Self-maps of spaces in fibrations (Q721498)

Let \(\mathcal E(X)\) denote the group of homotopy self-equivalences of a space \(X\) and let \[ N\mathcal E(X)=\min\Big\{k\,\mid\,\mathcal A_{\sharp}^{k}(X)=\mathcal E(X)\Big\} \] denote the self-closeness number of \(X\), where \(\mathcal A_{\sharp}^{k}(X)=\Big\{f\in[X,X]\,\mid\,f_{\sharp}:\pi_{i}(X)=\pi_{i}(X) \text{ for any } i\leq k\}\). Here \([X,X]\) denotes the monoid of homotopy classes of self-maps of the space \(X\). Let \(G\) be an abelian group and \(K(G, n)\) the Eilenberg-MacLane space. Consider the following type of fibration sequence: \[ \dots\to\Omega Y\overset{\Omega \gamma}{\longrightarrow}\Omega K(G,m+1)=K(G,m+1)\overset{q}{\longrightarrow}X\overset{i}{\longrightarrow}Y\overset{\gamma}{\longrightarrow}K(G,m+1) \] where \(m\geq 0\) and \(Y\) is a 0-connected CW-complex such that \(\pi_{d}(Y)=0\) for \(d\geq m+1\). The aim of the authors, in this paper under review, is to compare the two integers \(N\mathcal E(Y)\) and \(N\mathcal E(X)\). First, they prove the following fact: Assume that there exist maps \(f : Y \to Y , g : X \to X \) and \(\Omega_s:\Omega K(G,m+2)\to \Omega K(G,m+2)\) making the following diagram commutative, up to homotopy. \[ \begin{tikzcd} \Omega K(G,m+2) \arrow [r,"q"] \arrow[d,"\Omega_s"] & X \arrow[r,"i"] \arrow[d,"g"] & Y \arrow[r,"\gamma"] \arrow[d,"f"] & K(G,m+2)\arrow[d,"t"]\\ \Omega K(G,m+2) \arrow[r,"q"] & X \arrow[r,"i"] & Y \arrow[r,"\gamma"] & K(G,m+2)\end{tikzcd} \] Then there exists a map \(t: K(G,m+2)\to K(G,m+2)\) such that \(t\circ\gamma\simeq \gamma\circ f\) and \(s = a \circ t \circ b\) for some \(a, b\in\mathcal E(K(G,m + 2)).\) Next, as a consequence if we suppose that, \[ \gamma_{\sharp}([Y,Y])\subset\gamma^{\sharp}([K(G,m+2), K(G,m+2)]) \] then they show that \(N\mathcal E(Y)\leq N\mathcal E(X)\leq m+1\) and by adding some technical conditions, they prove that \(N\mathcal E(Y)= N\mathcal E(X)\). Finally using Sullivan models they obtain similar results for rational spaces.