An algebraic method for computing \(N^\epsilon(f)\) (Q386170): Difference between revisions

\textit{R. F. Brown} [Fixed Point Theory Appl. 2006, Spec. Iss., Article 29470, 10 p. (2006; Zbl 1093.55003)] defined the notion of \(\epsilon\)-Nielsen number for a self-map \(f\) on a Riemannian manifold. This number is difficult to compute since it is not homotopy-invariant. The author presents conditions under which an algebraic computation is possible. Let \(M\) be a compact connected Riemannian manifold with or without boundary and let \(d\) be a Riemannian metric on \(M\). Choose an \(\epsilon>0\) so that any two points \(p,q\in M\) with \(d(p,q)<\epsilon\) are joined by a unique geodesic. Denote by \(M^\epsilon(f)\) the minimum number of fixed points of maps \(g\) with \(d(f,g)<\epsilon\). Two points \(x,y\) are called equivalent if there is a component of \(\Delta^\epsilon(f):=\{z\in X|\, d(z,f(z))<\epsilon\}\) which contains both \(x\) and \(y\). This leads to the notion of \(\epsilon\)-fixed point class. An \(\epsilon\)-fixed point class is called essential if \(\text{ind}(f;C)\not=0\) were \(C\) is the component of \(\Delta^\epsilon(f)\) that contains \(F\). The number of essential \(\epsilon\)-fixed point classes is called the \(\epsilon\)-Nielsen number \(N^\epsilon(f)\). The \(\epsilon\)-core of \(f\) is defined as the set of maps \(g\) such that \(d(f,g)<\epsilon\) and \(d(f(x),x)<\epsilon\) iff \(d(g(x),x)<\epsilon\). It is easy to see that \(N^\epsilon(f)=N(f)\) iff every fixed point class of \(f\) contains at most one essential \(\epsilon\)-fixed point class of \(f\). Now fix a covering \(p:\tilde{X}\to X\). Then \(p\) is called sheet-wise isometric if each \(x\in X\) possesses an evenly covered neighbourhood such that \(p\) maps each of \(p^{-1}(U)\) isometrically onto \(U\). Finally, let \(\tilde{f}\) be a lift of \(f\) and let \(F\) be an \(\epsilon\)-fixed point class of \(f\) and \(\tilde{F}\) an \(\epsilon\)-fixed point class of \(\tilde{f}\) with \(p(\tilde{F})=F\). Define \(J_\epsilon(F)\) to be the cardinality of \(p^{-1}(\{x\})\cap\tilde{F}\) for \(x\in F\). Similarly, define \(S_\epsilon(\tilde{f})\) to be the number of \(\epsilon\)-fixed point classes of \(\tilde{F}\) which are mapped onto \(F\). A Reidemeister class is a conjugacy class of lifts of \(f\) and a lift taken from a Reidemeister class is called a Reidemeister representative. In the following situation the author obtains a formula computing \(N^\epsilon(f)\): Assume that \(f\) has only finitely many fixed points and let \(\tilde{f}_1,\dotsc,\tilde{f}_r\) be the Reidemeister representatives of \(f\) corresponding to the nonempty Nielsen fixed point classes of \(f\). Assume that: (1) \(p\) is a sheet-wise isometric covering map, (2) the path components of \(\Delta^\epsilon(\tilde{f}_i)\) are simply connected for \(i=1,\dotsc,r\), (3) \(J_\epsilon(F)\) is the same for all \(\epsilon\)-fixed point classes in the same Nielsen fixed point class. Then \(N^\epsilon(f)=\sum_{i=1}^r\frac{N^\epsilon(\tilde{f}_i)}{S_\epsilon(\tilde{f}_i)}\). Reviewer's remark: As a matter of fact, the author introduces a normal subgroup \(K\) of \(\pi_1(X)\) in the formulation of the theorem. There are, however, no conditions on \(K\), neither does the result depend on \(K\). A reader who wishes to find out what is really going on here should consult \textit{J. Jezierski}'s article [Fixed Point Theory Appl. 2006, Spec. Iss., Article 37807, 11 p. (2006; Zbl 1097.55002)].