑥 ± 𝑎 2 (𝑥2 + 𝑎2 )3/2 𝔼𝑓(𝑥) = 𝑓(𝑥 + 1).
64. ∫ = √𝑥2 ± 𝑎2 , 65. ∫ 4 𝑑𝑥 = ∓ ,
√𝑥2 ± 𝑎2 𝑥 3𝑎2 𝑥3
⎧ 1 2𝑎𝑥 + 𝑏 − √𝑏2 − 4𝑎𝑐 Fundamental Theorem:
{
{
ln ∣ ∣ , if 𝑏2 > 4𝑎𝑐, 𝑓(𝑥) = Δ𝐹(𝑥) ⇔ ∑ 𝑓(𝑥)𝛿𝑥 = 𝐹(𝑥) + 𝐶.
2
{ √𝑏 − 4𝑎𝑐 2𝑎𝑥 + 𝑏 + √𝑏 − 4𝑎𝑐 2
𝑑𝑥 𝑏 𝑏−1
66. ∫ 2 =⎨ , ∑ 𝑓(𝑥)𝛿𝑥 = ∑ 𝑓(𝑖).
𝑎𝑥 + 𝑏𝑥 + 𝑐 { 2 2𝑎𝑥 + 𝑏 2 < 4𝑎𝑐,
{ arctan , if 𝑏 𝑎 𝑖=𝑎
{ √4𝑎𝑐 − 𝑏2 √4𝑎𝑐 − 𝑏2
⎩ Differences:
⎧ 1 ln ∣2𝑎𝑥 + 𝑏 + 2√𝑎√𝑎𝑥2 + 𝑏𝑥 + 𝑐∣ , if 𝑎 > 0, Δ(𝑐𝑢) = 𝑐Δ𝑢, Δ(𝑢 + 𝑣) = Δ𝑢 + Δ𝑣,
{ √𝑎
𝑑𝑥 { Δ(𝑢𝑣) = 𝑢Δ𝑣 + 𝔼𝑣Δ𝑢, Δ(𝑥𝑛 ) = 𝑛𝑥𝑛−1 ,
67. ∫ =⎨ 1 −2𝑎𝑥 − 𝑏 ,
√𝑎𝑥2 + 𝑏𝑥 + 𝑐 { arcsin , if 𝑎 < 0, Δ(𝐻𝑥 ) = 𝑥−1 , Δ(2𝑥 ) = 2𝑥 ,
{ √−𝑎 √𝑏2 − 4𝑎𝑐
⎩ Δ(𝑐 𝑥 ) = (𝑐 − 1)𝑐 𝑥 , Δ( 𝑥 ) = ( 𝑥
).
𝑚 𝑚−1
2𝑎𝑥 + 𝑏 4𝑎𝑥 − 𝑏2 𝑑𝑥
68. ∫ √𝑎𝑥2 + 𝑏𝑥 + 𝑐 𝑑𝑥 = √𝑎𝑥2 + 𝑏𝑥 + 𝑐 + ∫ ,
4𝑎 8𝑎 √𝑎𝑥2 + 𝑏𝑥 + 𝑐 Sums:
𝑥 𝑑𝑥 √𝑎𝑥2 + 𝑏𝑥 + 𝑐 𝑏 𝑑𝑥 ∑ 𝑐𝑢 𝛿𝑥 = 𝑐 ∑ 𝑢 𝛿𝑥,
69. ∫ = − ∫ ,
2
√𝑎𝑥 + 𝑏𝑥 + 𝑐 𝑎 2𝑎 2
√𝑎𝑥 + 𝑏𝑥 + 𝑐
∑(𝑢 + 𝑣) 𝛿𝑥 = ∑ 𝑢 𝛿𝑥 + ∑ 𝑣 𝛿𝑥,
⎧ −1 2√𝑐√𝑎𝑥2 + 𝑏𝑥 + 𝑐 + 𝑏𝑥 + 2𝑐
{
{ √𝑐 ln ∣ ∣ , if 𝑐 > 0, ∑ 𝑢Δ𝑣 𝛿𝑥 = 𝑛+1
𝑢𝑣 − ∑ 𝔼𝑣Δ𝑢 𝛿𝑥,
𝑑𝑥 𝑥
70. ∫ =⎨ , 𝑛 𝑥 −1
𝑥√𝑎𝑥2 + 𝑏𝑥 + 𝑐 { 1 𝑏𝑥 + 2𝑐 ∑ 𝑥 𝛿𝑥 = , ∑ 𝑥 𝛿𝑥 = 𝐻𝑥 ,
{ √−𝑐 arcsin , if 𝑐 < 0, 𝑚+1
⎩ |𝑥|√𝑏2 − 4𝑎𝑐 𝑐𝑥 𝑥 𝑥
∑ 𝑐 𝑥 𝛿𝑥 = , ∑ ( 𝑚 ) 𝛿𝑥 = ( 𝑚+1 ) .
1 2 2 𝑐−1
71. ∫ 𝑥3 √𝑥2 + 𝑎2 𝑑𝑥 = ( 𝑥2 − 𝑎 )(𝑥2 + 𝑎2 )3/2 ,
3 15 Falling Factorial Powers:
1 𝑛
72. ∫ 𝑥𝑛 sin(𝑎𝑥) 𝑑𝑥 = − 𝑥𝑛 cos(𝑎𝑥) + ∫ 𝑥𝑛−1 cos(𝑎𝑥) 𝑑𝑥, 𝑥𝑛 = 𝑥(𝑥 − 1) ⋯ (𝑥 − 𝑛 + 1), 𝑛 > 0,
𝑎 𝑎
𝑥0 = 1,
1 𝑛 𝑛
73. ∫ 𝑥𝑛 cos(𝑎𝑥) 𝑑𝑥 = 𝑥 sin(𝑎𝑥) − ∫ 𝑥𝑛−1 sin(𝑎𝑥) 𝑑𝑥, 1
𝑎 𝑎 𝑥𝑛 = , 𝑛 < 0,
(𝑥 + 1) ⋯ (𝑥 + |𝑛|)
𝑥𝑛 𝑒 𝑎𝑥 𝑛 𝑥𝑛+𝑚 = 𝑥𝑚 (𝑥 − 𝑚)𝑛 .
74. ∫ 𝑥𝑛 𝑒𝑎𝑥 𝑑𝑥 = − ∫ 𝑥𝑛−1 𝑒𝑎𝑥 𝑑𝑥,
𝑎 𝑎

ln(𝑎𝑥) 1 Rising Factorial Powers:

75. ∫ 𝑥𝑛 ln(𝑎𝑥) 𝑑𝑥 = 𝑥𝑛+1 ( − ),
𝑛+1 (𝑛 + 1)2
𝑥𝑛 = 𝑥(𝑥 + 1) ⋯ (𝑥 + 𝑛 − 1), 𝑛 > 0,
𝑥𝑛+1 𝑚
76. ∫ 𝑥𝑛 (ln 𝑎𝑥)𝑚 𝑑𝑥 = (ln 𝑎𝑥)𝑚 − ∫ 𝑥𝑛 (ln 𝑎𝑥)𝑚−1 𝑑𝑥. 𝑥0 = 1,
𝑛+1 𝑛+1
1
𝑥𝑛 = , 𝑛 < 0,
(𝑥 − 1) ⋯ (𝑥 − |𝑛|)
𝑥1 = 𝑥1 = 𝑥1
𝑥𝑛+𝑚 = 𝑥𝑚 (𝑥 + 𝑚)𝑛 .
𝑥2 = 𝑥2 + 𝑥1 = 𝑥2 − 𝑥1
Conversion:
𝑥3 = 𝑥3 + 3𝑥2 + 𝑥1 = 𝑥3 − 3𝑥2 + 𝑥1
𝑥𝑛 = (−1)𝑛 (−𝑥)𝑛 = (𝑥 − 𝑛 + 1)𝑛 = 1/(𝑥 + 1)−𝑛 ,
𝑥4 = 𝑥4 + 6𝑥3 + 7𝑥2 + 𝑥1 = 𝑥4 − 6𝑥3 + 7𝑥2 − 𝑥1 𝑥𝑛 = (−1)𝑛 (−𝑥)𝑛 = (𝑥 + 𝑛 − 1)𝑛 = 1/(𝑥 − 1)−𝑛 ,
3
𝑥5 = 𝑥5 + 15𝑥4 + 25𝑥 + 10𝑥2 + 𝑥1 = 𝑥5 − 15𝑥4 + 25𝑥3 − 10𝑥2 + 𝑥1 𝑛
𝑛 𝑛
𝑛
𝑥𝑛 = ∑ [ ] 𝑥𝑘 = ∑ [ ] (−1)𝑛−𝑘 𝑥𝑘 ,
𝑘=1
𝑘 𝑘=1
𝑘
𝑥1 = 𝑥1 𝑥1 = 𝑥1 𝑛
𝑛
𝑥𝑛 = ∑ [ ] (−1)𝑛−𝑘 𝑥𝑘 ,
𝑘
𝑥2 = 𝑥2 + 𝑥1 𝑥2 = 𝑥2 − 𝑥1 𝑘=1
𝑛
𝑛
3 𝑥𝑛 = ∑ [ ] 𝑥𝑘 .
𝑥3 = 𝑥3 + 3𝑥2 + 2𝑥1 𝑥 = 𝑥3 − 3𝑥2 + 2𝑥1 𝑘=1
𝑘

𝑥4 = 𝑥4 + 6𝑥3 + 11𝑥2 + 6𝑥1 𝑥4 = 𝑥4 − 6𝑥3 + 11𝑥2 − 6𝑥1

𝑥5 = 𝑥5 + 10𝑥4 + 35𝑥3 + 50𝑥2 + 24𝑥1 𝑥5 = 𝑥5 − 10𝑥4 + 35𝑥3 − 50𝑥2 + 24𝑥1


 Mathematics Cheat Sheet
Series
Taylor’s series: Ordinary power series:
∞ ∞
(𝑥 − 𝑎)2 ″ (𝑥 − 𝑎)𝑖 (𝑖)
𝑓(𝑥) = 𝑓(𝑎) + (𝑥 − 𝑎)𝑓′ (𝑎) + 𝑓 (𝑎) + ⋯ = ∑ 𝑓 (𝑎). 𝐴(𝑥) = ∑ 𝑎 𝑖 𝑥𝑖 .
2 𝑖=0
𝑖! 𝑖=0
Expansions: Exponential power series:
1 ∞
= 1 + 𝑥 + 𝑥2 + 𝑥3 + 𝑥4 + ⋯ =
∞
∑𝑖=0 𝑥 ,
𝑖 𝑥𝑖
1−𝑥 𝐴(𝑥) = ∑ 𝑎 𝑖 .
𝑖=0
𝑖!
1 ∞
= 1 + 𝑐𝑥 + 𝑐 2 𝑥2 + 𝑐 3 𝑥3 + ⋯ = ∑𝑖=0 𝑐 𝑥 ,
𝑖 𝑖
1−𝑐𝑥 Dirichlet power series:
∞
1 ∞ 𝑛𝑖 𝑎
= 1 + 𝑥𝑛 + 𝑥2𝑛 + 𝑥3𝑛 + ⋯ = ∑𝑖=0 𝑥 , 𝐴(𝑥) = ∑ 𝑥𝑖 .
1−𝑥𝑛 𝑖
𝑥 ∞ 𝑖=1
= 𝑥 + 2𝑥2 + 3𝑥3 + 4𝑥4 + ⋯ = 𝑖
∑𝑖=0 𝑖𝑥 ,
(1−𝑥)2 Binomial theorem:
𝑛
𝑘!𝑧 𝑘 𝑛
𝑛 𝑛
= 𝑥 + 2𝑛 𝑥 2 + 3𝑛 𝑥 3 + 4𝑛 𝑥 4 +⋯ =
∞ 𝑛 𝑖 (𝑥 + 𝑦)𝑛 = ∑ ( ) 𝑥𝑛−𝑘 𝑦 𝑘 .
∑𝑘=0 {𝑘} (1−𝑧)𝑘+1 ∑𝑖=0 𝑖 𝑥 , 𝑘
𝑘=0
1 1 ∞ 𝑥𝑖 Difference of like powers:
𝑒𝑥 = 1 + 𝑥 + 𝑥2 + 𝑥3 + ⋯ = ∑𝑖=0 , 𝑛−1
2 6 𝑖!
1 2 1 3 1 4 ∞ 𝑖+1 𝑥
𝑖 𝑥𝑛 − 𝑦 𝑛 = (𝑥 − 𝑦) ∑ 𝑥𝑛−1−𝑘 𝑦 𝑘 .
ln(1 + 𝑥) = 𝑥− 𝑥 + 𝑥 − 𝑥 −⋯ = ∑𝑖=1 (−1) 𝑖 , 𝑘=0
2 3 4
1 1 1 1 ∞ 𝑥𝑖 For ordinary power series:
ln = 𝑥 + 𝑥2 + 𝑥3 + 𝑥4 + ⋯ = ∑𝑖=1 , ∞
𝛼𝐴(𝑥) + 𝛽𝐵(𝑥) = ∑𝑖=0 (𝛼𝑎 𝑖 + 𝛽𝑏𝑖 )𝑥𝑖 ,
1−𝑥 2 3 4 𝑖
2𝑖+1 ∞
sin 𝑥 =
1
𝑥 − !𝑥3 + !𝑥5 − !𝑥7 + ⋯
1 1
=
∞
∑𝑖=0 (−1)
𝑖 𝑥 , 𝑥𝑘 𝐴(𝑥) = ∑𝑖=𝑘 𝑎 𝑖−𝑘 𝑥 ,
𝑖
3 5 7 (2𝑖+1)!
1 1 1 ∞ 2𝑖
𝑖 𝑥 , 𝐴(𝑥)−∑𝑘−1
𝑖=0 𝑎 𝑖 𝑥
𝑖
∞ 𝑖
cos 𝑥 = 1 − !𝑥2 + !𝑥4 − !𝑥6 + ⋯ = ∑𝑖=0 (−1) = ∑𝑖=0 𝑎 𝑖+𝑘 𝑥 ,
2 4 6 (2𝑖)! 𝑥𝑘
2𝑖+1 ∞ 𝑖 𝑖
1 1 1 ∞ 𝑖 𝑥 𝐴(𝑐𝑥) = ∑𝑖=0 𝑐 𝑎 𝑖 𝑥 ,
tan−1 𝑥 = 𝑥 − 𝑥3 + 𝑥5 − 𝑥7 + ⋯ = ∑𝑖=0 (−1) (2𝑖+1) ,
3 5 7 ∞
𝑛 ∞ 𝐴′ (𝑥) = ∑𝑖=0 (𝑖 + 1)𝑎 𝑖+1 𝑥 ,
𝑖
(1 + 𝑥)𝑛 = 1 + 𝑛𝑥 + 𝑛 − 1)2𝑥2 + ⋯ = ∑𝑖=0 ( 𝑛 ) 𝑥𝑖 ,
( 𝑖 ∞
1 ∞
𝑥𝐴′ (𝑥) = ∑𝑖=1 𝑖𝑎 𝑖 𝑥 ,
𝑖
= 1 + (𝑛 + 1)𝑥 + ( 𝑛+2 ) 𝑥2 +⋯ = 𝑖+𝑛 𝑖
∑𝑖=0 ( 𝑖 ) 𝑥 ,
(1−𝑥)𝑛+1 2 ∞ 𝑎 𝑖−1 𝑖
∫ 𝐴(𝑥) 𝑑𝑥 = ∑𝑖=1 𝑖
𝑥,
𝑥 1 1 2 1 4 ∞ 𝐵𝑖 𝑥𝑖
= 1− 𝑥+ 𝑥 − 𝑥 +⋯ = ∑𝑖=0 , 𝐴(𝑥)+𝐴(−𝑥) ∞
𝑒 𝑥 −1 2 12 720 𝑖! = 2𝑖
2 ∑𝑖=0 𝑎 2𝑖 𝑥 ,
1 ∞ 1 2𝑖
(1 − √1 − 4𝑥) = 1 + 𝑥 + 2𝑥2 + 5𝑥3 +⋯ = 𝑖
∑𝑖=0 𝑖+1 ( 𝑖 ) 𝑥 , 𝐴(𝑥)−𝐴(−𝑥) ∞
2𝑥 = 2𝑖+1 .
2 ∑𝑖=0 𝑎 2𝑖+1 𝑥
1 ∞ 2𝑖
= 1 + 2𝑥 + 6𝑥2 + 20𝑥3 +⋯ = 𝑖
∑𝑖=0 ( 𝑖 ) 𝑥 ,
√1−4𝑥 𝑖
𝑛 Summation: If 𝑏𝑖 = ∑𝑗=0 𝑎 𝑖 then
1 1−√1−4𝑥 ∞
( ) = 1 + (2 + 𝑛)𝑥 + ( 4+𝑛 ) 𝑥2 + ⋯ = ∑𝑖=0 (
2𝑖+𝑛
) 𝑥𝑖 , 𝐵(𝑥) =
1
𝐴(𝑥).
√1−4𝑥 2𝑥 2 𝑖
1−𝑥
1 1 3 11 25 4 ∞
ln = 𝑥 + 𝑥2 + 𝑥3 + 𝑥 +⋯ = ∑𝑖=1 𝐻𝑖 𝑥 ,
𝑖 Convolution:
1−𝑥 1−𝑥 2 6 12 ∞ 𝑖
1 1 2 1 2 3 3
11 ∞ 𝐻𝑖−1 𝑥𝑖 𝐴(𝑥)𝐵(𝑥) = ∑ ( ∑ 𝑎 𝑗 𝑏𝑖−𝑗 ) 𝑥𝑖 .
(ln 1−𝑥 ) = 𝑥 + + 𝑥4 + ⋯
𝑥 = ∑𝑖=2 𝑖 , 𝑖=0 𝑗=0
2 2 424
𝑥 ∞
= 𝑥 + 𝑥2 + 2𝑥3 + 3𝑥4 + ⋯ = 𝑖
∑𝑖=0 𝐹𝑖 𝑥 ,
1−𝑥−𝑥2
God made the natural numbers; all the rest is the
𝐹𝑛 𝑥 ∞
= 𝐹𝑛 𝑥 + 𝐹2𝑛 𝑥2 + 𝐹3𝑛 𝑥3 + ⋯ = ∑𝑖=0 𝐹𝑛𝑖 𝑥𝑖 . work of man.
1−(𝐹𝑛−1 +𝐹𝑛+1 )𝑥−(−1)𝑛 𝑥2
– Leopold Kronecker


 Mathematics Cheat Sheet
Series Escher’s impossible brick

Expansions:
1 1 ∞ 1 −𝑛 ∞
ln = ∑𝑖=0 (𝐻𝑛+𝑖 − 𝐻𝑛 ) ( 𝑛+𝑖 ) 𝑥𝑖 , (𝑥) = ∑𝑖=0 [ 𝑛 ] 𝑥 ,
𝑖 𝑖
(1−𝑥)𝑛+1 1−𝑥 𝑖
∞ ∞ 𝑛!𝑥𝑖
𝑥𝑛 = ∑𝑖=0 [ 𝑛 ] 𝑥𝑖 , (𝑒 𝑥 − 1)𝑛 = ∑𝑖=0 [ 𝑛 ]
𝑖
,
𝑖 𝑖!
1 𝑛 ∞ 𝑛!𝑥𝑖 ∞ (−4)𝑖 𝐵2𝑖 𝑥2𝑖
(ln 1−𝑥 ) = ∑𝑖=0 [ 𝑖 ] , 𝑥 cot 𝑥 = ∑𝑖=0 ,
𝑛 𝑖! (2𝑖)!
∞ 22𝑖 (22𝑖 −1)𝐵2𝑖 𝑥2𝑖−1 ∞1
tan 𝑥 = ∑𝑖=1 (−1)𝑖−1 , 𝜁(𝑥) = ∑𝑖=1 ,
(2𝑖)! 𝑖𝑥
1 ∞ 𝜇(𝑖) 𝜁(𝑥−1) ∞ 𝜙(𝑖)
= ∑𝑖=1 , = ∑𝑖=1 𝑖 𝑥 ,
𝜁(𝑥) 𝑖𝑥 𝜁(𝑥)
1
𝜁(𝑥) = ∏𝑝 , Stieltjes Integration
1−𝑝−𝑥
∞ 𝑑(𝑖) If 𝐺 is continuous in the interval [𝑎, 𝑏] and 𝐹 is nondecreasing then
𝜁 2 (𝑥) = ∑𝑖=1 𝑥𝑖 where 𝑑(𝑛) = ∑𝑑|𝑛 1, 𝑏
∞ 𝑆(𝑖) ∫ 𝐺(𝑥) 𝑑𝐹(𝑥)
𝜁(𝑥)𝜁(𝑥 − 1) = ∑𝑖=1 𝑥𝑖 where 𝑆(𝑛) = ∑𝑑|𝑛 𝑑, 𝑎
exists. If 𝑎 ≤ 𝑏 ≤ 𝑐 then
22𝑛−1 |𝐵2𝑛 | 2𝑛 𝑐 𝑏 𝑐
𝜁(2𝑛) = 𝜋 , 𝑛 ∈ ℕ, ∫ 𝐺(𝑥) 𝑑𝐹(𝑥) = ∫ 𝐺(𝑥) 𝑑𝐹(𝑥) + ∫ 𝐺(𝑥) 𝑑𝐹(𝑥).
(2𝑛)!
𝑎 𝑎 𝑏
𝑥 𝑖 2𝑖 If the integrals involved exist
=
∞ 𝑖−1 (4 −2)𝐵2𝑖 𝑥 ,
sin 𝑥 ∑𝑖=0 (−1) (2𝑖)!
𝑏 𝑏 𝑏
𝑛 ∫ (𝐺(𝑥) + 𝐻(𝑥)) 𝑑𝐹(𝑥) = ∫ 𝐺(𝑥) 𝑑𝐹(𝑥) + ∫ 𝐻(𝑥) 𝑑𝐹(𝑥),
1−√1−4𝑥 ∞𝑛(2𝑖+𝑛−1)! 𝑖 𝑎 𝑎 𝑎
( ) = ∑𝑖=0 𝑥, 𝑏 𝑏 𝑏
2𝑥 𝑖!(𝑛+𝑖)!
𝑖/2 𝑖𝜋 ∫ 𝐺(𝑥) 𝑑(𝐹(𝑥) + 𝐻(𝑥)) = ∫ 𝐺(𝑥) 𝑑𝐹(𝑥) + ∫ 𝐺(𝑥) 𝑑𝐻(𝑥),
∞ 2 sin 4 𝑖 𝑎 𝑎 𝑎
𝑒𝑥 sin 𝑥 = ∑𝑖=1 𝑥,
𝑖!
𝑏 𝑏 𝑏
√1−√1−𝑥 ∞ (4𝑖)! 𝑖
= ∑𝑖=0 16𝑖 √2(2𝑖)!(2𝑖+1)! 𝑥 , ∫ 𝑐 ⋅ 𝐺(𝑥) 𝑑𝐹(𝑥) = ∫ 𝐺(𝑥) 𝑑(𝑐 ⋅ 𝐹(𝑥)) = 𝑐 ∫ 𝐺(𝑥) 𝑑𝐹(𝑥),
𝑥 𝑎 𝑎 𝑎
𝑏 𝑏
arcsin 𝑥 2 ∞ 4𝑖 𝑖!2 2𝑖
( 𝑥 ) = ∑𝑖=0 (𝑖+1)(2𝑖+1)! 𝑥 . ∫ 𝐺(𝑥) 𝑑𝐹(𝑥) = 𝐺(𝑏)𝐹(𝑏) − 𝐺(𝑎)𝐹(𝑎) − ∫ 𝐹(𝑥) 𝑑𝐺(𝑥).
𝑎 𝑎
If the integrals involved exist, and 𝐹 possesses a derivative 𝐹 ′ at every point in [𝑎, 𝑏] then
𝑏 𝑏
∫ 𝐺(𝑥) 𝑑𝐹(𝑥) = ∫ 𝐺(𝑥)𝐹 ′ (𝑥) 𝑑𝑥. Fibonacci numbers
𝑎 𝑎

Cramer’s rule 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, …

         
If we have equations:           Definitions:
𝑎 1,1 𝑥1 + 𝑎 1,2 𝑥2 + ⋯ + 𝑎 1,𝑛 𝑥𝑛 = 𝑏1           𝐹𝑖 = 𝐹𝑖−1 + 𝐹𝑖−2 , 𝐹0 = 𝐹1 = 1,
         
𝑎 2,1 𝑥1 + 𝑎 2,2 𝑥2 + ⋯ + 𝑎 2,𝑛 𝑥𝑛 = 𝑏2           𝐹−𝑖 = (−1)𝑖−1 𝐹𝑖 ,
⋮ ⋮ ⋮           1
          𝐹𝑖 =
√5
(𝜙 𝑖 − 𝜙̂𝑖 ) ,
𝑎 𝑛,1 𝑥1 + 𝑎 𝑛,2 𝑥2 + ⋯ + 𝑎 𝑛,𝑛 𝑥𝑛 = 𝑏𝑛          
          Cassini’s identity: for 𝑖 > 0:
Let 𝐴 = (𝑎 𝑖,𝑗 ) and 𝐵 be the column matrix (𝑏𝑖 ). Then
         
𝐹𝑖+1 𝐹𝑖−1 − 𝐹𝑖 2 = (−1)𝑖 .
there is a unique solution iff det 𝐴 ≠ 0. Let 𝐴𝑖 be 𝐴 with Additive rule:
column 𝑖 replaced by 𝐵. Then The Fibonacci number system: Every integer 𝑛 has a unique 𝐹
𝑛+𝑘 = 𝐹𝑘 𝐹𝑛+1 + 𝐹𝑘−1 𝐹𝑛 ,
det 𝐴𝑖 representation
𝑥𝑖 = . 𝐹2𝑛 = 𝐹𝑛 𝐹𝑛+1 + 𝐹𝑛−1 𝐹𝑛 .
det 𝐴 𝑛 = 𝐹𝑘1 + 𝐹𝑘2 + ⋯ + 𝐹𝑘𝑚 ,
where 𝑘𝑖 ≥ 𝑘𝑖+1 + 2 for all 𝑖, 1 ≤ 𝑖 < 𝑚 and 𝑘𝑚 ≥ 2. Calculation by matrices:
Improvement makes strait roads, but the crooked roads 𝑛
𝐹𝑛−2 𝐹𝑛−1 0 1
without Improvement, are roads of Genius. (𝐹 = .
– William Blake (The Marriage of Heaven and Hell) 𝑛−1 𝐹𝑛 ) ( 1 1)

