Paul M. N. Feehan,Thomas G. Leness
Abstract:We prove an analogue of the Kotschick-Morgan conjecture in the context of SO(3) monopoles, obtaining a formula relating the Donaldson and Seiberg-Witten invariants of smooth four-manifolds using the SO(3)-monopole cobordism. The main technical difficulty in the SO(3)-monopole program relating the Seiberg-Witten and Donaldson invariants has been to compute intersection pairings on links of strata of reducible SO(3) monopoles, namely the moduli spaces of Seiberg-Witten monopoles lying in lower-level strata of the Uhlenbeck compactification of the moduli space of SO(3) monopoles [<a class="link-https" data-arxiv-id="dg-ga/9710032" href="https://arxiv.org/abs/dg-ga/9710032">arXiv:dg-ga/9710032</a>]. In this monograph, we prove --- modulo a gluing theorem which is an extension of our earlier work in [<a class="link-https" data-arxiv-id="math/9907107" href="https://arxiv.org/abs/math/9907107">arXiv:math/9907107</a>] --- that these intersection pairings can be expressed in terms of topological data and Seiberg-Witten invariants of the four-manifold. This conclusion is analogous to the Kotschick-Morgan conjecture concerning the wall-crossing formula for Donaldson invariants of a four-manifold with $b_2^+=1$; that wall-crossing formula and the resulting structure of Donaldson invariants for four-manifolds with $b_2^+=1$ were established, assuming the Kotschick-Morgan conjecture, by Goettsche [<a class="link-https" data-arxiv-id="alg-geom/9506018" href="https://arxiv.org/abs/alg-geom/9506018">arXiv:alg-geom/9506018</a>] and Goettsche and Zagier [<a class="link-https" data-arxiv-id="alg-geom/9612020" href="https://arxiv.org/abs/alg-geom/9612020">arXiv:alg-geom/9612020</a>]. In this monograph, we reduce the proof of the Kotschick-Morgan conjecture to an extension of previously established gluing theorems for anti-self-dual SO(3) connections (see [<a class="link-https" data-arxiv-id="math/9812060" href="https://arxiv.org/abs/math/9812060">arXiv:math/9812060</a>] and references therein). Since the first version of our monograph was circulated, applications of our results have appeared in the proof of Property P for knots by Kronheimer and Mrowka [<a class="link-https" data-arxiv-id="math/0311489" href="https://arxiv.org/abs/math/0311489">arXiv:math/0311489</a>] and work of Sivek on Donaldson invariants for symplectic four-manifolds [<a class="link-https" data-arxiv-id="1301.0377" href="https://arxiv.org/abs/1301.0377">arXiv:1301.0377</a>].
What problem does this paper attempt to address?
This paper aims to solve the problem of the relationship between topological invariants of four - dimensional manifolds in the field of mathematics, especially the relationship between Donaldson invariants and Seiberg - Witten invariants. Specifically, the authors attempt to prove an SO(3)-monopole cobordism formula, which can relate the Donaldson invariants of a smooth four - dimensional manifold to the Seiberg - Witten invariants of this manifold. This formula is as follows:
\[ D^w_X(h^{\delta-2m}x^m) = \sum_{s \in \text{Spin}^c(X)} (-1)^{\frac{1}{4}(w-\Lambda+c_1(s))^2} \text{SW}_X(s) \times \sum_{i = 0}^{\min(\ell, [\delta/2]-m)} p_{\delta,\ell,m,i}(c_1(s),\Lambda) Q_X^i(h) \]
where:
- \( X \) is a closed, connected, oriented smooth four - dimensional manifold satisfying \( b_1(X)=0 \), and \( b_+(X) \) is odd and greater than 1.
- \( \Lambda, w\in H^2(X; \mathbb{Z}) \) satisfy \( w - \Lambda\equiv w_2(X)\ (\text{mod}\ 2) \).
- \( \delta, m \) are non - negative integers satisfying \( m\leq [\delta/2] \), and \( \delta\equiv -w^2-\frac{3}{4}(\chi+\sigma)\ (\text{mod}\ 4) \), while \( \delta < i(\Lambda) \), where \( i(\Lambda)=\frac{\Lambda^2 - 1}{4}(3\chi + 7\sigma) \).
- \( h\in H^2(X; \mathbb{R}) \), \( x\in H^0(X; \mathbb{Z}) \) are generators.
- \( Q_X \) is the intersection form on \( H^2(X; \mathbb{R}) \).
- \( \ell=\frac{1}{4}(\delta+(c_1(s)-\Lambda)^2+\frac{3}{4}(\chi+\sigma)) \).
- \( p_{\delta,\ell,m,i}(\cdot,\cdot) \) is a homogeneous polynomial whose coefficients are universal functions of \( \chi, \sigma, c_1(s)^2, \Lambda^2, c_1(s)\cdot\Lambda, \delta, m, \ell \).
The main technical difficulty in the paper lies in calculating the intersection numbers of the links of certain singular strata in the Uhlenbeck compactification moduli space. By introducing the local gluing theorem, the authors solve this problem and finally prove the above formula. This result not only provides a strict mathematical proof for the Witten conjecture, but also lays the foundation for the proof of other related conjectures (such as the Kotschick - Morgan conjecture).