A Parametric, Second-Order Cone Representable Model of Fairness for Decision-Making Problems

Deepjyoti Deka; Kaarthik Sundar; Russell Bent

REVIEW 2 major objections 4 minor 1 cited by

A Parametric, Second-Order Cone Representable Model of Fairness for Decision-Making Problems

T0 review · 2 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Fairness in any optimization problem can be imposed by a single second-order cone constraint that is exactly equivalent to a lower bound on the Jain et al. fairness index.

desk verdict The paper supplies a clean, correct SOC form of a Jain-index threshold, but the stated novelty is a bit inflated and Proposition 3 glosses over the inequality/equality distinction. read the letter →

arxiv 2412.05143 v1 pith:K2GPWSMP submitted 2024-12-06 math.OC cs.CY

Kaarthik Sundar , Deepjyoti Deka , Russell Bent This is my paper · ORCID

classification math.OCcs.CY MSC 90C2590C90

keywords fairnesssecond-orderconeJainindexnormequivalenceefficiency-fairnesstrade-offpowersystemsconvexoptimization

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes a parametric notion of fairness, called $\varepsilon$-fairness, that can be added to any decision-making optimization problem as a single second-order cone constraint. For a nonnegative utility vector $u \in \mathbb{R}^n_{\ge 0}$, the constraint is $(1-\varepsilon+\varepsilon\sqrt{n})\|u\|_2 \le \|u\|_1$, where $\varepsilon \in [0,1]$ interpolates between the most unfair distribution ($\varepsilon=0$, all utility on one agent) and perfect equality ($\varepsilon=1$, all components equal). The paper proves that this constraint is equivalent to requiring the Jain et al. index to be at least $(1-\varepsilon+\varepsilon\sqrt{n})^2/n$, giving a closed-form translation between a tunable parameter and an established fairness metric. It also proves monotonicity of feasibility and optimal value in $\varepsilon$, so the efficiency–fairness trade-off can be quantified and the right $\varepsilon$ found by bisection. A power-system case study on minimum load shedding shows the constraint in action, with at most a few percent loss in efficiency even at high fairness levels.

What carries the argument

The load-bearing object is the $\varepsilon$-fairness constraint $g(u,\varepsilon) \le 0$, defined for $u \in \mathbb{R}^n_{\ge 0}$ by $(1-\varepsilon+\varepsilon\sqrt{n})\|u\|_2 \le \|u\|_1$. This is a single second-order cone constraint, and it is built directly on the norm-equivalence inequality $\|u\|_2 \le \|u\|_1 \le \sqrt{n}\|u\|_2$, whose two endpoints are the most unfair vector (one nonzero component) and the perfectly fair vector (all components equal). In statistical terms, the constraint is equivalent to $c_v^2 \le h(\varepsilon)$, where $c_v$ is the coefficient of variation and $h(\varepsilon)$ decreases strictly from $\sqrt{n}$ to $0$ as $\varepsilon$ goes from $0$ to $1$, so the mechanism by which fairness is imposed is a cap on the relative dispersion of utilities. The bridge to the Jain index, $JI(u) \ge (1-\varepsilon+\varepsilon\sqrt{n})^2/n$, is what turns $\varepsilon$ into an interpretable, closed-form fairness parameter.

What would settle it

For fixed $n$ and $\varepsilon$, find a nonnegative vector $u$ such that $(1-\varepsilon+\varepsilon\sqrt{n})\|u\|_2 \le \|u\|_1$ but $\frac{1}{n}\left(\sum_i u_i\right)^2 / \sum_i u_i^2 < (1-\varepsilon+\varepsilon\sqrt{n})^2/n$; Proposition 3 asserts no such vector exists.

Watch

Extended reading notes

Core claim

The central claim is that fairness in the distribution of utilities is fully captured by the position of $\|u\|_1$ between $\|u\|_2$ and $\sqrt{n}\|u\|_2$, and that requiring $\|u\|_1$ to lie at or above the convex combination $(1-\varepsilon+\varepsilon\sqrt{n})\|u\|_2$ is a single second-order cone constraint. Because second-order cone constraints are convex, adding the $\varepsilon$-fairness constraint to a convex optimization problem preserves convexity and keeps the problem solvable by the same algorithms. The paper's key theoretical result, Proposition 3, states that enforcing this constraint is exactly equivalent to enforcing $JI(u) \ge (1-\varepsilon+\varepsilon\sqrt{n})^2/n$, where $JI(u)$ is the Jain et al. index; no other parametric fairness model in the literature is known to have such a closed-form relationship to a fairness index. The same inequality is shown to be equivalent to an upper bound on the coefficient of variation of the utilities, which makes precise the sense in which $\varepsilon$ controls the dispersion of the utility distribution. The paper also establishes that the feasible region shrinks and the optimal objective decreases monotonically as $\varepsilon$ increases.

Load-bearing premise

The construction assumes that all utilities are nonnegative and that fairness is completely characterized by the ratio of the 1-norm to the 2-norm, which is exactly the Jain et al. index.

Editorial extensions

If this is right

Any optimization problem whose objective depends on a nonnegative utility vector can be made fair by appending one convex constraint; if the original problem is convex, the fair version remains convex and solvable with the same algorithms.
A decision-maker can set $\varepsilon$ directly from a desired Jain index value by inverting the closed-form formula $w(\varepsilon)=(1-\varepsilon+\varepsilon\sqrt{n})^2/n$.
The feasibility domain of the fair problem is an interval $[0,\varepsilon_{\max}]$, and $\varepsilon_{\max}$ can be located by bisection; the paper argues no such systematic procedure exists for the parameters $\alpha$ or $p$ in existing fairness models.
The monotone decrease of $z(\varepsilon)$ yields a rigorous efficiency–fairness trade-off curve; in the minimum load-shedding case study, the maximum efficiency loss at $\varepsilon=0.9$ is about 4%.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

One immediate extension the paper does not spell out is that the cone constraint could serve as a generic dispersion constraint in any allocation problem, since it is equivalent to bounding the coefficient of variation; this would apply to facility location, portfolio selection, or communication-network rate allocation.
Because the closed-form bridge to the Jain index is built only on the norm ratio, the same construction might be adapted to other indices that are monotone functions of $\|u\|_1/\|u\|_2$, although the paper proves the equivalence only for the Jain index.
The nonnegativity assumption is restrictive for settings with signed utilities such as profits or net benefits; extending $\varepsilon$-fairness to vectors with mixed signs would require a different norm-based construction, and the paper leaves this open.
The monotonicity results suggest that $\varepsilon$-fairness could be used as a regularizer in multi-objective or bi-level formulations, where the trade-off curve itself is the object of interest; this is a testable algorithmic direction beyond the case study.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Core thought: the ε-fairness constraint is a nice observation—enforcing (1−ε+ε√n)||u||2 ≤ ||u||1 is a single SOC constraint and, for nonnegative utilities, equivalent to JI(u) ≥ w(ε). The math is correct, the proof is simple, and the monotonicity results (feasible set shrinks, z(ε) decreases) follow directly. This is genuinely useful for practitioners who want a fairness knob that maps to a Jain-index threshold without solving a parametric family of problems and computing indices a posteriori. The bisection on ε is a sensible add-on.

What is actually new is the parameterization and the explicit closed-form bridge to Jain's index. The norm-equivalence derivation gives a clean geometric story, and the paper correctly credits the classical inequality. It also does honest empirical work on the power-system case, showing the trade-off curve and the infeasibility counts. That part is fine as an illustration, not as a benchmark.

Soft spots, in order of importance. First, the novelty claim in the intro is too strong. The statement that no direct or indirect relationship exists between α/p-norm utilities and fairness indices is debatable—p-norm fairness with total utility fixed is essentially the same ratio, and the connection is not hard to make. The stress-test note caught this; the reader did too. It is a positioning issue, not a correctness gap. Second, Proposition 3's statement "Enforcing u to be ε-fair is equivalent to setting JI(u)=w(ε)" conflates the equality version (Definition 1) with the inequality version used everywhere else. The proof then says "at least ε-fair" is equivalent to JI(u) ≥ w(ε), which is the right claim. It's a minor presentational flaw, but it could confuse readers. Third, the case study has no reproducibility artifacts; the data and code are not provided. That is a limitation, not a fatal one, but it does weaken the empirical validation.

For the record, I agree with the stress-test: the central argument holds up. The nonnegativity and zero-vector caveats are explicit, so they are limitations, not hidden flaws.

Who is this for? Someone who wants to add a fairness constraint to an existing convex (or even nonconvex) optimization model and wants to control fairness via a single parameter that maps to Jain's index. It is not a deep theoretical advance; it is a well-packaged, practical tool.

Recommendation: send it to peer review. A serious referee will ask for the novelty claim to be toned down, Proposition 3 to be reworded, and ideally a reproducibility statement, but the core is sound and the contribution is real. This is a conditional accept, not a desk reject.

Referee Report

2 major / 4 minor

Summary. The paper proposes a parametric fairness model called ε-fairness. For a nonnegative utility vector u ∈ R^n_≥0 and ε ∈ [0,1], u is declared 'at least ε-fair' when (1−ε+ε√n)‖u‖2 ≤ ‖u‖1; because this is a second-order cone inequality, it can be appended to any optimization problem in which u depends affinely on the decision variables. The authors show that this constraint is equivalent to a lower bound on the Jain fairness index, JI(u) ≥ (1−ε+ε√n)^2/n, prove monotonicity of the feasible sets and of the optimal value z(ε), and illustrate the model on a minimum-load-shedding problem for a damaged IEEE 14-bus network. The central mathematical derivation is elementary and internally consistent, with the caveat that Proposition 3 must be stated for the inequality form of ε-fairness.

Significance. If corrected, the model would provide a practical and attractive device: a single convex constraint with a closed-form translation between ε and the Jain index, enabling principled parameter selection and a bisection search for attainable fairness levels. The derivations are transparent and require no fitted parameters; the monotonicity results are simple and directly checkable, and the case study gives concrete empirical support. The main limitations are that the fairness notion is tied to the Jain index, the u=0 case needs an explicit caveat, and the claim that no existing parametric utility family admits any relationship to fairness indices is overstated. None of these issues undermines the core SOC representability result, but the presentation of Proposition 3 and the novelty claim need revision.

major comments (2)

[Section 3.1, Proposition 3] Proposition 3 states that enforcing u to be ε-fair is equivalent to setting JI(u) = (1−ε+ε√n)^2/n, but the constraint used throughout the paper, in equations (6), (8), (9), and (16), is the inequality form (1−ε+ε√n)‖u‖2 ≤ ‖u‖1. For any nonzero u ∈ R^n_≥0, this inequality is equivalent to JI(u) ≥ (1−ε+ε√n)^2/n, with equality only when the norm inequality is tight. The proposition and the proof should be rewritten in terms of 'at least ε-fair' and JI(u) ≥ w(ε); otherwise the central statement misstates the relationship between the proposed model and the Jain index.
[Section 1, after Eq. (4)] The sentence 'no direct or indirect relationship exists between either of the utility functions in (3) or (4) and any of the fairness indices used in the literature' is unsupported and, as stated, false: when the total utility is fixed, maximizing Up(u) for p=2 is equivalent to minimizing ‖u‖2, which is equivalent to maximizing the Jain index JI(u). The cited survey cannot establish an absence result. The claim should be weakened to a statement about closed-form parameter-to-index mappings that are known to the authors, or should be formulated with a precise and provable notion of 'no relationship'.

minor comments (4)

[Definition 1 and Section 3.1] The u=0 case is not handled: the inequality (8) holds trivially at u=0, while JI(u) is undefined because the denominator in (5) vanishes; the equivalence in Proposition 3 should explicitly exclude u=0 or define a convention for it.
[Section 4.3 and Table 1] The text states that (16) was solved for each of the 9765 scenarios at ε=0.9, but Table 1 reports 1715 infeasible scenarios at that value; the treatment of infeasible scenarios in Figures 2 and 5 should be clarified.
[Section 4.1] The notation Xdc in (16) is introduced as XDC earlier in the same section; please unify the notation.
[Throughout] There are several typographical and spacing errors, including 'withn agents' and 'apriori', and the mathematical prose would benefit from a careful proofreading pass.

Circularity Check

1 steps flagged · score 1.0 of 10

Only minor definitional circularity: the Jain-index relationship is a restatement of the epsilon-fairness definition; the SOC model and monotonicity results stand on their own.

self definitional [Section 3.1, Proposition 3 (Eq. 14); Definition 1 (Eq. 8)]
"Proposition 3 Enforcing u to be ε-fair is equivalent to setting JI(u) = (1 − ε + ε√n)2/n. Proof Utilizing the non-negativity of the utility values, Jain et al. index in (5) can be equivalently rewritten as JI(u) = ||u||1^2/(n · ||u||2^2). Combining the above equation with the definition of enforcing ε-fairness of u leads to JI(u) = (1 − ε + ε√n)2/n ≜ w(ε)."

Definition 1 defines 'at least ε-fair' as the inequality c(ε)||u||2 ≤ ||u||1 with c(ε) = 1 − ε + ε√n. Jain's index in Eq. (5) is, for u ≥ 0, exactly (||u||1/||u||2)^2/n. Squaring and rearranging the defining inequality therefore yields JI(u) ≥ c(ε)^2/n = w(ε), so Proposition 3 merely rewrites the definition in Jain-index notation rather than deriving an independent connection. Consequently, the case-study 'validation' that feasible solutions have Jain values at least w(ε) is enforced by the constraint itself, not independently observed. The SOC representability and monotonicity results remain substantive, so this is a low-severity definitional circularity.

full rationale

The paper's central derivation is self-contained: ε-fairness is defined directly from the norm-equivalence inequality (7), and the SOC constraint g(u,ε) ≤ 0 in (6)/(9) is exactly the inequality (8). Propositions 1 and 2 follow from the monotonicity of c(ε) in the constraint and from the fact that (6) is a maximization problem; no fitted parameter is used and no load-bearing result is imported from a self-citation. The closed-form relationship to Jain's index (Proposition 3) is the one place where the claimed result is, by construction, a restatement: since both ε-fairness and JI(u) are defined through ||u||1/||u||2, the equivalence JI(u) ≥ w(ε) is immediate algebra. The paper is transparent about this, and the SOC representability claim (the main novel contribution) does not depend on the Jain-index identification. The MLS case study uses an external benchmark without fitting; its 'validation' of the theoretical result is partly definitional because feasible solutions must satisfy JI ≥ w(ε). Overall, no load-bearing circularity; the score reflects only the minor self-definitional character of the Jain-index equivalence.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The model introduces no new physical or mathematical entities and no fitted constants. It relies only on standard norm inequalities, the nonnegativity of utilities, and the chosen fairness metric.

assumptions (5)

standard math For any u in R^n_>=0, ||u||_2 <= ||u||_1 <= sqrt(n) * ||u||_2.
This is the classical equivalence of norms applied to the 1- and 2-norms, which the paper uses to define the epsilon-fairness constraint in Section 2.
domain assumption Utilities are nonnegative (u in R^n_>=0).
The definition of epsilon-fairness and the linearity of the 1-norm rely on nonnegative utilities; this is stated after Eq. (1).
domain assumption Jain's index is an appropriate measure of fairness.
The paper adopts the Jain index as the fairness metric without justification of its superiority over other indices, and builds the closed-form relationship on it.
domain assumption The utility vector is a linear function of decisions (u = W(x)).
The problem formulation in Eq. (1)-(2) assumes a linear map from decisions to utilities.
domain assumption The DC power flow model is accurate for the case study.
The case study uses the DC approximation of the power network, which ignores voltage and reactive power; this is a standard modeling assumption.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A Parametric, Second-Order Cone Representable Model of Fairness for Decision-Making Problems." pith.science (2026). https://pith.science/paper/K2GPWSMP

@misc{pith2026241205143,
  author       = {Pith},
  title        = {Pith review of: A Parametric, Second-Order Cone Representable Model of Fairness for Decision-Making Problems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/K2GPWSMP}},
  note         = {Machine review of arXiv:2412.05143}
}

abstract

The article develops a parametric model of fairness called "$\varepsilon$-fairness" that can be represented using a single second-order cone constraint and incorporated into existing decision-making problem formulations without impacting the complexity of solution techniques. We develop the model from the fundamental result of finite-dimensional norm equivalence in linear algebra and show that this model has a closed-form relationship to an existing metric for measuring fairness widely used in the literature. Finally, a simple case study on the optimal operation of a damaged power transmission network illustrates its effectiveness.

Figures

Figures reproduced from arXiv: 2412.05143 by the authors.

**Figure 1.** Plot of JI(u) = w(ε) when n = 10. 3.1 Relationship to Jain et al. index We now show that enforcing ε-fairness of the utilities is equivalent to setting the Jain et al. index of the utilities to a bijective function of ε. This will translate any value of the Jain et al. index sought to an equivalent ε value to enforce ε-fairness. To the best of our knowledge, this cannot be done using any of the existing models of fa… view at source ↗

**Figure 2.** Box plot of load shed obtained by solving the fair version of the MLS problem ( [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗

**Figure 3.** Jain et al. index of the load shed values for varying values of [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: Jain et al. index of the load shed values for varying values of [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

**Figure 5.** Figure 5: Relative loss in efficiency (%) as a function of [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]

**Figure 6.** Figure 6: Relative loss in efficiency (%) as a function of [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Fairness-aware Dynamic Hosting Capacity and the Impacts of Strategic Solar PV Curtailment
eess.SY 2025-04 conditional novelty 6.0 of 10

A tighter convex inner approximation plus fairness-aware allocation indicates that roughly 5% annual solar curtailment can increase a test feeder's solar hosting capacity by about 50% without line or voltage violation...

Reference graph

Works this paper leans on

30 extracted references · 29 canonical work pages · cited by 1 Pith paper

[1]

The power grid library for benchmarking AC optimal power flow algorithms

Babaeinejadsarookolaee, Sogol et al. (2019). “The power grid library for benchmarking AC optimal power flow algorithms”. In:arXiv preprint arXiv:1908.02788(cit. on p. 7)

arXiv 2019
[2]

Equality measures properties for location problems

Barbati, Maria and Carmela Piccolo (2016). “Equality measures properties for location problems”. In:Opti- mization Letters10, pp. 903–920 (cit. on p. 2). Bektaş, Tolga and Adam N Letchford (2020). “Usingℓp-norms for fairness in combinatorial optimisation”. In: Computers & Operations Research120, p. 104975 (cit. on pp. 1, 2)

work page 2016
[3]

Athena Scientific (cit

Bertsekas, Dimitri and Robert Gallager (2021).Data networks. Athena Scientific (cit. on p. 1)

work page 2021
[4]

Julia: A fresh approach to numerical computing

Bezanson, Jeff et al. (2017). “Julia: A fresh approach to numerical computing”. In:SIAM review59.1, pp. 65– 98 (cit. on p. 7)

work page 2017
[5]

Relaxations of AC maximal load delivery for severe contingency analysis

Coffrin, Carleton et al. (2018). “Relaxations of AC maximal load delivery for severe contingency analysis”. In: IEEE Transactions on Power Systems34.2, pp. 1450–1458 (cit. on p. 6)

work page 2018
[6]

V22.1.1: User’s Manual for CPLEX

CPLEX, IBM ILOG (2022). “V22.1.1: User’s Manual for CPLEX”. In:International Business Machines Cor- poration (cit. on p. 7)

work page 2022
[7]

JuMP: A modeling language for mathematical opti- mization

Dunning, Iain, Joey Huchette, and Miles Lubin (2017). “JuMP: A modeling language for mathematical opti- mization”. In:SIAM review59.2, pp. 295–320 (cit. on p. 7)

work page 2017
[8]

The Cambridge dictionary of statistics

Everitt, Brian S and Anders Skrondal (2010). “The Cambridge dictionary of statistics”. In: (cit. on p. 4)

work page 2010

Show all 30 references

[9]

On the measure of concentration with special reference to income and statistics

Gini, Corrado (1936). “On the measure of concentration with special reference to income and statistics”. In: Colorado College Publication, General Series208.1, pp. 73–79 (cit. on p. 3)

1936
[10]

Throughput maximization with short-term and long- term Jain’s index constraints in downlink OFDMA systems

Guo, Chongtao, Min Sheng, Xijun Wang, et al. (2014). “Throughput maximization with short-term and long- term Jain’s index constraints in downlink OFDMA systems”. In:IEEE transactions on communications 62.5, pp. 1503–1517 (cit. on p. 2)

2014
[11]

A Jain’s index perspective onα-fairness resource allocation over slow fading channels

Guo, Chongtao, Min Sheng, Yan Zhang, et al. (2013). “A Jain’s index perspective onα-fairness resource allocation over slow fading channels”. In:IEEE communications letters17.4, pp. 705–708 (cit. on p. 2)

2013
[12]

Can locational disparity of prosumer energy optimization due to inverter rules be limited?

Hashmi, Md Umar et al. (2022). “Can locational disparity of prosumer energy optimization due to inverter rules be limited?” In:IEEE Transactions on Power Systems(cit. on p. 6)

2022
[13]

Cambridge university press (cit

Horn, Roger A and Charles R Johnson (2012).Matrix analysis. Cambridge university press (cit. on p. 3)

2012
[14]

Inequalities across cooling and heating in households: Energy equity gaps

Huang, Luling et al. (2023). “Inequalities across cooling and heating in households: Energy equity gaps”. In: Energy Policy182, p. 113748 (cit. on p. 6). SOC Representable Model of Fairness 11

2023
[15]

A quantitative measure of fairness and discrimination

Jain, Rajendra K, Dah-Ming W Chiu, William R Hawe, et al. (1984). “A quantitative measure of fairness and discrimination”. In: Eastern Research Laboratory, Digital Equipment Corporation, Hudson, MA21 (cit. on pp. 3, 5)

1984
[16]

Charging and rate control for elastic traffic

Kelly, Frank (1997). “Charging and rate control for elastic traffic”. In:European transactions on Telecommu- nications 8.1, pp. 33–37 (cit. on p. 2)

1997
[17]

Income distribution and mortality: cross sectional ecological study of the Robin Hood index in the United States

Kennedy, Bruce P, Ichiro Kawachi, and Deborah Prothrow-Stith (1996). “Income distribution and mortality: cross sectional ecological study of the Robin Hood index in the United States”. In:British Medical Journal 312.7037, pp. 1004–1007 (cit. on p. 3)

1996
[18]

Performance Enhancement and Scheduling in Communication Net- works—AReviewintoVariousApproaches

Kumari, Priya and Nitin Jain (2023). “Performance Enhancement and Scheduling in Communication Net- works—AReviewintoVariousApproaches”.In: International Conference on Micro-Electronics and Telecom- munication Engineering. Springer, pp. 661–672 (cit. on p. 2)

2023
[19]

Cooperative search for fair nurse rosters

Martin, Simon et al. (2013). “Cooperative search for fair nurse rosters”. In:Expert Systems with Applications 40.16, pp. 6674–6683 (cit. on p. 2)

2013
[20]

Fairness analysis of indoor multi-user communications through steerable IR-beam

Masoumi, Hamed et al. (2023). “Fairness analysis of indoor multi-user communications through steerable IR-beam”. In:IET Optoelectronics17.2-3, pp. 77–86 (cit. on p. 2)

2023
[21]

Algorithms for Energy Justice

Mathieu, Johanna L (2023). “Algorithms for Energy Justice”. In:Women in Power: Research and Development Advances in Electric Power Systems. Springer, pp. 67–83 (cit. on p. 5)

2023
[22]

Workload equity in vehicle routing problems: A survey and analysis

Matl, Piotr, Richard F Hartl, and Thibaut Vidal (2018). “Workload equity in vehicle routing problems: A survey and analysis”. In:Transportation Science52.2, pp. 239–260 (cit. on p. 2)

2018
[23]

Fair end-to-end window-based congestion control

Mo, Jeonghoon and Jean Walrand (2000). “Fair end-to-end window-based congestion control”. In:IEEE/ACM Transactions on networking8.5, pp. 556–567 (cit. on p. 2). Nash Jr, John F (1950). “The bargaining problem”. In:Econometrica: Journal of the econometric society, pp. 155–162 (...

2000
[24]

A theory of justice

Rawls, John (1971). “A theory of justice”. In:Cambridge (Mass.)(cit. on p. 2)

1971
[25]

Throughput maximizing multiuser schedul- ing with adjustable fairness

Schwarz, Stefan, Christian Mehlfuhrer, and Markus Rupp (2011). “Throughput maximizing multiuser schedul- ing with adjustable fairness”. In:2011 IEEE International Conference on Communications (ICC). IEEE, pp. 1–5 (cit. on p. 2)

2011
[26]

Network optimization and control

Shakkottai, Srinivas, Rayadurgam Srikant, et al. (2008). “Network optimization and control”. In:Foundations and Trends® in Networking 2.3, pp. 271–379 (cit. on pp. 1, 2)

2008
[27]

Multi-portfolio optimization and fairness in allocation of trades

Stubbs, Robert A and Dieter Vandenbussche (2009). “Multi-portfolio optimization and fairness in allocation of trades”. In:White paper, Axioma Inc. Research Paper13 (cit. on p. 1)

2009
[28]

Managing Wildfire Risk and Promoting Equity through Optimal Configuration of Networked Microgrids

Taylor, Sofia et al. (2023). “Managing Wildfire Risk and Promoting Equity through Optimal Configuration of Networked Microgrids”. In:Proceedings of the 14th ACM International Conference on Future Energy Systems, pp. 189–199 (cit. on p. 5)

2023
[29]

Local and utility-wide cost allocations for a more equitable wildfire-resilient distribution grid

Wang, Zhecheng et al. (2023). “Local and utility-wide cost allocations for a more equitable wildfire-resilient distribution grid”. In:Nature Energy, pp. 1–12 (cit. on p. 6). Xinying Chen, Violet and JN Hooker (2023). “A guide to formulating fairness in an optimization model”. ...

2023
[30]

Alpha-Fair Routing in Urban Air Mobility with Risk-Aware Constraints

Yu, Yue et al. (2023). “Alpha-Fair Routing in Urban Air Mobility with Risk-Aware Constraints”. In:arXiv preprint arXiv:2310.00135 (cit. on p. 2)

2023 arXiv

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [1]

The power grid library for benchmarking AC optimal power flow algorithms

Babaeinejadsarookolaee, Sogol et al. (2019). “The power grid library for benchmarking AC optimal power flow algorithms”. In:arXiv preprint arXiv:1908.02788(cit. on p. 7)

arXiv 2019

[2] [2]

Equality measures properties for location problems

Barbati, Maria and Carmela Piccolo (2016). “Equality measures properties for location problems”. In:Opti- mization Letters10, pp. 903–920 (cit. on p. 2). Bektaş, Tolga and Adam N Letchford (2020). “Usingℓp-norms for fairness in combinatorial optimisation”. In: Computers & Operations Research120, p. 104975 (cit. on pp. 1, 2)

work page 2016

[3] [3]

Athena Scientific (cit

Bertsekas, Dimitri and Robert Gallager (2021).Data networks. Athena Scientific (cit. on p. 1)

work page 2021

[4] [4]

Julia: A fresh approach to numerical computing

Bezanson, Jeff et al. (2017). “Julia: A fresh approach to numerical computing”. In:SIAM review59.1, pp. 65– 98 (cit. on p. 7)

work page 2017

[5] [5]

Relaxations of AC maximal load delivery for severe contingency analysis

Coffrin, Carleton et al. (2018). “Relaxations of AC maximal load delivery for severe contingency analysis”. In: IEEE Transactions on Power Systems34.2, pp. 1450–1458 (cit. on p. 6)

work page 2018

[6] [6]

V22.1.1: User’s Manual for CPLEX

CPLEX, IBM ILOG (2022). “V22.1.1: User’s Manual for CPLEX”. In:International Business Machines Cor- poration (cit. on p. 7)

work page 2022

[7] [7]

JuMP: A modeling language for mathematical opti- mization

Dunning, Iain, Joey Huchette, and Miles Lubin (2017). “JuMP: A modeling language for mathematical opti- mization”. In:SIAM review59.2, pp. 295–320 (cit. on p. 7)

work page 2017

[8] [8]

The Cambridge dictionary of statistics

Everitt, Brian S and Anders Skrondal (2010). “The Cambridge dictionary of statistics”. In: (cit. on p. 4)

work page 2010