12.07.2015 Views

Ivancevic_Applied-Diff-Geom

Ivancevic_Applied-Diff-Geom

Ivancevic_Applied-Diff-Geom

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

760 <strong>Applied</strong> <strong>Diff</strong>erential <strong>Geom</strong>etry: A Modern Introductionline has a natural orientation. The d + d ∗ operator is independent of A andM (as the gauge group is Abelian), and its determinant line is trivializedonce and for all by picking an orientation of H 1 (X, R) ⊕ H 2,+ (X, R). Notethat this is the same data needed by Donaldson [Donaldson (1987)] to orientinstanton moduli spaces for SU(2); this is an aspect of the relation betweenthe two theories.If one replaces L by L −1 , A by -A, and M by ¯M, the monopole equationsare invariant, but the trivialization of the determinant line is multiplied by(−1) δ with δ the Dirac index. Hence the invariants for L and L −1 arerelated byn −x = (−1) ∆ n x .For W < 0, the moduli space is generically empty. For W > 0 one cantry, as in Donaldson theory, to define topological invariants that involveintegration over the moduli space. Donaldson theory does not detect thoseinvariants at least in known situations. We will see below that even whenW > 0, the moduli space is empty for almost all x.4.14.10.1 Topological InvarianceIn general, the number of solutions of a system of equations weighted bythe sign of the determinant of the operator analogous to T is always atopological invariant if a suitable compactness holds. If as in the caseat hand one has a gauge invariant system of equations, and one wishesto count gauge orbits of solutions up to gauge transformations, then onerequires (i) compactness; and (ii) free action of the gauge group on thespace of solutions.Compactness fails if a field or its derivatives can go to infinity. Toexplain the contrast with Donaldson theory, note that for SU(2) instantonscompactness fails precisely because an instanton can shrink to zero size.This is possible because(i) the equations are conformally invariant,(ii) they have non–trivial solutions on a flat R 4 , and(iii) embedding such a solution, scaled to very small size, on any fourmanifoldgives a highly localized approximate solution of the instantonequations (which can sometimes be perturbed to an exact solution). Themonopole equations by contrast are scale invariant but they have no nonconstantL 2 solutions on flat R 4 (or after dimensional reduction on flat R nwith 1 ≤ n ≤ 3). So there is no analog for the monopole equations of the

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!