---
title: Riesz-Kantorovich Formulas in Operator Lattices
url: https://www.emergentmind.com/topics/riesz-kantorovich-formulas
type: topic
---

# Riesz-Kantorovich Formulas in Operator Lattices

The Riesz-Kantorovich formulas provide explicit lattice operation expressions—suprema, infima, and absolute value—for order-bounded linear or module homomorphisms between vector lattices or broader ordered algebraic structures. These formulas encapsulate the supremum and infimum of two operators in terms of pointwise suprema or infima over decompositions of input elements. Their validity and extensions illuminate the structure of spaces of operators and underpin more advanced developments in ordered functional analysis, module theory, and multi-wedged vector spaces.

## 1. Structural Background and Key Definitions

Let $\mathbb{L}$ be a Dedekind complete unital $f$-algebra—a real vector lattice with associative, commutative multiplication and a multiplicative unit such that, for $0 \leq a \leq b$ and $c \geq 0$ in $\mathbb{L}$, $ca \leq cb$. A partially ordered $\mathbb{L}$-module $X$ is an $\mathbb{L}$-module that is also a partially ordered real vector space, compatible in the sense that $x \leq y$ implies $x+z \leq y+z$ and $0 \leq \lambda \in \mathbb{L}$, $x \leq y$ implies $\lambda x \leq \lambda y$. $X$ is *directed* if every pair has an upper bound.

The **Riesz Decomposition Property (RDP)** holds for $X$ if for all $x,y \in X^+$, $[0,x+y] = [0,x] + [0,y]$, mirroring the classical requirement from vector lattice theory.

An $\mathbb{L}$-vector lattice is an $\mathbb{L}$-module $X$ that is also a lattice with respect to its partial order and compatible with $\mathbb{L}$-module operations.

## 2. The Riesz–Kantorovich Theorem and Lattice Formulas

Let $X$ be a directed partially ordered $\mathbb{L}$-module with RDP and $Y$ a Dedekind-complete $\mathbb{L}$-vector lattice satisfying either:
- Sequentially $P$-Archimedean (for every decreasing sequence of idempotents in $\mathbb{L}$),
- or $\mathbb{L} = \mathbb{L}_1$, the ideal generated by the unit.

The set $\mathbb{L}_b(X,Y)$ of order-bounded $\mathbb{L}$-module homomorphisms coincides with the space of regular operators $\mathbb{L}_r(X,Y)$ (differences of positive operators) and is itself a Dedekind-complete $\mathbb{L}$-vector lattice under the pointwise order.

For $S, T \in \mathbb{L}_b(X,Y)$ and $x \in X^+$, the lattice operations are given by the Riesz–Kantorovich formulas:
\[
\begin{align*}
(S\vee T)(x) &= \sup \{ S(y) + T(x-y) : 0 \leq y \leq x \} \\
(S\wedge T)(x) &= \inf \{ S(y) + T(x-y) : 0 \leq y \leq x \} \\
S^+(x) &= (S\vee 0)(x) = \sup\{ S(y) : 0 \leq y \leq x \} \\
S^-(x) &= -(S\wedge 0)(x) = \inf\{ S(y) : 0 \leq y \leq x \} \\
|S|(x) &= S^+(x) + S^-(x) = \sup\{ S(y) : -x \leq y \leq x \}
\end{align*}
\]
If $(T_\alpha)$ is an increasing net in $\mathbb{L}_b(X,Y)$ bounded above by $R$, the supremum $T = \sup_\alpha T_\alpha$ satisfies $T(x) = \sup_\alpha T_\alpha(x)$ for all $x \in X^+$ [2602.03206].

## 3. Proof Techniques and Extension Lemmas

The proof of the Riesz–Kantorovich formulas in the $\mathbb{L}$-module context extends the classical outline, introducing crucial new elements:

- **Extension Lemma (Theorem 4.1 of [2602.03206])**: For directed $\mathbb{L}$-modules $X$, $Y$ (with $Y$ sequentially almost Archimedean or $\mathbb{L} = \mathbb{L}_1$), every additive, $P$-homogeneous map $T: X^+ \to Y^+$ extends uniquely to a positive $\mathbb{L}$-module homomorphism $\widehat{T}: X \to Y$.
    - The proof establishes $Q^+$-homogeneity from additivity and $P$-homogeneity, utilizes the Freudenthal Spectral Theorem, and approximates arbitrary positive $\lambda \in \mathbb{L}^{+}$ by idempotents and rational-step functions.
- The supremum operator $R(x) = \sup \{S(y) + T(x-y): 0 \leq y \leq x\}$ is additive and $P$-homogeneous on $X^+$, and its extension corresponds to the least upper bound of $S$ and $T$. The infimum follows dually or via negation.

Dedekind completeness of $\mathbb{L}_b(X,Y)$ is ensured by taking suprema over nets pointwise and applying the extension lemma.

## 4. Classical and Generalized Forms

When $\mathbb{L} = \mathbb{R}$, $P = \{0,1\}$ trivially. The result recovers the classical Riesz–Kantorovich context: if $X$ is a directed real vector space with RDP and $Y$ is a Dedekind-complete (Archimedean) vector lattice, $\mathfrak{L}_b(X,Y)$ is a Dedekind-complete vector lattice with formulae:
\[
(S\vee T)(x) = \sup \{ S(y) + T(x-y) : 0 \leq y \leq x \}
\]
and analogues for infima, positive and negative parts, and modulus [1212.6243].

The hypotheses have been further relaxed:
- **Local Riesz Decomposition**: If $X$ satisfies merely an $\mathcal{L}$-Riesz Decomposition Property for a suitable subspace of operators, the RK formula may still hold—expanding the range of operator lattices.
- **Topological Weakening**: For continuous functionals ($Y = \mathbb{R}$), under weak compactness of order intervals and the presence of an order unit or suitable topology, lattice structures and RK formulas persist [1212.6243].
- **Economics Applications**: When $X_+$ has an interior point (order unit), RK formulas yield linearity on the interior, crucial for decomposition theorems in economic theory.

## 5. Riesz–Kantorovich in Multi-Wedged and Generalized Settings

Schwanke–Wortel [1609.05833] have generalized the RK formulas to multi-wedged spaces—vector spaces equipped with collections of wedges (not necessarily cones), each associated with its own pre-order:
- **Multi-Wedged Space** $(E, \mathcal{W})$: $E$ a real vector space, $\mathcal{W}$ a family of wedges.
- **Multi-suprema and Multi-infima**: Generalizations of suprema and infima with respect to multiple wedges; the set of all multi-suprema of $\{(x_i, W_i)\}$ is characterized geometrically.
- **$(\alpha, \beta)$-Riesz Decomposition Property**: For finite or countable decompositions among multiple wedges, ensures the decomposition property generalizes as required for multi-orders.

The main result (Theorem 3.5 of [1609.05833]) asserts:
- If $(E,\mathcal{W})$ has the $(2, \kappa)$-Riesz decomposition property and $(F, \{V\})$ is Dedekind complete, then the space of $(W,V)$-positive operators $L_{W,V}(E, F)$ forms a $\kappa$-multi-lattice.
- Explicit formulas for multi-suprema of operators generalize the classical RK formula:
\[
\bigl(\msup_{i\in I}T_i\bigr)(x) = \sup\left\{\sum_{i\in I} T_i(y_i): y_i \in W_i,\, \sum_i y_i = x\right\}
\]
for $x$ in the appropriate multi-wedge sum.

## 6. Examples, Structural Failures, and Extensions

Instances where the RK formula fails directly illustrate the necessity of structural conditions:
- If $X$ lacks RDP, there may exist $z \in [0, x+y] \setminus ([0,x] + [0,y])$; hence, the superlinear map $R$ (from the RK formula) might not be additive, invalidating the identification of $S\vee T$ via formula.
- $\mathcal{L}$–RDP but not RDP: In $\mathbb{R}^3$ with an appropriately chosen cone, $\mathbb{R}^3$ exhibits an $\mathcal{L}$–RDP but fails the classical RDP; thus, the RK formula does not generally hold [1212.6243].

The RK formulas extend to dual multi-lattices (e.g., algebraic duals endowed with dual wedges) and function spaces with pointwise positivity structures [1609.05833].

## 7. Corollaries and Impact on Operator Lattices

The RK formulas and associated structure results imply:
- Order-bounded homomorphism spaces become Dedekind complete lattices when domain and codomain possess appropriate Riesz decomposition and completeness properties.
- Lattice operations (suprema/infima, modulus, positive/negative parts) in these operator lattices have concrete, manageable expressions in terms of action on positive cones.
- There are equivalences among various properties: lattice structure of operator spaces, decomposition/interpolation properties in the domain, and linearity of the RK transform.

Recent research demonstrates that substantial weakening of classical ordering and completeness conditions still preserves the utility and form of the Riesz–Kantorovich formulas, both for module-valued and multi-ordered contexts, and that these results yield foundational insights into the algebraic and order-theoretic underpinnings of spaces of operators [2602.03206, 1212.6243, 1609.05833].

Source: https://www.emergentmind.com/topics/riesz-kantorovich-formulas