Conical combination¶

A finite nonnegative linear combination of vectors; the collection of all such combinations is their conical hull, the minimal convex cone generated from the origin.

Version: v1 · 2026-09-28 · History
Domain-specific #: 8648
Domain group: Formal Sciences
Origin domain: Mathematics
Subdomains: Convex Geometry, Linear Algebra → Mathematics

Core Idea¶

Conical combination retains addition and positive scaling while forbidding signed cancellation. Each result is assembled from declared generators with coefficients at least zero. The empty or all-zero combination places the origin in the generated hull.

The construction sits between a union of rays and a linear span. Addition fills mixtures between rays, but absence of negative weights preserves directionality. Normalizing a nonzero coefficient vector shows the close relationship to convex combinations without collapsing the distinction.

How would you explain it like I'm…

Stretch and Add, No Backwards

You have a few arrows. You can make each arrow longer or shorter, but you can never flip it to point backwards, and you can put arrows end to end. Everywhere you can reach this way, plus staying right at the start by using no arrows, is made by a conical combination.

Adding Arrows With No Minus Signs

Imagine a few arrows all starting at the same spot. A conical combination means you stretch or shrink each arrow by some amount that is zero or more, then add them together tip to tail. You are not allowed to use negative amounts, so you can never flip an arrow around to cancel out another one. Using zero of everything keeps you at the starting point. All the spots you can reach form a region that fans out between the arrows.

Nonnegative Linear Combination

A conical combination of vectors v_1, ..., v_k is a sum a_1 v_1 + ... + a_k v_k where every coefficient a_i is at least zero. It keeps addition and positive scaling but forbids negative coefficients, so signed cancellation is impossible. The set of all conical combinations of some generators is their conical hull, which always contains the origin, since all coefficients can be zero. It sits between two neighbors: a union of rays, which lets you scale each vector but not add them, and a linear span, which allows negative coefficients and so fills whole lines and planes. It is also close to a convex combination, whose coefficients are nonnegative and add to one; dividing a nonzero set of conical coefficients by their sum gives such a combination, but the two ideas are not the same.

A Conical combination of generators v₁, …, v_k is a vector of the form Σ λ_i v_i with every λ_i ≥ 0. It preserves addition and nonnegative scaling but forbids signed cancellation, which is what distinguishes it from a general linear combination. Taking all coefficients zero (or the empty combination) gives the origin, so the conical hull of any set of generators contains the origin. Structurally, the conical hull lies between the union of the rays generated by each vector and the linear span: addition fills in the mixtures between rays, while the absence of negative weights preserves directionality. Convex combinations are the special case where the coefficients also sum to one, and any nonzero conical coefficient vector can be normalized to that form, which shows the close relationship without erasing the difference, since the conical hull is closed under positive scaling while the convex hull is not.

Scope of Application¶

Convex optimization. Represents feasible directions and dual certificates.
Polyhedral geometry. Generates finitely described cones and studies their extreme rays.
Economics. Models nonnegative production combinations or activity levels.
Signal and data models. Expresses additive mixtures with nonnegative amplitudes.
Theorem alternatives. States membership in a generated cone versus separation by a linear functional.

Clarity¶

State the ambient vector space, scalar field, generator set, coefficient sign constraint, and whether finite, closed, or pointed cones are intended. For membership, exhibit nonnegative coefficients; for exclusion, use separation or an invariant incompatible with the generated cone. Inclusion test: Require a fixed real vector space, finite sums of stated generators, and nonnegative scalar weights with no normalization of their total. Exclusion test: Exclude affine combinations whose weights sum to one, convex combinations with nonnegative weights summing to one, arbitrary linear combinations with signed weights, and unions of generating rays not closed under addition. Nearest boundary: A convex combination fixes the coefficient sum at one and remains within the convex hull; a conical combination permits any nonnegative total and therefore scales the convex combination radially from the origin. Exit condition: The identity changes when negative coefficients are permitted or when total weight is constrained to one. Common misclassifications: It is not an arbitrary linear combination. It is not a convex combination unless the weights also sum to one. It is not merely the union of rays through each generator. It is not automatically a topologically closed cone. Nearest named distinctions: Convex Combination: Both use nonnegative weights, but a convex combination requires their sum to equal one. Linear Span: A span permits negative coefficients and therefore cancellation and opposite directions. Convex Cone: A convex cone is any set closed under conical combinations; the conical hull is the smallest one generated by a given set. Positive Linear Dependence: Dependence concerns a nontrivial conical relation summing to zero rather than any generated point.

Manages Complexity¶

The abstraction compresses infinitely many feasible mixtures into a finite or described generator set plus a nonnegativity rule. It separates directional generation from normalized interpolation and signed span, allowing the correct geometry and algorithms to be selected.

Abstract Reasoning¶

Fix the ambient vector space and source generators.
Introduce one nonnegative coefficient per selected generator.
Form their finite weighted sum.
Collect all possible sums to obtain the conical hull.
Test closure under addition and nonnegative scaling and verify minimality.
Distinguish algebraic hull from its topological closure when limits matter.

Knowledge Transfer¶

The transferable cargo is closure under addition and nonnegative scaling from chosen generators. It transfers to resources, rays, and additive mixtures with meaningful directionality; it stops when negative cancellation or unit-normalized interpolation is constitutive.

Neighborhood in Abstraction Space¶

Conical combination sits in a moderately populated region (47^th percentile for distinctiveness): it has near-neighbors but no dense thicket of look-alikes.

Family — Polynomial Rings & Rational Approximants (15 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-08