Positive representations of \(L^{1}\) of a vector measure (Q2454886): Difference between revisions

For a positive measure \(\mu\) and \(f\in L_1 (\mu)\), it is elementary that \(\| f\| =\int | f| \:d\mu\). If \(\mu\) has its values in a Banach lattice, it is called positive if its range is a subset of the positive cone \(C_X^+\). In this case, we have \(\| f\| =\| \int | f| \:d\mu\| _X\). The integral \(\int f\:d\mu\) is to be considered as the Bartle-Dunford-Schwarz integral with respect to the countably additive \(X\)-valued measure \(\mu\). If \(\mu\) is no longer positive, it could still be true that the formula above gives an equivalent norm (or quasi-norm) on \(L_1 (\mu)\). In Section 2, it is proved that the formula gives an equivalent quasi-norm if and only if \(\mu:\Sigma\rightarrow X\) is a cone-open transformation of a positive measure, that is, there are a Banach lattice \(Z\), a positive measure \(\nu:\Sigma\rightarrow Z\) and a cone-open operator \(S:Z\rightarrow X\) such that \(\mu=S\circ \nu\). If the formula gives an equivalent (quasi-) norm on \(L_1 (\mu)\), one may wonder whether \(X\) can be re-ordered in some way such that \(\mu\) becomes positive with respect to this new order structure. A measure which becomes positive under a re-ordering that turns \(X\) into an \(M\)-normed Riesz space is called pseudo-positive. In Section 3, it is proved that a measure is pseudo-positive if and only if it is a cone-open and lattice-generating transformation of a positive measure. In particular, thus, for pseudo-positive measures \(\mu\), \(\| \int | f| \:d\mu\| _X\) gives an equivalent quasi-norm for \(L_1(\mu)\).