Coincidences of fibrewise maps between sphere bundles over the circle (Q2921042)

The authors continue the work of the first two authors on Nielsen coincidence theory. Here, they consider sphere bundles \(M\) and \(N\) over \(B=S^1\) with orthogonal gluing maps \(A_M:F_M\to F_M\) and \(A_N:F_N\to F_N\) of the fibres. Let \(m-1\) and \(n-1\) denote the dimensions of the spheres in the respective fibers. Let \(f_1,f_2:M\to N\) be fibrewise maps and let \(C(f_1,f_2):=\{x\in M|\;f_1(x)=f_2(x)\}\) and call \(\text{MCC}_B(f_1,f_2)\) the minimum number of path components of \(C(f'_1,f'_2)\) where \(f'_1\) (\(f'_2\)) is obtained by a fibrewise deformation from \(f_1\) (\(f_2\)). The authors address three problems: (1) Can \(C(f_1,f_2)\) be made empty by fibrewise homotopies of \(f_1\) and \(f_2\)? (2) Does \(\text{MCC}_B(f_1,f_2)\) equal the Nielsen number \(N_B(f_1,f_2)\) introduced by the first two authors in [Topol. Methods Nonlinear Anal. 33, No. 1, 85--103 (2009; Zbl 1178.55002)] for all fibrewise maps \(f_1,f_2:M\to N\)? (3) Can one classify the fibrewise homotopy classes of fibrewise maps or get estimates or bounds for their number? As to (1) the answer is positive if \(f_2\) is fibrewise homotopic to \(a\circ f_1\) where \(a\) denotes the fibrewise antipodal map. As to (2) denote by \(\mathcal{F}\) the set of fibrewise maps \(M\to N\) modulo fibrewise homotopy. The first two authors [loc. cit.] introduced an obstruction \(\omega_B(f_1,f_2)\) in the normal bordism class \(\Omega_{m-n+1}(M;\phi)\).NEWLINENEWLINE Then \(\text{MCC}_B(f_1,f_2)=N_B(f_1,f_2)\) for all \([f_1],[f_2]\in\mathcal{F}\) iff the map \(\text{deg}_B:\mathcal{F}\to\Omega_{m-n+1}(M;\phi)\) which sends \([f]\) to \(\omega_B(f,a\circ f_0)\) is injective where \(f_0\) is the map which maps each fibre in \(M\) to the point \((1,0,\dotsc,0)\) in the corresponding fibre of \(N\). The details are highly technical and require a lot of machinery from homotopy theory.