Subvalue Maps: Valuation, Geometry & Games

• Subvalue maps are context-dependent mathematical constructs that decompose or refine global valuation information, enabling controlled loss of detail.
• They are applied in classical valuation theory for coarsening valuations, in simplicial settings for summing local contributions, and in hybrid games for specifying winning policies.
• Their versatile use across algebra, geometry, and control synthesis illustrates a common pattern of reconstructing global properties from structured local data.

Subvalue maps are context-dependent mathematical objects used to decompose, refine, or constrain valuation-like information. In the literature considered here, the term occurs in at least three closely related but non-identical settings: classical valuation theory, where a subvalue map composes with a valuation to produce a coarser one; supertropical algebra, where the lattice of supervaluations is presented as a refinement of the classical poset of subvalue maps; and hybrid game control, where a subvalue map assigns formulas to subgames and thereby represents a nondeterministic winning policy or control envelope. Related work on valuations on simplicial maps uses the language of “subvalue map” for local contributions on open simplices whose sum yields the global valuation, providing a geometric and topological analogue of decomposition into subparts (Izhakian et al., 2010, Staecker et al., 2014, Kabra et al., 8 Aug 2025).

1. Terminological scope and principal uses

The supplied literature does not treat “subvalue map” as a single universally fixed formalism. Instead, it uses the term in multiple technical settings, each organized around a decomposition or coarsening principle. In valuation theory, a subvalue map acts on the codomain of a valuation. In simplicial valuation theory, it refers to contributions attached to open simplices. In hybrid games, it is a compositional map from subgames to formulas describing winning regions.

Setting	Object called or related to a subvalue map	Role
Classical valuation theory	A map from a value object such as $T$ to a coarser bipotent monoid $N$	Produces $w=\chi\circ v$
Valuations on simplicial maps	Contributions on open simplices	Sum to the valuation on the whole complex
Hybrid games	A map from subgames to formulas	Represents a nondeterministic winning policy

This distribution of usage suggests that “subvalue map” is best understood functionally rather than terminologically: in each setting it captures structure below the level of a global value, whether by local contribution, codomain coarsening, or subgame-wise winning characterization.

2. Classical valuation-theoretic usage

In the valuation-theoretic setting discussed in the supertropical literature, subvalue maps belong to the classical theory of coarsenings of valuations. The stated formulation is that, in classical valuation theory, especially for totally ordered groups $T$, subvalue maps from $T$ to another bipotent monoid $N$ classify coarser valuations dominated by $v$. Given a valuation $v:R\to T$, a subvalue map $\chi:T\to N$ leads to $w=\chi\circ v:R\to N$ (Izhakian et al., 2010).

This places subvalue maps at the level of value objects rather than at the level of the original ring or semiring. Their function is not to alter the algebraic source $N$0, but to compress or quotient the information carried by $N$1. The resulting valuation $N$2 is “less refined” in the precise sense that it forgets distinctions visible in $N$3 but not retained in $N$4. A plausible implication is that subvalue maps encode a controlled loss of resolution in valuation data.

The same source explicitly contrasts this classical perspective with later supertropical refinements. There, the poset of subvalue maps is treated as a baseline object that can be lifted into a richer algebraic environment. This contrast is important because it prevents a common misconception: subvalue maps are not, in this literature, synonymous with supervaluations. Rather, supervaluations refine the classical picture built from valuations and their coarsenings.

3. Supertropical refinement and the lattice of covers

The paper on supertropical semirings and supervaluations develops a much richer object than the classical subvalue map. A valuation $N$5 is interpreted as a map into a bipotent semiring $N$6, and a supervaluation $N$7 covers $N$8 when its ghost map satisfies $N$9 for all $w=\chi\circ v$0. The set $w=\chi\circ v$1 of equivalence classes of such supervaluations carries a natural ordering by dominance: $w=\chi\circ v$2 if there exists a transmission $w=\chi\circ v$3 with $w=\chi\circ v$4. With this order, $w=\chi\circ v$5 is a complete lattice; the bottom element is the valuation $w=\chi\circ v$6 itself, and the top element is an initial cover $w=\chi\circ v$7 (Izhakian et al., 2010).

Within this framework, the relation to subvalue maps is stated explicitly: the entire lattice $w=\chi\circ v$8 is a “lifting” of the classical poset of subvalue maps, i.e. quotients of the value group by convex subgroups, to a richer algebraic context that records tangible versus ghost elements and allows multiple tangible fibers over a given value (Izhakian et al., 2010). For fields, the refinement becomes especially concrete. If $w=\chi\circ v$9 is a field and $T$0 is a surjective Krull valuation, every equivalence class in $T$1 corresponds to a subgroup $T$2 of the unit group $T$3 of the valuation ring, and the lattice $T$4 is anti-isomorphic to the lattice of subgroups of $T$5, augmented by one element at the top.

The quotient construction by MFCE-relations makes this refinement explicit. Given a tangibly surjective supervaluation $T$6, every dominated cover arises as $T$7 for an MFCE-relation $T$8 on $T$9. The paper further states that, for fields, all MFCE-relations except the coarsest are orbital. In this sense, supervaluations do not replace subvalue maps; they generalize the coarsening picture from value-group quotients to semiring-level structural quotients. This suggests that the classical notion remains the appropriate coarse shadow of a more elaborate supertropical geometry.

4. Simplicial-map valuations and decomposition into local contributions

The paper “A Hadwiger Theorem for Simplicial Maps” does not define “subvalue map” as a standalone formal object, but it gives a directly related decomposition principle. A valuation on admissible pairs $T$0, where $T$1 is a polyhedron and $T$2 is a simplicial map, is additive over polyhedra: $T$3 Such a valuation extends to open simplices by

$T$4

and then satisfies

$T$5

The paper’s summary explicitly connects this with classical subvalue-map language: classically, a valuation on sets gives rise to a “subvalue map,” in the sense that each subset, such as an open simplex, contributes a summand and the valuation on the whole is the sum (Staecker et al., 2014).

The simplicial-map setting generalizes this viewpoint by attaching not merely a geometric contribution, but a signed contribution modulated by the map $T$6. The simplex property reads

$T$7

where $T$8 is $T$9 if $N$0, $N$1 if $N$2 preserves orientation, and $N$3 if $N$4 reverses orientation. The resulting Lefschetz volumes are defined by

$N$5

with $N$6 the intrinsic volume of the closed simplex $N$7 (Staecker et al., 2014).

Two identifications organize their significance. First, when $N$8, $N$9 is the classical Lefschetz number. Second, when $v$0, $v$1, recovering intrinsic volumes. The paper therefore states that Lefschetz volumes generalize both intrinsic volumes and Lefschetz numbers, and that they “substantially generalize” the idea of subvalue maps by incorporating data from how $v$2 acts on simplices, not just the combinatorics of the complex (Staecker et al., 2014). The associated Hadwiger-style theorem classifies invariant, continuous valuations with the simplex property as linear combinations of the Lefschetz volumes, making the local-contribution picture complete at the level of basis theorems.

5. Hybrid-game subvalue maps

In the hybrid-game setting, subvalue maps are defined directly and centrally. The paper “Hybrid Game Control Envelope Synthesis” introduces them as compositional representations of nondeterministic winning policies for hybrid games. Control envelopes are families of safe control solutions and correspond to nondeterministic winning policies; each deterministic specialization of such a policy is a control solution. A subvalue map represents such a policy by mapping each subgame of a hybrid game to a logical formula describing a set of states from which the player can still win, starting from the corresponding point in the game (Kabra et al., 8 Aug 2025).

This use of the term is syntactic and semantic at once. Syntactically, the domain is the AST of the hybrid game together with a distinguished “end” label. Semantically, the codomain consists of formulas specifying the winning region for the suffix of play that remains after a given subgame. The subvalue at “end” is the set of states for which the winning condition holds. The paper states that subvalue maps are analogous to value functions in reinforcement learning, but are defined compositionally over the subgames of an imperative, hybrid game (Kabra et al., 8 Aug 2025).

The policy interpretation is recursive over game constructs. For angelic assignment $v$3, Angel may choose any value $v$4 such that the successor formula holds after the assignment. For choice $v$5, Angel can choose left, right, or both depending on which successor formulas hold. For loops $v$6, Angel can iterate if the loop formula holds or exit if the successor formula holds. The same paper develops two projections for logical characterization: an existential projection $v$7, meaning there exists a strategy constrained by the subvalue map that wins, and a universal projection $v$8, meaning that all constrained strategies win. The projection construction inserts tests or guards so that all control decisions respect the formulas assigned by the subvalue map (Kabra et al., 8 Aug 2025).

The resulting interpretation differs sharply from the valuation-theoretic one. Here a subvalue map is not a coarsening of values but a compositional specification of the admissible future of play. This suggests that the shared name reflects a common methodological pattern—local objects indexed by substructure that determine a global notion—rather than a common ontology.

6. Inductiveness, maximality, ordering, and synthesis

The hybrid-game paper distinguishes plain subvalue maps from inductive subvalue maps. Plain subvalue maps guarantee that every choice taken so far was safe, but they do not by themselves guarantee that the agent can continue without becoming stuck. Inductiveness addresses precisely this issue. The paper states that an inductive subvalue map requires, for every subgame, that the formula at a control point implies at least one branch or action remains open; for a choice $v$9, this is written as

$v:R\to T$0

For loops, $v:R\to T$1 must serve as an inductive invariant and must ensure a finite strategy within $v:R\to T$2 to exit the loop into the postcondition (Kabra et al., 8 Aug 2025).

Among all inductive subvalue maps, the paper defines a maximal, or model predictive, subvalue map and proves that it exists. This maximal solution is the most permissive one under the paper’s partial order on subvalue maps. That order compares permissiveness pointwise over subgames through reachability under the constrained game: $v:R\to T$3 is at least as permissive as $v:R\to T$4 when all states reachable at each subgame under $v:R\to T$5 and satisfying $v:R\to T$6 also satisfy $v:R\to T$7. The maximal subvalue map therefore corresponds to the largest winning envelopes at each subgame (Kabra et al., 8 Aug 2025).

These structural results support algorithmic synthesis. The paper describes a recursive synthesis algorithm: atomic cases compute and simplify the winning-region formula, possibly via quantifier elimination; compositional cases recursively solve subgames and combine the results according to the inductive rules; loop invariants are generated by heuristics or rewriting and then checked for inductiveness and validity; and policy extraction derives the nondeterministic policy from the subvalue formulas. The algorithm can be parameterized so that subformulas remain in efficient fragments such as propositional real arithmetic (Kabra et al., 8 Aug 2025). The paper further states that inductive subvalue maps always yield sound and runtime-enforceable control envelopes.

A common misconception would be to regard these hybrid-game subvalue maps as merely descriptive annotations. The source instead treats them as a full semantic interface between dGL verification and controller synthesis: they characterize winning regions subgame by subgame, induce nondeterministic winning policies, admit existential and universal logical projections, and support the existence theorem for a maximal permissive envelope.

7. Comparative significance

Across the three settings, subvalue maps and subvalue-map-like constructions serve as intermediate representations between local structure and global invariants. In classical valuation theory, the local/global interface is codomain-level coarsening: $v:R\to T$8 compresses the values of $v:R\to T$9. In simplicial valuation theory, it is decomposition into contributions from open simplices, with signed intrinsic-volume terms yielding global Lefschetz volumes. In hybrid games, it is structural recursion over subgames, where formulas at nodes describe winning regions for the remaining game.

The similarities are methodological rather than formal. Each framework uses a compositional object attached to “sub-” structure—subvalues of a valuation, sub-simplex contributions, or subgame formulas—to recover or constrain a global object. Each framework also supports an order or classification principle: classical subvalue maps classify coarser valuations; supervaluation theory organizes covers in the complete lattice $\chi:T\to N$0; simplicial-map valuations admit a Hadwiger-style basis theorem; and hybrid-game subvalue maps admit a maximal inductive solution under a permissiveness order (Izhakian et al., 2010, Staecker et al., 2014, Kabra et al., 8 Aug 2025).

At the same time, the differences are substantive. Classical and supertropical uses are algebraic; the simplicial use is geometric-topological; the hybrid-game use is logical and algorithmic. The literature therefore supports a careful, context-sensitive reading of the term. A plausible synthesis is that “subvalue map” names a recurrent research pattern: a structured assignment on subobjects that either reconstructs a global quantity or defines the admissible space of globally correct behavior.