N-groups and fusion systems. (Q907342)

\textit{J. G. Thompson}'s \(N\)-group paper [Bull. Am. Math. Soc. 74, 383-437 (1968; Zbl 0159.30804); Pac. J. Math. 33, 451-536 (1970; Zbl 0256.20024); ibid. 39, 483-534 (1971; Zbl 0256.20025); ibid. 48, 511-592 (1973; Zbl 0291.20014); ibid. 50, 215-297 (1974; Zbl 0291.20015); ibid. 51, 573-630 (1974; Zbl 0306.20012)], which classified all finite simple groups all of whose local subgroups are solvable, was a game plane for the classification of the finite simple groups. In [J. Algebra 190, No. 1, 11-67 (1997; Zbl 0877.20008)] \textit{B. Stellmacher} based on the amalgam method classified the amalgams belonging to \(N_2\)-groups, which are groups whose 2-local subgroups are solvable. This to a certain extend was the starting point of the MSS-program, which aims to understand the \(p\)-local structure of groups of local characteristic 2 and at least for \(p=2\) to classify the simple groups all of whose 2-local subgroups are of characteristic 2 [see Mem. Am. Math. Soc. 1147 (2016)]. The author of the paper under review aims in future to give a classification of the finite simple groups based on fusion systems. Hence it is natural to ask if there is a theorem similar to the \(N\)-group theorem or the \(N_2\)-group theorem in the language of fusion systems. For the prime 2 there is an answer given in this paper. The author defines a saturated fusion system \(\mathcal F\) on a \(2\)-group \(S\) to be an \(N\)-system if for each nontrivial fully normalized subgroup \(P\) of \(S\), \(N_{\mathcal F}(P)\) is constrained and \(\Aut_{\mathcal F}(P)\) is solvable. In this paper all fusion systems of \(N\)-type are classified. This again might be a test case for a classification of all 2-fusion systems of characteristic 2-type. Instead of stating this result it might be more informative to state a corollary, which is the combination of the main theorem with the pushing up results obtained by the author [in Invent. Math. 180, No. 2, 225-299 (2010; Zbl 1234.20025)] and Theorem 2 of [J. Algebra 378, 217-262 (2013; Zbl 1297.20017)]: It says: There is a finite group \(G\) such that \(\mathcal F=\mathcal F_S(G)\), the fusion system of \(G\) and one of the following holds: (i) \(G\) is solvable with \(F^*(G)=O_2(G)\), (ii) \(S\) is dihedral of order at least 16, and \(G\cong L_2(q)\) or \(PGL_2(q)\) for some odd prime power \(q\). (iii) \(S\) is semidihedral and \(F^*(G)\cong L_2(q)\) for some odd prime power \(q\), which is a square, (iv) \(S\) is semidihedral of order 16 and \(G\cong L_3(3)\), (v) \(|S|=32\) and \(G\cong\Aut(A_6)\) or \(\Aut(L_3(3))\), (vi) \(G\cong L_3(2)\), \(Sp_4(2)\), \(U_3(3)\), \(M_{12}\), \(\Aut(M_{12})\), \(^2F_4(2)'\), or \(^2F_4(2)\). Using classifications of simple groups by Sylow 2-subgroups, the solvable signalizer theorem and Goldschmidt's theorem on strongly closed subgroups one then receives as a corollary the classification of the \(N_2\)-groups due to J. Thompson, \textit{Z. Janko} [J. Algebra 21, 458-517 (1972; Zbl 0243.20013)], \textit{F. Smith} [J. Algebra 34, 481-520 (1975; Zbl 0325.20009)], \textit{D. Gorenstein} and \textit{R. Lyons} [J. Algebra 38, 453-522 (1976; Zbl 0402.20012)]. The crucial point in the proof is to show that in the generic case if we do not have pushing up, then \(\mathcal F\) is generated by \(C_{\mathcal F}(Z)\), \(Z=\Omega_1(Z(S))\) and \(N_{\mathcal F}(V)\), where \(V\) is either \(\langle Z^{\Aut_{\mathcal F}(B)}\rangle\), with \(B\) the Baumann subgroup of \(S\) or \(V=\langle Z^{\mathcal F_M}\rangle\), where \(\mathcal F_M\) is a unique maximal local subsystem on \(S\) such that \(J(S)\) does not centralize \(\langle Z^{\mathcal F_M}\rangle\).