No label defined (Q4249376)

A space \(F\) is called elliptic if \(H^*(F; Q)\) and \(\pi_*(F)\otimes Q\) are finite dimensional \( Q\)-vector spaces, and it is called positively elliptic if it is elliptic and has positive Euler characteristic. The conjecture of Halperin says that any fibration \(F \to E \to B\) with fibre \(F\) positively elliptic is totally non-cohomologous to zero, i.e.~the induced map \(H^*(E; Q) \to H^*(F; Q)\) is onto. In the paper the author does not prove the conjecture but shows various related theorems. For example, he shows that the Halperin conjecture would imply that any fibration with base a wedge of odd-dimensional spheres and fibre \(F\) positively elliptic is rationally trivial. He also shows for this kind of fibrations that the three numerical invariants: rational Toomer's invariant \(e_0(E)\), rational Lyusternik-Shnirel'man category \(\text{cat}_0(E)\) and rational cone-length \(cl_0(E)\) all agree (and differ by one from the corresponding value for \(F\)). Using the previously mentioned result this would follow from Halperin's conjecture. Therefore the last result can be regarded as a proof of a weakened version of the Halperin conjecture. The author also gives two new equivalent formulations of the Halperin conjecture using the concept of formal spaces. For the entire collection see [Zbl 0906.00019].