Energy-Based Cobb-Douglas Models

• Energy-Based Cobb-Douglas (EBDC) represents a family of models that modify the traditional Cobb-Douglas function by incorporating energy as a direct input, physical constraint, or information proxy.
• Formulations range from logistic energy input in growth models to information-geometric aggregators and maximum entropy reconstructions, emphasizing bounded growth and sustainable thresholds.
• EBDC applications extend to economic growth, exoplanet habitability scoring, and cloud data center cost analysis, providing actionable insights into system optimization and sustainability.

Searching arXiv for the cited papers to ground the response in current paper records. Energy Based Cobb-Douglas (EBDC) denotes a family of Cobb-Douglas-type constructions in which energy enters the model either as an explicit production input, as a physical or thermodynamic constraint on micro-technology, as an information-theoretic or geometric “energy” associated with distortion minimization, or as an energy-related proxy within a multiplicative scoring function. In the most explicit usage, EBDC is the single-input logistic production function $X=f(E)$ introduced to extend the Solow-Swan model by replacing capital with energy production/conversion (Power et al., 12 Jun 2025). In adjacent literatures, the same label or closely related formulations refer to Cobb-Douglas as a low-distortion aggregator induced by Bregman divergences and KL-type generators (0801.0390, 0901.2586), to a macro regularity reconstructed from maximum entropy with physical/energy constraints (Liuh, 3 Dec 2025), and to domain-specific multiplicative scores in which escape velocity and surface temperature act as energy-related variables (Krishna et al., 2020). This suggests that EBDC is not a single standardized formalism but a cluster of energy-centered reinterpretations of Cobb-Douglas structure.

1. Conceptual range of the term

The cited literature uses “energy-based” in several non-equivalent senses. In growth theory, energy is the primary state variable of production. In information geometry, “energy” refers to generalized information quantities or transition costs. In statistical-physical reconstruction, energy appears through physical bounds on micro-level technical coefficients. In exoplanet habitability scoring, energy enters through gravitational and thermal proxies. The shared element is the retention of a Cobb-Douglas aggregation logic while changing the semantic status of its inputs or its underlying justification (Power et al., 12 Jun 2025, 0801.0390, Liuh, 3 Dec 2025, Krishna et al., 2020).

Context	Representative formula	Role of energy
Growth extension	$$X = \frac{N_1 E^\alpha}{C|N_2- E|^\alpha + E^\alpha}$$	Explicit input
Information geometry	$$\mu_\varphi = \nabla_\varphi^{-1}\!\left(\frac{1}{\Gamma}\sum_i \gamma_i \nabla_\varphi(x_i)\right)$$	Distortion or transition cost
Maximum entropy reconstruction	$$Y = \mathcal{A} X_1^{\gamma_1} X_2^{\gamma_2} K^{\gamma_K}$$	Physical/energy constraints on micro coefficients
Habitability scoring	$Y = R^\alpha D^\beta V_e^\delta T_s^\gamma$	Gravitational and thermal proxies

A recurrent misconception is to treat EBDC as synonymous with “Cobb-Douglas plus an energy variable.” The supplied literature does support that usage, but it also supports broader interpretations in which energy is encoded through thermodynamic ceilings, entropy-based inference, or information-geometric distortion (Power et al., 12 Jun 2025, Liuh, 3 Dec 2025, 0801.0390).

2. Logistic energy-input production function

The most direct formulation appears in the extension of the Solow-Swan model, where the standard Cobb-Douglas production function $Y = A K^\alpha L^\beta$ is replaced by a one-input “Energy-Based Cobb-Douglas (EBDC) Function” (Power et al., 12 Jun 2025). The paper first presents a multi-input logistic generalization,

$$Y = \frac{N_Y L^\alpha K^\beta}{C|N_L-L|^\alpha|N_K-K|^\beta + L^\alpha K^\beta},$$

and then specializes to the single-input energy version

$$X = f(E) = \frac{N_1 E^\alpha}{C|N_2- E|^\alpha + E^\alpha}.$$

Here $X$ is aggregate output/production, $E$ is energy production/conversion, and $$X = \frac{N_1 E^\alpha}{C|N_2- E|^\alpha + E^\alpha}$$0 are positive parameters; $$X = \frac{N_1 E^\alpha}{C|N_2- E|^\alpha + E^\alpha}$$1 acts as a carrying capacity, ensuring that the production function is bounded and logistic in shape (Power et al., 12 Jun 2025). This boundedness is central to the model’s departure from the unbounded classical Cobb-Douglas specification.

The corresponding dynamic law is

$$X = \frac{N_1 E^\alpha}{C|N_2- E|^\alpha + E^\alpha}$$2

where $$X = \frac{N_1 E^\alpha}{C|N_2- E|^\alpha + E^\alpha}$$3 is the savings or investment rate and $$X = \frac{N_1 E^\alpha}{C|N_2- E|^\alpha + E^\alpha}$$4 is depreciation or loss (Power et al., 12 Jun 2025). In this formulation, energy becomes “analogously central” to growth as capital was in the original Solow-Swan model. The paper also defines

$$X = \frac{N_1 E^\alpha}{C|N_2- E|^\alpha + E^\alpha}$$5

and, for $$X = \frac{N_1 E^\alpha}{C|N_2- E|^\alpha + E^\alpha}$$6 and $$X = \frac{N_1 E^\alpha}{C|N_2- E|^\alpha + E^\alpha}$$7, a net energy gain expression

$$X = \frac{N_1 E^\alpha}{C|N_2- E|^\alpha + E^\alpha}$$8

Steady states satisfy $$X = \frac{N_1 E^\alpha}{C|N_2- E|^\alpha + E^\alpha}$$9. For $$\mu_\varphi = \nabla_\varphi^{-1}\!\left(\frac{1}{\Gamma}\sum_i \gamma_i \nabla_\varphi(x_i)\right)$$0, the paper gives explicit nontrivial equilibria $$\mu_\varphi = \nabla_\varphi^{-1}\!\left(\frac{1}{\Gamma}\sum_i \gamma_i \nabla_\varphi(x_i)\right)$$1 in addition to the trivial state $$\mu_\varphi = \nabla_\varphi^{-1}\!\left(\frac{1}{\Gamma}\sum_i \gamma_i \nabla_\varphi(x_i)\right)$$2 (Power et al., 12 Jun 2025). The substantive interpretation is that growth becomes bounded, indefinite exponential growth is ruled out, and sustainable growth requires remaining above a critical net energy threshold. The paper states that over-exploitation leads to collapse and that “EROI regime” analysis identifies quantitative intervals in which energy investment is effective (Power et al., 12 Jun 2025).

3. Information-geometric and low-distortion formulations

A second line of work treats Cobb-Douglas not primarily as a technological law but as an information-theoretically optimal aggregator. In this framework, the central object is the Bregman divergence

$$\mu_\varphi = \nabla_\varphi^{-1}\!\left(\frac{1}{\Gamma}\sum_i \gamma_i \nabla_\varphi(x_i)\right)$$3

and the associated low-distortion aggregator (LDA)

$$\mu_\varphi = \nabla_\varphi^{-1}\!\left(\frac{1}{\Gamma}\sum_i \gamma_i \nabla_\varphi(x_i)\right)$$4

The claim is that the LDA is the unique minimizer of expected distortion, so it is information-optimal under the chosen Bregman geometry (0801.0390).

Within this scheme, Cobb-Douglas appears as a special case of an LDA. One paper identifies the Cobb-Douglas aggregator

$$\mu_\varphi = \nabla_\varphi^{-1}\!\left(\frac{1}{\Gamma}\sum_i \gamma_i \nabla_\varphi(x_i)\right)$$5

with the generator $$\mu_\varphi = \nabla_\varphi^{-1}\!\left(\frac{1}{\Gamma}\sum_i \gamma_i \nabla_\varphi(x_i)\right)$$6, linking it to the KL divergence (0801.0390). A closely related treatment states that Cobb-Douglas corresponds to the generator $$\mu_\varphi = \nabla_\varphi^{-1}\!\left(\frac{1}{\Gamma}\sum_i \gamma_i \nabla_\varphi(x_i)\right)$$7, and more generally that mainstream economic production functions are lowest distortion aggregators induced by appropriate convex generators (0901.2586).

This literature sharpens the relation between information geometry and classical economic restrictions. The supplied summary reports several structural results: if dual aggregators satisfy a fixed-point relation, both must be CES; if the sum of partial elasticities equals one, the aggregator must be CES; if substitution elasticity equals one, the aggregator must be Cobb-Douglas; and if homogeneity constraints are imposed, the admissible set collapses to Cobb-Douglas or CES (0801.0390). In this sense, the “energy-based” reading does not refer to a physical energy input, but to generalized information quantities that function as distortions, transition costs, or, by analogy in the paper, “work” (0901.2586).

The same geometric framework is extended to transition costs,

$$\mu_\varphi = \nabla_\varphi^{-1}\!\left(\frac{1}{\Gamma}\sum_i \gamma_i \nabla_\varphi(x_i)\right)$$8

and to geometric duality between Marshallian and Hicksian demands, Slutsky-type properties, and Roy-type properties (0901.2586). A plausible implication is that EBDC, in this strand, is best understood as an information-geometric reinterpretation of Cobb-Douglas rather than an energy-input production specification.

4. Maximum entropy, scale invariance, and physical constraints

A third formulation reconstructs macro Cobb-Douglas from statistical physics. The starting point is not a representative production technology but a distribution of micro-level fixed-proportion technologies characterized by technical coefficients $$\mu_\varphi = \nabla_\varphi^{-1}\!\left(\frac{1}{\Gamma}\sum_i \gamma_i \nabla_\varphi(x_i)\right)$$9 (Liuh, 3 Dec 2025). The reconstruction relies on three ingredients stated in the summary: scale invariance, physical/energy constraints, and the Maximum Entropy Principle.

Under constraints on logarithmic means and bounded support $$Y = \mathcal{A} X_1^{\gamma_1} X_2^{\gamma_2} K^{\gamma_K}$$0, maximizing Shannon entropy yields a truncated power-law density

$$Y = \mathcal{A} X_1^{\gamma_1} X_2^{\gamma_2} K^{\gamma_K}$$1

with

$$Y = \mathcal{A} X_1^{\gamma_1} X_2^{\gamma_2} K^{\gamma_K}$$2

The paper interprets the support bounds in physical terms: there is a maximal potential efficiency of energy use per unit output, or a minimal energy input per unit output, reflecting energy or physical limits in production (Liuh, 3 Dec 2025).

Aggregation over the profitably active region then produces

$$Y = \mathcal{A} X_1^{\gamma_1} X_2^{\gamma_2} K^{\gamma_K}$$3

and inversion yields the aggregate production function

$$Y = \mathcal{A} X_1^{\gamma_1} X_2^{\gamma_2} K^{\gamma_K}$$4

where

$$Y = \mathcal{A} X_1^{\gamma_1} X_2^{\gamma_2} K^{\gamma_K}$$5

with $$Y = \mathcal{A} X_1^{\gamma_1} X_2^{\gamma_2} K^{\gamma_K}$$6 (Liuh, 3 Dec 2025). Factor shares are constant: $$Y = \mathcal{A} X_1^{\gamma_1} X_2^{\gamma_2} K^{\gamma_K}$$7

A major theoretical consequence is that macro Cobb-Douglas-like substitution arises “entirely from the entry/exit of firms at different technical levels as relative prices change”—an extensive-margin effect rather than micro-level continuous substitution (Liuh, 3 Dec 2025). The same paper explicitly argues that exponents are not output elasticities in a physical or technological sense, but projections of micro-distributional parameters, and that macro production functions are “lossy compressions” of micro-level heterogeneity. This is the main critical counterpoint to naïve readings of EBDC as a direct description of production technology.

5. Energy-related Cobb-Douglas constructions outside macro growth

Energy-based Cobb-Douglas reasoning is also used outside macroeconomic production theory. In exoplanet science, the Cobb-Douglas Habitability Function is defined as

$$Y = \mathcal{A} X_1^{\gamma_1} X_2^{\gamma_2} K^{\gamma_K}$$8

where $$Y = \mathcal{A} X_1^{\gamma_1} X_2^{\gamma_2} K^{\gamma_K}$$9 is planetary radius, $Y = R^\alpha D^\beta V_e^\delta T_s^\gamma$0 planetary density, $Y = R^\alpha D^\beta V_e^\delta T_s^\gamma$1 escape velocity, and $Y = R^\alpha D^\beta V_e^\delta T_s^\gamma$2 mean surface temperature, all in Earth units (Krishna et al., 2020). The paper separates the score into

$Y = R^\alpha D^\beta V_e^\delta T_s^\gamma$3

subject to

$Y = R^\alpha D^\beta V_e^\delta T_s^\gamma$4

and combines them as

$Y = R^\alpha D^\beta V_e^\delta T_s^\gamma$5

typically with $Y = R^\alpha D^\beta V_e^\delta T_s^\gamma$6 (Krishna et al., 2020).

The energy content of this model is explicit but indirect. Escape velocity is said to combine the planet’s gravitational energy with its radius and mass, while surface temperature reflects the energy available at a planet’s surface (Krishna et al., 2020). The paper therefore treats $Y = R^\alpha D^\beta V_e^\delta T_s^\gamma$7 and $Y = R^\alpha D^\beta V_e^\delta T_s^\gamma$8 as energetic proxies and uses genetic algorithms and NSGA-II to optimize single-objective and bi-objective formulations. The result is a Pareto-front analysis of trade-offs between interior and surface scores, rather than a single fixed optimum (Krishna et al., 2020).

A separate applied literature uses a quasi Cobb-Douglas model for cloud data centers. There the relevant formulation is

$Y = R^\alpha D^\beta V_e^\delta T_s^\gamma$9

where $Y = A K^\alpha L^\beta$0 is recurring cost and $Y = A K^\alpha L^\beta$1 is infrastructure cost (Sarkar et al., 2016). Energy enters through recurring operational cost, including power and cooling. The paper also introduces a stochastic frontier form,

$Y = A K^\alpha L^\beta$2

with $Y = A K^\alpha L^\beta$3 as random noise and $Y = A K^\alpha L^\beta$4 as technical inefficiency (Sarkar et al., 2016). In this setting, the “energy-based” component is cost-oriented rather than thermodynamic or informational: energy consumption is modeled as a key input affecting revenue, cost, profit, and efficiency.

6. Interpretation, scope, and controversies

The main interpretive issue surrounding EBDC is the ontological status of the Cobb-Douglas form. One line of work treats it as an analytically useful production function that can be generalized to include energy explicitly and bounded by carrying capacity (Power et al., 12 Jun 2025). Another treats it as the unique or near-unique low-distortion aggregator that survives standard economic assumptions when the problem is cast in information geometry (0801.0390, 0901.2586). A third treats it as a statistical macro-regularity emerging from maximum entropy, scale invariance, and physical/energy constraints on heterogeneous micro-technologies (Liuh, 3 Dec 2025).

These perspectives are not identical. The maximum-entropy reconstruction explicitly criticizes the conventional interpretation of Cobb-Douglas as a technical law, arguing instead that its empirical success reflects “high-entropy statistical equilibrium,” constant factor shares, and lossy compression of heterogeneity (Liuh, 3 Dec 2025). By contrast, the Solow-Swan extension adopts Cobb