university of copenhagen Burnside rings and fusion systems Sune Precht Reeh Centre for Symmetry and Deformation Department of Mathematical Sciences YTM 2013, July 11 Slide 1/13 university of copenhagen Outline 1 Burnside rings and the Segal conjecture The Burnside ring of a finite group The Segal conjecture 2 Fusion stable elements The Burnside ring of a saturated fusion system A transfer map 3 Segal conjecture for fusion systems Slide 2/13 university of copenhagen The Burnside ring of a finite group Let G be a finite group. The isomorphism classes of finite G-sets form an abelian monoid with disjoint union as addition. The Grothendieck group of this monoid is the Burnside ring of G, denoted A(G). The multiplication in A(G) is given by Cartesian products. Every finite G is the disjoint union of its orbits, hence the transitive G-sets [G/H] for H ≤ G form an additive basis for A(G). The basis element [G/H] only depends on H up to conjugation in G. Slide 3/13 university of copenhagen For each finite G-set X and any subgroup H ≤ G, we can count the number of fixed points |X H |. Taking fixed points respects the structure of A(G): |(X t Y )H | = |X H | + |Y H | and |(X × Y )H | = |X H | · |Y H |, so the fixed point maps X 7→ |X H | extend to ring homomorphisms A(G) → Z. |X H | only depends on H up to G-conjugation, and the collection of numbers |X H | for H ≤ G determines X ∈ A(G) uniquely. Slide 4/13 university of copenhagen Connections to topology Theorem (Segal conjecture, proven by Carlsson) Let G be a finite group. The 0’th stable cohomotopy group πS0 (BG+ ) = [Σ∞ (BG+ ), S] is isomorphic to the completion A(G)∧ I of the Burnside ring at a certain ideal I ≤ A(G). This completion is in general difficult to describe. However, if S is a p-group, then A(S)∧ I is (almost) the p-completion A(S)∧ . p Question: What is the corresponding result for πS0 (BG∧ p )? Slide 5/13 university of copenhagen Fusion systems A fusion system over a finite p-group S is a category F where the objects at the subgroups P ≤ S and the morphisms satisfy: • HomS (P, Q) ⊆ F(P, Q) ⊆ Inj(P, Q) for all P, Q ≤ S. • Every ϕ ∈ F(P, Q) factors in F as an isomorphism P → ϕP followed by an inclusion ϕP ,→ Q. A saturated fusion system satisfies a few additional axioms that play the role of Sylow’s theorems. The canonical example of a saturated fusion system is FS (G) defined for S ∈ Sylp (G) with morphisms HomFS (G) (P, Q) := HomG (P, Q). for P, Q ≤ S. Slide 6/13 university of copenhagen The Burnside ring of a fusion system Let F be a saturated fusion system over S. A finite S-set X (or an element of A(S)) is said to be F-stable if |X Q | = |X P | whenever Q, P are isomorphic/conjugate in F. Theorem (R.) The isomorphism classes of F-stable finite S-sets form a free abelian monoid. The irreducible stable sets are in 1-to-1 correspondence with the subgroups of S up to F-conjugation. The Grothendieck group of this monoid is the Burnside ring of F, denoted A(F), and is a subring of A(S). Slide 7/13 university of copenhagen A transfer map Theorem (R.) Let F be a saturated fusion system over a finite p-group S. ∧ Then there is a transfer map π : A(S)∧ p → A(F)p satisfying π(X)Q = X 1 X Q0 0 #{Q ∼F Q} 0 Q ∼F Q for all X ∈ A(S)∧ p and Q ≤ S. Furthermore, π is a homomorphism of A(F)∧ p -modules and restricts to the ∧ identity on A(F)p . Slide 8/13 university of copenhagen The characteristic idempotent A generalization of the Segal conjecture for p-groups S, T states that homotopy classes of stable maps [BS, BT ] is (almost) the p-completion of A(S, T ) – which is constructed like a Burnside ring, but from sets that have both an S-action and a free T -action that commute. To each saturated fusion system F there is a unique associated element 0 6= ωF ∈ A(S, S)∧ p such that • ωF is F-stable with respect to both S-actions, • every stabilizer of ωF has the form {(p, ϕp) | p ∈ P } for some F(P, S), • the corresponding stable homotopy class ωF ∈ [BS, BS] is idempotent. Slide 9/13 university of copenhagen The Segal conjecture for fusion systems The classifying spectrum BF can be constructed as the infinite mapping telescope ω ω F F BF := hocolim(BS −−→ BS −−→ BS → · · · ). If F = FS (G) with S ∈ Sylp (G), then BF = Σ∞ (BG∧ p ). Theorem (Ragnarsson) We can view π 0 (BF) as a subring of π 0 (BS), and then we have π 0 (BF) = (ωF )∗ (π 0 (BS)). In particular, π 0 (BF) is (almost) isomorphic to ∧ ∼ (ωF )∗ (A(S)∧ p ) = A(F)p . Slide 10/13 university of copenhagen Theorem (R.) ∧ The map A(S)∧ p → A(F)p induced by the characteristic ∧ idempotent ωF ∈ A(S, S)p coincides with the transfer map from earlier: Q X 1 (ωF )∗ (X) = X Q0 . #{Q0 ∼F Q} 0 Q ∼F Q Slide 11/13 university of copenhagen Thanks for your attention! Slide 12/13 university of copenhagen References [1] Kári Ragnarsson, Classifying spectra of saturated fusion systems, Algebr. Geom. Topol. 6 (2006), 195–252. MR2199459 (2007f:55013) [2] Kári Ragnarsson and Radu Stancu, Saturated fusion systems as idempotents in the double Burnside ring, 54 pp., preprint, available at arXiv:0911.0085. [3] Sune Precht Reeh, The abelian monoid of fusion-stable finite sets is free, 14 pp., preprint, available at arXiv:1302.4628. [4] Sune Precht Reeh, Transfer and characteristic idempotents for saturated fusion systems, 39 pp., preprint, available at arXiv:1306.4162. Slide 13/13