MTH3003 Group Theory - Final Exam Cheat Sheet

Created by William Fayers

Good luck and have fun!! :)) - start at the §1 dispatcher, match your question, follow the arrows to a method, run the numbered steps, sanity-check (§14). Every recipe links to the definitions it needs. Straight from Simon’s notes (theorem numbers kept). Ez 100%.

§0. Notation, Groups, Key Facts

Notation & terms - every cross-reference bottoms out here.

Symbol / termPlain meaning
, , group with operation ; identity = the element that does nothing in (e.g. in , in ); order = number of elements
order of an element: least with (times to multiply to reach )
abelian - order of multiplication doesn’t matter
simple (Def 4.2.1)only normal subgroups are and
subgroup: a smaller group inside (same operation)
, generated: all you can build by multiplying; = powers of , with ; cyclic = whole group is some
, coset = shifted by , ; index = number of cosets
normal subgroup: - the kind you can “divide by”
quotient group: the cosets of , made a group by (needs normal)
homomorphism: ( usually a group action)
, kernel = elements sent to ; image = elements actually hit
isomorphic: a perfect 1-1 homomorphism (“same group, relabelled”)
conjugation, replace by ; centraliser = elements that commute with
action, , , action = each permutes a set ; orbit = where can be sent (transitive = one orbit); stabiliser = the ‘s fixing ; fixed set = points fixed by one
Sylow -subgroupbiggest subgroup whose size is a power of the prime
direct product: pairs multiplied slot-by-slot
misc divides ” ( = doesn’t); nonzero reals; integers mod ; cycle

Axioms : Closure, Associativity, Identity, Inverses.

Identities (Thm 2.1.11): ; cancel ; ; ; identity/inverses unique.

GroupOrderAbelian?Notes
, yescyclic
/ any finite abelianyes iff (§13)
yes; smallest non-cyclic
yes
no ()all perms of
no ()even perms; ; simple
no ()-gon symmetries ( tri, square, pent)
no;
nofinitary perms;

§1. Dispatcher - Find Your Question

Q1-type (permutations / symmetry)

  • Cycle notation, product, inverse, order, find element of order , solve §2
  • Odd/even (signature), is , list §9
  • in cycles / simplify word / / symmetry group of a shape / “is it abelian” → §5; prove §5+§8

Q2-type (structure: Lagrange, homs, iso theorems, simple)

  • Prove / group identity → §3; list / ” cyclic” / §4
  • State+prove Lagrange / rule out a subgroup order / cosets / Cauchy → §6
  • Prove / / quotient group → §7
  • Is it simple? / hom out of a simple group → §7
  • Prove/disprove homomorphism / build an isomorphism → §8
  • State+prove First Iso / identify a quotient via a hom / 3rd-theorem cancellation → §9 (theorems) + §8 (build map)

Q3/Q4-type (actions / Sylow / abelian)

  • Is an action / verify-or-refute / Cayley → §10
  • Orbit & stabiliser / Orbit-Stabiliser / rule out orbit size / orbit counting → §11
  • State Sylow / Sylow example / count / “not simple” / subgroup of order §12
  • Classify abelian groups of order / subgroups of abelian / element of order / / abelian → §13

Each question: part (a) ~9 marks (often “state & prove”), (b)~8, (c)~8 (stretch). Answer best 3 of 4.

§2. Permutations

Cycle : , , rest fixed. Disjoint cycles commute (1.2.7); every = unique product of disjoint cycles (1.3.4). Products act right-to-left.

  • Two-row → cycles: start at smallest unused , follow till back to (closes a cycle), repeat on unused; drop fixed points.
  • Cycles → two-row: read off ; unlisted points fixed.
  • Product : chase each point right-to-left through every factor, close each cycle, repeat. (Disjoint factors can be done independently.) ex (do first): , , , .
  • Inverse: reverse each cycle, ; invert a product factor-by-factor.
  • Order: of cycle lengths.
  • Find element of order in : choose disjoint cycle lengths summing to with (e.g. , lengths ). None exists if it would need a cycle longer than (e.g. order in : needs a -cycle).
  • Solve : cancel outward - (left factor inverts on left, right factor on right). Then compute and verify by substituting back. ex (WP1.4): .
  • Odd/even → §9.

ex (Mock A1a): , ; .

§3. Subgroups

  • Prove (Quick Subgroup Test, Thm 2.2.5): show 1) , 2) , 3) (or combine 2-3: ). Standard targets: (use both), (§8), (§11), (§9), (support stays finite, §7).
  • Prove a group identity: expand and cancel . Key ones: ; .
  • Order facts: ; finite (fails if infinite: ).

§4. Cyclic Groups

, .

  • List : compute until repeats ( elements). likewise.
  • Prove : division algorithm , so exactly .
  • Prove cyclic: if prime, any has (Lagrange §6) so , thus . For : so .
  • Disprove "": counterexample : , order .

§5. Dihedral & Symmetry Groups

, order ; , (so ). , = reflection through vertex .

  • Write in cycle notation: rotations “shift by ”; reflections - draw the -gon and read off, or compute .
  • Prove : is a reflection so , i.e. ; rearrange .
  • Simplify a word to or : push every to the left via , then reduce mod , .
  • Prove : every word reduces (above) to , , = the listed elements.
  • Symmetry group of a labelled shape: write each rotation and reflection as a permutation of the labels; collect them.
  • Construct a non--gon shape with symmetry: regular -gon + an identical decoration on each edge (e.g. hexagon + square per edge ).
  • Is the symmetry group abelian? Identify it by counting symmetries: rotations and no reflections → (cyclic, abelian); rotations and reflections → (non-abelian for ). ex (Mock B1b): a pinwheel (4-fold rotation, no reflections) → → abelian. (Which groups are abelian: §0.)
  • Prove : build the iso (§8) on generators ; check relations hold and images are all 6 of .

§6. Cosets & Lagrange

Coset (Def 4.1.1); = distinct cosets. (4.1.3); cosets are disjoint-or-equal, all size (4.1.4).

  • List cosets / find index: start ; pick listed, form ; repeat until used up; = count. ex , : , , .
  • Lagrange (Thm 4.2.5) - “state & prove”:
    • Statement: if with finite, then ; in particular .
    • Terms: = subgroup; = orders (number of elements); coset ; index = number of distinct cosets of in .
    • Proof: every lies in a coset (4.1.4), so with the distinct cosets; distinct cosets are disjoint (4.1.4) and all have size (bijection ); hence , so .
  • Rule out a subgroup/element of order : if , impossible. ( via , §4.) Converse is false (, no order-6).
  • Cauchy (4.2.6): prime an element of order exists.

§7. Normal Subgroups, Quotients & Simple Groups

(Def 4.1.5): ; equiv. .

  • Prove - pick the cheapest route:
    1. Direct: fix , show (do it for all ).
    2. Abelian ⇒ automatic (which are abelian: §0).
    3. for some hom ⇒ normal (§8) - the slickest; e.g. , .
    4. Index 2 (, §6) ⇒ normal.
    5. Invariant under conjugation by a formula: , so and are normal in ; and (support = points moved), so .
    6. Unique Sylow -subgroup ⇒ normal (§12).
  • Order Switching (Lem 4.2.3) : , (). Key Lemma (4.2.4): , , . (Use these to prove via Quick Subgroup Test, swapping orders with Order Switching.)
  • Quotient (needs ): elements , , . Identify it: find a surjective hom with kernel and apply FIT (§9) - e.g. .
  • Is simple? (only normal subgroups ):
    • ( prime) yes (subgroup orders only).
    • No if abelian with a proper nontrivial subgroup (e.g. - all subgroups normal).
    • () no (, §9).
    • no (, §13). E.g. .
    • General order: use Sylow to force a unique (normal) Sylow ⇒ not simple (§12).
  • Hom out of a simple group (Mock B2c, worked):
    • Claim: if is simple and is a homomorphism, then either for all (trivial), or is injective.
    • Proof: (§8). Since is simple, its only normal subgroups are and , so or . If , every maps to (trivial); if , then is injective (§8).

§8. Homomorphisms & Isomorphisms

Hom: . Props (5.2.4): , , , , .

  • Prove a map is a hom: check for general . Disprove: one counterexample. ( is a hom abelian, since commute.)
  • Prove : , cancel. : induction. : . : Quick Subgroup Test (§3).
  • Prove an isomorphism : build (define on generators, extend), then check 1) hom, 2) injective (), 3) onto. Shortcut: if finite, injective onto (check one). ex (here with = a rotation and = a reflection, §5): set ; the defining relations carry over (, , ) so is a hom; its 6 images are all of , so (equal orders) it’s a bijection ⇒ iso.
  • Power map on is an iso .
  • normal is the engine for §7 route 3 and §9.

§9. Isomorphism Theorems & the Alternating Group

  • First Isomorphism Theorem (Thm 6.0.1) - “state & prove”:
    • Statement: if is a homomorphism with kernel and image , then - via the map .
    • Terms: homomorphism ; kernel ; image ; quotient = cosets of with ; = isomorphic (a bijective homomorphism).
    • Proof (4 checks): well-defined ; hom ; onto each element of is ; 1-1 .
  • Use FIT to identify a quotient : spot a surjective hom with ; then . ex : is a hom, onto, with ; FIT ⇒ . Table: (); (signature); (); .
  • Second (6.0.4): (map , kernel ).
  • Third (6.0.5): , both (“cancellation”; map ).

Signature (Def 7.2.1, via ). (Def 7.3.1). Cycle shape (7.1.1).

  • Compute : over disjoint cycles (Cor 7.2.3). Quick: count even-length cycles - odd count ⇒ odd. (even iff odd). ex (Mock B1a): = 4-cycle (even length → sign ) · 3-cycle (sign ) → , odd. For a product , just multiply signs: .
  • is a homomorphism (Prop 7.2.2, examinable). Statement: . Proof: , so . Also (Prop 7.2.5), ; for a product, multiply signs (write as disjoint cycles first if computing directly).
  • Is ? .
  • and - “state & prove”:
    • Statement: the even permutations form a normal subgroup of , with .
    • Proof: subgroup (Quick Subgroup Test): , , . normal (Thm 7.3.5): (as ), so . order (Thm 7.3.2): is onto with , so FIT gives , whence .
    • Simple for (proof not examinable); is not simple ().
  • List via cycle shapes (shape: sign, count): . even: . even: [, , ].

§10. Group Actions & Cayley

Action (Def 8.1.1) : (1) ( = the identity of the group ), (2) . Same thing as a hom (8.1.4). .

  • Verify a map is an action: check 0) closure , 1) identity (use ‘s own identity - e.g. in , not ), 2) compatibility .
  • Show it’s NOT an action: exhibit the failing axiom. Closure (WP8.4: on , ); identity (Mock A3a: on by - here , and ); compatibility (8.1.10: ).
  • A hom is automatically an action (and vice versa) - the two are the same data (WP8.8).
  • Standard actions on : regular ; (left) conjugation . On subsets: (check it lands in the right-sized subsets).
  • Cayley: every a subgroup of . Proof: regular action has (), so FIT (§9) gives .

§11. Orbits, Stabilisers, Counting

Orbit (8.3.1); transitive if . Stabiliser (8.3.2). Orbits partition (9.1.0). Fixed set (9.2.1).

  • Compute orbit & stabiliser of : work out for every , then:
    • orbit = the set of distinct images - a subset of (the places lands);
    • stabiliser = the set of those that give back - a subset of (the group elements that fix ).
  • Deduce if the action is transitive (from the orbit): transitive the orbit is all of , i.e. (equivalently ). So compute the orbit, then compare its size to : equal → transitive; smaller → not transitive.
  • Prove : Quick Subgroup Test (§3): ; fix fixes ; fixes does.
  • Orbit-Stabiliser (Thm 9.1.1) - “state & prove”:
    • Statement: for all , (finite ).
    • Terms: orbit ; stabiliser ; = index (number of cosets).
    • Proof: the map () is a bijection cosets of , since (coset rule §6); so , and Lagrange gives .
  • Find an orbit/stabiliser size when you know the other: divide .
  • Rule out an orbit of size / a transitive action on : if (resp. ), impossible (Orbit-Stab forces integer). E.g. no orbit of size 4 in (); not transitive on a triangle ().
  • Conjugation action: = centraliser (things commuting with ); orbit = conjugacy orbit. ex (Mock A3): in , , orbit (not transitive).
  • Stabiliser of a subset (Mock B4): , order ; orbit = all -subsets (size 10); . ✓
  • (Mock B4c): quick - it’s the kernel of the action-hom (§10), and any kernel is normal (§8). Direct - fixes every point, so for any : ; thus , so .
  • Count orbits (Orbit Counting / Burnside, 9.2.3): . Recipe for colourings with colours: each fixes ; tabulate, sum, divide by . Octagon edges, 3 col, : . Cube faces, 3 col, rotations (24, no reflections): . [Proof: double-count : by , ; by , (orbits partition , each contributes ). Equate .]

§12. Sylow’s Theorems

with ( = the part of coprime to ). -group (10.1.1): a group of order . Sylow -subgroup (10.1.3): a subgroup of order (the largest power of dividing ). = the number of Sylow -subgroups.

  • State Sylow (define every term, Mock A4a):
    1. First (10.1.5): a Sylow -subgroup exists. (Generalises Cauchy §6.)
    2. Second (10.1.6): every -subgroup lies in a conjugate of a Sylow -subgroup; all Sylow -subgroups conjugate.
    3. Third (10.1.7): and . (Proofs not examinable.)
  • Give a Sylow example: factor for the order , then write down any subgroup of that order built from elements of -power order. Its type depends on - not always dihedral:
    • order (): cyclic - e.g. Sylow 5 of is .
    • order : or - e.g. Sylow 3 of () is (abelian, not dihedral).
    • order 8 (, in ): the square’s symmetries - this is the dihedral case (§5).
    • (only need it to exist? don’t construct - just cite Sylow First, next bullet.)
  • Prove a subgroup of order exists: factor , quote Sylow First.
  • Count / bound (via Sylow Third): write with (so = the part of coprime to , and = number of Sylow -subgroups). Then and . So list the divisors of , keep those - that’s the candidate set. To show : exhibit distinct Sylow subgroups (e.g. two squares on different 4-point sets ⇒ , so if the candidates are ).
  • Prove ” is not simple” (i.e. find a normal subgroup that is neither nor ):
    1. Why this works: simple = the only normal subgroups are and . So you just need one proper, nontrivial normal subgroup. Best source: a Sylow -subgroup that is unique, because it’s normal (§7).
    2. Factorise and pick a prime dividing it - try the largest prime first (most likely to force ).
    3. Pin down with Sylow Third: write the divisors of , then keep only those .
    4. If only survives: the Sylow -subgroup is unique, hence ; and lies strictly between and , so is proper and nontrivial ⇒ is not simple. (If is not forced to 1, try another prime.)
    • Worked example, : factor . Take (largest): then , divisors ; their remainders mod 7 are , so only survives. So the Sylow 7-subgroup () is unique ⇒ ; since , is proper and nontrivial ⇒ not simple.
    • (Why not ? Then , divisors , and both are , so - not forced to 1, so it proves nothing. That’s why we picked the larger prime 7.)
  • Classify small order: unique normal Sylows with , (§13). E.g. order 15 ⇒ .

§13. Direct Products & Finite Abelian Groups

External : ; ; (11.1.1).

  • Prove : ; check hom + bijection (§8).
  • Prove (): Quick Subgroup Test componentwise (§3).
  • Internal direct product (Thm 11.1.3, examinable) - “state & prove”:
    • Statement: if with , then (Key Lemma §7) and .
    • Terms: = normal subgroups (§7); = trivial subgroup ; ; = external direct product (above).
    • Proof: (commuting, Lemma 11.1.2) ( normal) so ; ( normal) so ; thus , giving . Then each is uniquely (if then ), so is well-defined, a homomorphism (by commuting), 1-1 and onto.
    • To detect : find normal with trivial intersection and .
  • Prove abelian: (each factor abelian).
  • (11.2.1/2). So .
  • Fundamental Theorem (11.3.2): every finite abelian (prime-power cyclic factors), unique up to order. [Fin. gen. (11.3.3): torsion, rank.]
  • Classify all abelian groups of order (FTfAG): factor ; for each prime, list the partitions of the exponent (ways to write as a sum of positive integers, order ignored); a partition gives factors ; combine across primes. ( abelian groups of order = partitions of .) ex (Mock B3a) : partitions of 3 are , , - three groups.
  • Find all subgroups of an abelian group (up to isomorphism) - three steps:
    • (1) Possible orders: by Lagrange (§6), each subgroup’s order divides .
    • (2) Exhibit one subgroup of each of those orders.
    • (3) Identify up to isomorphism - the step that completes the list (don’t skip): show each order has only one possible type, so any subgroup of that order matches one you found. Tools: order 1 = trivial group; order = itself; prime order cyclic (§4).
    • All subgroups here are automatically normal (since is abelian, §7).
    • Example - , order 6 (write , ):
      • Orders dividing 6: .
      • Exhibit: ; ; ; .
      • Iso step: any order-2 subgroup , any order-3 (prime ⇒ cyclic); order 1 = trivial; order 6 .
      • Four up to isomorphism: trivial, , , .
  • Count abelian groups of order with an element of order :
    1. Key fact: the order of an element of a direct product = lcm of the orders of its components. So an element of order exists iff, prime by prime, the factors can supply each prime-power in .
    2. For each prime power exactly dividing : you need a cyclic factor with in the -part (only such a factor holds an element of order ). Primes not in are unconstrained.
    3. So count, per relevant prime, the partitions of its exponent that include a part ; multiply across primes (free primes contribute all their partitions).
    • Worked: order , want an element of order (so need order from the 3-part and order from the 5-part):
      • 3-part (partitions of 3): ✓, ✓ (each has a factor ), ✗ (max element order ). → 2 of 3.
      • 5-part : only - has an element of order 5 ✓. → 1.
      • 2-part : irrelevant to order 45, so free: or . → 2.
      • Total (out of the abelian groups of order 540).
  • for (take ); hence abelian converse to Lagrange: abelian , has a subgroup of order (build it factor-by-factor via FTfAG).

§14. Pitfalls & Sanity Checks

Pitfalls

  • Permutation products act right-to-left.
  • Converse of Lagrange is false ().
  • is a hom only if abelian.
  • Quotient needs normal - check first.
  • Action checks: don’t forget closure , then identity, then compatibility.
  • Sylow needs both and .
  • Cube colouring = rotations only (24); octagon (2D) allows reflections (16).
  • FTfAG factors are prime powers; element order = lcm of component orders.

Sanity checks

  • Orders divide: , , class/orbit sizes .
  • and are integers.
  • Isomorphism check: , and it preserves order/abelian/cyclic.
  • Abelian-group count = product of partition-counts of the prime exponents.
  • After solving or building an iso: substitute back.