Some remarks on Garay's conjecture (Q1187244)

Let \(M\) be a compact subset of a metric space \(X\). A function \(f: X\to [0,+\infty)\) is called distance-like with respect to \(M\) iff it vanishes precisely on \(M\) and for any \(\{x_ n\}\subset X\) the condition \(\{f(x_ n)\to 0\) implies that \(d(x_ n,M)=\inf\{d(x_ n,y): y\in M\}\to 0\). We will say that the pair \((X,M)\) satisfies the condition (G), (Garay's condition), iff for every function \(f\) distance-like with respect to \(M\) there exists a metric \(\rho\) on \(X\) equivalent to the given \(d\) such that \(\rho\mid_ M=d\mid_ M\) and \(\rho(x,M)=f(x)\) for every \(x\in X\). \textit{B. M. Garay} [Nonlinear Anal., Theory Methods Appl. 8, 1033-1042 (1984; Zbl 0552.54030)] conjectured that the compact \(M\) satisfies (G) iff it is a retract of \(X\). In the present paper a partial positive answer to Garay's conjecture is given: namely the author proved that if \(M\) is compact in a Banach space \(X\) satisfying the condition (G), then it is acyclic in the sense of the Alexander-Spanier cohomology. Moreover, the condition (G') stronger than (G) is studied.