6.23.5 Tensor Component Growth Pattern

Tensor Component Growth Pattern is the qualitative and quantitative character of the sequence of component counts N = d^n viewed as a whole, focusing on the shape of its growth curve — constant ratio between successive terms, strictly convex increase, and eventual dominance over polynomial sequences — rather than on the specific formula or on the separate roles of dimension and order individually. Where earlier topics established the formula and examined dimension and order one at a time, this topic characterizes the resulting sequence itself as a mathematical object, the kind of pattern it belongs to, and how it compares to other familiar growth patterns.

Identifying the Pattern as Geometric Growth

Constant Ratio Between Successive Terms

The defining feature of the component count sequence, for fixed dimension d and increasing order n, is that the ratio of consecutive terms is constant:

This constant-ratio property is precisely the definition of a geometric sequence, and it is what distinguishes the component count pattern from, for instance, a linear or quadratic growth pattern, where the ratio of successive terms changes as the sequence progresses.

Convexity of the Growth Curve

Because a geometric sequence with ratio d > 1 grows faster and faster in absolute terms even though its relative growth rate is fixed, plotting N(n) against n produces a strictly convex curve: each step upward is larger than the previous step, in contrast to a linear pattern (constant absolute increase) or a pattern with a diminishing rate of increase.

Comparing the Pattern to Other Growth Families

Geometric Versus Polynomial

A polynomial growth pattern, such as the count d² growing as a function of d at fixed order 2, has a shrinking ratio of consecutive terms as the base grows, ((d+1)/d)² → 1, whereas the geometric pattern in order maintains the exact ratio d forever. This is the formal reason that, for large enough order, the exponential-in-order growth of component count always eventually overtakes any polynomial-in-dimension growth, no matter how large the dimension.

Geometric Versus Factorial

The unconstrained component count dⁿ grows more slowly than a factorial-type sequence such as n!, but the component count of totally antisymmetric tensors — which involves a binomial coefficient C(d, n) — grows more slowly still and eventually reaches zero once n exceeds d. Placing the plain geometric pattern between these two extremes situates it precisely: unconstrained tensors grow purely geometrically in order, symmetry-constrained tensors grow sub-geometrically, and no natural tensor construction produces super-geometric (factorial-like) component growth in order.

Diagram of the Growth Pattern on a Logarithmic Scale

Why Recognizing the Pattern Matters

Predicting Behavior Without Recomputing

Once the growth pattern is recognized as geometric, the component count at any order can be predicted immediately from the count at any other order, without recomputing the underlying power formula from scratch: N(n₂) = N(n₁) · d^(n₂ − n₁). This shortcut is a direct consequence of the constant-ratio property and is useful whenever component counts must be compared or estimated across several related tensors of differing order.

Informing the Choice of Computational Strategy

Recognizing that component count follows a geometric — rather than merely "large" — growth pattern is what justifies specific algorithmic responses: geometric growth in a single controllable parameter (order) is the textbook signal for seeking a recursive or divide-and-conquer algorithm, or for adopting compressed tensor representations that avoid the geometric blow-up altogether, since a pattern this regular is far more amenable to structural exploitation than an irregular or unpredictable growth curve would be.