An alternative approach to well-posedness of McKean-Vlasov equations arising in Consensus-Based Optimization

Alessandro Baldi

REVIEW 1 major objections 3 minor 34 references

An alternative approach to well-posedness of McKean-Vlasov equations arising in Consensus-Based Optimization

T0 review · 1 major / 3 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read A cut-off trick tames the non-Lipschitz mean-field equation for Consensus-Based Optimization, proving strong existence and pathwise uniqueness under a weaker condition.

desk verdict The truncation proof is a clean alternative route to known well-posedness, but the advertised uniqueness extension is vacuous: bounded consensus already forces continuity, so the contribution is technique, not a new theorem. read the letter →

arxiv 2512.19446 v4 pith:4SFYHEXJ submitted 2025-12-22 math.OC math.APmath.PR

Alessandro Baldi This is my paper · ORCID

classification math.OCmath.APmath.PR MSC 60H1060K3593A1635Q9365C35

keywords Consensus-BasedOptimizationMcKean–Vlasovequationswell-posednesspathwiseuniquenesstruncationargumentSznitmanfixedpointmean-fieldlimitnon-Lipschitzcoefficients

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

This paper proves that the mean-field equation describing Consensus-Based Optimization (CBO), a particle method for global optimization, has a unique strong solution under mild assumptions. Because the equation's coefficients are not globally Lipschitz, the classical Sznitman existence theory does not apply directly; the author introduces a truncation function that cuts off the consensus point when the p-th moment of the law grows too large. The truncated equation fits into the standard Lipschitz framework, and a uniform moment bound shows that the cut-off is never active on a large enough radius. The central result relaxes the pathwise-uniqueness assumption: the map sending time to the consensus point only needs to be bounded, not continuous, on [0,T].

What carries the argument

The key object is the cut-off function φ_R(μ) = η_R(m_p(μ)), defined on the Wasserstein space (P_p(R^d), W_p), where m_p(μ) is the p-th moment and η_R is a smooth bump taking value 1 for arguments ≤ R and 0 for arguments ≥ R+1. Because m_p(μ) equals W_p(μ, δ_0), the cut-off is 1-Lipschitz in W_p. Multiplying the consensus point M_β(μ) by φ_R(μ) in both the drift and diffusion fields tames the non-global Lipschitz behaviour of the original fields, reducing the equation to the globally Lipschitz case where Sznitman's fixed-point theorem yields existence and pathwise uniqueness.

What would settle it

Check whether the consensus map M_β, for some objective f in O(s,ℓ), violates the sublinear growth bound ||M_β(μ)|| ≤ C_M m_p(μ) for large moments, or whether the local Lipschitz constants L_{M,R} grow faster than R; if either fails, the cut-off fields in (4.5) would not satisfy the global Lipschitz estimates needed for Sznitman's theorem.

Watch

Extended reading notes

Core claim

Theorem 1.1 establishes that, for objective functions in the class O(s,ℓ) and for p ≥ 2 ∨ p_M(s,ℓ), given an L^p initial state, there exists a strong solution to the mean-field CBO equation (1.4), and this solution is pathwise unique among strong solutions for which the consensus-point map t ↦ M_β(ρ_t) is bounded over [0,T]. The proof works by defining a cut-off function φ_R on the space of probability measures that equals 1 when the p-th moment is ≤ R and 0 when it is ≥ R+1, then showing that the truncated drift and diffusion fields are globally Lipschitz, so Sznitman's fixed-point argument applies. A Grönwall estimate uniform in R then guarantees that for R large enough the cut-off is neve

Load-bearing premise

The entire proof leans on the imported result (Proposition 3.2) that the consensus map M_β is locally Lipschitz on bounded-moment sets and sublinear in the p-th moment; if these constants failed or scaled unfavourably with R, the truncated fields would not be globally Lipschitz and the cut-off argument would collapse.

Editorial extensions

If this is right

The uniqueness class for the mean-field CBO equation is enlarged: only boundedness of the consensus-point trajectory is needed, not continuity, which is a weaker and more practical condition on solutions.
The truncation technique provides a self-contained route to well-posedness that avoids the Leray–Schauder fixed-point argument previously used for this equation.
The same cut-off construction could be applied to other mean-field equations whose coefficients are only locally Lipschitz in the measure argument, as long as a sublinearity bound like ||M_β(μ)|| ≤ C_M m_p(μ) holds.
The proof gives a quantitative Grönwall-type control on the growth of p-th moments that is uniform in the truncation radius, which may be of independent use in propagation-of-chaos or stability analysis.

Reading between the lines

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

The cut-off argument suggests a generic template for mean-field equations with non-globally-Lipschitz coefficients: localize via a moment-based bump, solve in the Lipschitz regime, then use a uniform moment estimate to remove the cut-off; this could be tested on other CBO variants with state constraints or jump diffusions.
The same technique may yield a simpler proof of existence of weak solutions via a martingale-problem formulation, since the truncated equations are Lipschitz and the uniform bound controls tightness.
A testable extension would be to check whether the boundedness condition on t ↦ M_β(ρ_t) can be replaced by an explicit a priori bound depending only on the initial moment; this would make uniqueness statement more directly applicable in mean-field limit arguments.
The paper's reliance on the imported Lipschitz/sublinearity estimates of the consensus map M_β suggests a simpler standalone proof of these estimates might exist for the specific class O(s,ℓ), which would make the approach fully self-contained.

Share X Bluesky LinkedIn Reddit HN

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

The paper is a solid, readable alternative proof of known well-posedness for the mean-field CBO equation. The truncation on Wasserstein space is the real novelty: it restores global Lipschitz continuity, and the uniform moment bound that removes the cut-off is done carefully. The Sznitman framework is spelled out in an appendix, which is useful for students. The author is honest that the existence content of Theorem 1.1 is well-known.

The advertised extension of pathwise uniqueness is not real. Your stress-test note is correct. For any strong solution with bounded t↦Mβ(ρt), Proposition 5.1 gives E supt∥Xt∥p < ∞. Since X has continuous paths, dominated convergence gives Wp(ρt, ρs) ≤ (E∥Xt−Xs∥p)1/p → 0, so t↦ρt is continuous in Wp. The moment bound puts all ρt in a common Pp,R, so the imported local Lipschitz estimate (Prop 3.2) yields continuity of Mβ(ρt). Thus the bounded class coincides with the continuous class of [20]; uniqueness is not extended. The abstract and §2.3 overstate the result. This is a genuine soft spot, not a nitpick, because the only new theorem beyond the truncation method is the one that evaporates.

The existence proof itself checks out. Proposition 4.2 is clean, and the Grönwall argument in Prop 4.4 is standard. The dependency on [20] for the key local Lipschitz estimate is explicitly flagged; that is fine, though it means the proof is not self-contained. No circularity or invented entities.

Verdict: the paper deserves a serious referee, but the referee should ask the author to withdraw or heavily qualify the uniqueness claim. As it stands, the contribution is a valid alternative proof of a known result. That can still be worth publishing—the cut-off idea may be reused elsewhere—but the framing should match the actual advance. I would not cite it for a new result; I might point to it as a method if I needed the truncation trick. It is coherent, serious work, with one overclaim that needs fixing.

Referee Report

1 major / 3 minor

Summary. The paper proposes an alternative proof of well-posedness for the mean-field Consensus-Based Optimization equation (1.4), where the drift and diffusion fields lack global Lipschitz continuity. The method introduces a measure-dependent cut-off function φ_R, proves that the truncated fields (4.5) are globally Lipschitz (Prop. 4.2), obtains a strong solution to the truncated problem via Sznitman's fixed-point framework (Prop. 4.3), and establishes a uniform-in-R moment bound (Prop. 4.4) that allows removing the truncation. Existence of a strong solution to (1.4) follows. For uniqueness, the paper combines an a priori L^p bound (Prop. 5.1) with the truncated-problem uniqueness to obtain pathwise uniqueness in the class of strong solutions whose consensus map t ↦ Mβ(ρ_t) is bounded (Cor. 5.2, Thm. 1.1). The paper explicitly states that its main goal is to illustrate an alternative technique and that the core well-posedness result is known from [10,20].

Significance. The truncation-on-measure-space argument is a clean and potentially reusable technique: Proposition 4.2 is carefully proven, the uniform moment bound of Proposition 4.4 is elegant, and the appendix gives a self-contained account of Sznitman's argument. The reliance on the external local-Lipschitz estimate (Prop. 3.2 from [20]) is transparent and not circular. However, the advertised extension of the uniqueness class from continuous consensus maps to merely bounded consensus maps is vacuous: under the paper's own estimates, bounded consensus implies continuous consensus. Thus the genuinely new content is the alternative existence proof, which is valid, but the claimed improvement over [20, Thm. 2.3] is overstated and must be corrected.

major comments (1)

[Abstract; §2.3; Theorem 1.1; Corollary 5.2] The claimed extension of pathwise uniqueness from the class of solutions with continuous consensus map to those with merely bounded consensus map is not real. Indeed, take any strong solution X of (1.4) with t ↦ Mβ(ρ_t) bounded. Proposition 5.1 gives E[sup_t ||X_t||^p] < ∞. Since X is a.s. continuous and sup_t ||X_t||^p is integrable, dominated convergence gives W_p(ρ_t,ρ_s)^p ≤ E||X_t − X_s||^p → 0 as s → t. Let R := sup_t m_p(ρ_t) < ∞. The local Lipschitz estimate (3.3) of Prop. 3.2 then yields ||Mβ(ρ_t) − Mβ(ρ_s)|| ≤ L_{M,R} W_p(ρ_t,ρ_s) → 0, so t ↦ Mβ(ρ_t) is continuous. Hence the 'bounded' class coincides with the 'continuous' class already treated in [20, Thm. 2.3]. The uniqueness statement in Theorem 1.1 is correct but is not an extension. The abstract, §2.3, and the remark after Theorem 1.1 should withdraw the claim of an extended uniqueness class or explicitly qualify that the c

minor comments (3)

[Proposition 4.4] In the proof of Proposition 4.4, the derivative d/dt m_p^p(ρ_t^R) is formally computed via Itô's formula. It would be helpful to briefly justify that t ↦ m_p^p(ρ_t^R) is absolutely continuous under the integrability condition (4.8), e.g., by writing Itô's formula and noting that the stochastic integral has zero mean and the remaining terms are integrable.
[Appendix A] In the proof of Theorem 2.1, the map S^n is sometimes denoted 'S n'; please use consistent notation for composition. Also, 'sparable' should be 'separable'.
[Section 5] Proposition 5.1 invokes [4, Theorem 9.1] for L^p estimates. This is standard, but it would improve clarity to state explicitly that the constant C_ρ may depend on M_ρ = sup_t ||Mβ(ρ_t)||, which is finite by hypothesis.

Circularity Check

1 steps flagged · score 2.0 of 10

Existence proof via truncation is not circular, but the advertised uniqueness-class extension is vacuous: bounded consensus already implies continuity.

renaming known result [Abstract; §2.3, first paragraph; Theorem 1.1; Corollary 5.2 (with Props. 5.1 and 3.2)]
"we prove that pathwise uniqueness holds in the class of strong solutions for which the map t 7→ Mβ(ρt) is bounded over [0, T], thus relaxing the continuity requirement on the aforementioned map (see, e.g., [20, Theorem 2.3] for comparison)."

By Prop. 5.1, any strong solution with bounded consensus satisfies E sup_t ||X_t||^p < ∞. Since X is a.s. continuous, dominated convergence gives W_p(ρ_t,ρ_s)→0; the imported local Lipschitz bound (3.3) then forces t↦Mβ(ρ_t) to be continuous. Conversely, a continuous map on [0,T] is bounded. Hence the 'bounded' uniqueness class coincides with the 'continuous' class of [20, Thm 2.3]. The claimed relaxation is therefore equivalent by construction to the known result, not an extension.

full rationale

The central derivation is self-contained: Prop. 4.2 proves global Lipschitzness of the truncated fields using the externally proved estimates (3.3)-(3.4) from [20]; Prop. 4.3 applies Sznitman's theorem; Prop. 4.4 gives an R-uniform moment bound, and R is chosen after that bound, not fitted to force the conclusion. The only self-citation [1] is motivational and not load-bearing. The imported Prop. 3.2 is independent prior work, not a self-citation. The one circularity-adjacent issue is the uniqueness-class claim: the paper presents 'bounded consensus' as a relaxation of 'continuous consensus', but its own Prop. 5.1 plus Prop. 3.2 imply bounded consensus ⇒ continuous consensus, so the class is the same and the extension is vacuous. This is a real overclaim, but it does not infect the existence argument; the truncation-based proof remains an independent alternative technique.

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

The paper introduces no new physical or mathematical entities and fits no parameters to data. The cut-off radius R and the smooth function η_R are proof devices chosen after the estimates; they do not enter the final theorem. The load-bearing input from outside the paper is Proposition 3.2 of [20] (local Lipschitz and sublinearity of Mβ), plus the standard Sznitman well-posedness framework.

assumptions (5)

domain assumption f ∈ O(s,ℓ) (Definition 3.1): f satisfies the continuity bound (3.1a) and the polynomial growth bounds (3.1b).
Standing assumption on the objective function; standard in the CBO literature and used throughout.
domain assumption The noise coefficient S is globally Lipschitz and S(0)=0 (Remark 3.3).
Guarantees linear growth and Lipschitz continuity of the diffusion field.
domain assumption Proposition 3.2 of [20]: for f∈O(s,ℓ), p≥p_M(s,ℓ), the consensus map Mβ satisfies (3.3) and (3.4).
Load-bearing external estimate: local Lipschitz continuity on P_{p,R} and sublinear growth in m_p. Without it the truncation argument fails.
standard math Sznitman's well-posedness theorem (Theorem 2.1) for McKean–Vlasov SDEs with globally Lipschitz fields.
Classical framework; the paper proves it in Appendix A using BDG inequality and a contraction argument.
standard math Itô's formula and standard SDE moment estimates (e.g., [4, Theorem 9.1]) for SDEs with Lipschitz-in-x coefficients.
Used in the moment estimates of Propositions 4.4 and 5.1.

how reviews work

0 comments

Cite this review

Pith. "Pith review of An alternative approach to well-posedness of McKean-Vlasov equations arising in Consensus-Based Optimization." pith.science (2026). https://pith.science/paper/4SFYHEXJ

@misc{pith2026251219446,
  author       = {Pith},
  title        = {Pith review of: An alternative approach to well-posedness of McKean-Vlasov equations arising in Consensus-Based Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4SFYHEXJ}},
  note         = {Machine review of arXiv:2512.19446}
}

read the original abstract

In this work we study the mean-field description of Consensus-Based Optimization (CBO), a derivative-free particle optimization method. Such a description is provided by a non-local SDE of McKean-Vlasov type, whose fields lack of global Lipschitz continuity. We propose a novel approach to prove the well-posedness of the mean-field CBO equation based on a truncation argument. The latter is performed through the introduction of a cut-off function, defined on the space of probability measures, acting on the fields. This procedure allows us to study the well-posedness problem in the classical framework of Sznitman. Through this argument, we recover the established result on the existence of strong solutions, and we extend the class of solutions for which pathwise uniqueness holds.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 1 linked inside Pith

[20]

N. J. Gerber, F. Hoffmann, and U. Vaes,Mean-field limits for Consensus-Based Optimization and Sampling, ESAIM Control Optim. Calc. Var., 31 (2025), p. Paper No. 74

2025
[1]

S. Almi, A. Baldi, M. Morandotti, and F. Solombrino,A general perspective on CBO methods with stochastic rate of information, preprint arXiv:2507.20029, 2025

arXiv 2025
[2]

Ambrosio, E

L. Ambrosio, E. Brué, and D. Semola,Lectures on optimal transport, vol. 130 of Unitext, Springer, Cham, 2021

2021
[3]

Ambrosio, N

L. Ambrosio, N. Gigli, and G. Savaré,Gradient flows in metric spaces and in the space of probability measures, Lectures in Mathematics ETH Zürich, Birkhäuser Verlag, Basel, second ed., 2008

2008
[4]

Baldi,Stochastic calculus

P. Baldi,Stochastic calculus. An introduction through theory and exercises, Universitext, Springer, Cham, 2017

2017
[5]

Bellman,Introduction to matrix analysis, vol

R. Bellman,Introduction to matrix analysis, vol. 19 of Classics in Applied Mathematics, Society for Industrial and Applied Mathematics (SIAM), Philadelphia, PA, 1997. Reprint of the second (1970) edition, With a foreword by Gene Golub

1997
[6]

Borghi and M

G. Borghi and M. Herty,Model predictive control strategies using consensus-based optimization, Mathematical Control and Related Fields, 15 (2025), pp. 876–894

2025
[7]

Borghi, M

G. Borghi, M. Herty, and L. Pareschi,An adaptive consensus based method for multi-objective optimization with uniform Pareto front approximation, Appl. Math. Optim., 88 (2023), pp. Paper No. 58, 43

2023

Show all 34 references

[8]

Optim., 33 (2023), pp

,Constrained consensus-based optimization, SIAM J. Optim., 33 (2023), pp. 211–236

2023
[9]

Carmona,Lectures on BSDEs, stochastic control, and stochastic differential games with finan- cial applications, vol

R. Carmona,Lectures on BSDEs, stochastic control, and stochastic differential games with finan- cial applications, vol. 1 of Financial Mathematics, Society for Industrial and Applied Mathematics (SIAM), Philadelphia, PA, 2016

2016
[10]

J. A. Carrillo, Y.-P. Choi, C. Totzeck, and O. Tse,An analytical framework for consensus- based global optimization method, Math. Models Methods Appl. Sci., 28 (2018), pp. 1037–1066. AN ALTERNATIVE APPROACH TO WELL-POSEDNESS OF MEAN-FIELD CBO 16

2018
[11]

J. A. Carrillo, S. Jin, L. Li, and Y. Zhu,A consensus-based global optimization method for high dimensional machine learning problems, ESAIM Control Optim. Calc. Var., 27 (2021), pp. Paper No. S5, 22

2021
[12]

J. A. Carrillo, N. G. Trillos, S. Li, and Y. Zhu,FedCBO: Reaching Group Consensus in Clustered Federated Learning through Consensus-based Optimization, Journal of Machine Learning Research, 25 (2024), pp. 1–51

2024
[13]

Chaintron and A

L.-P. Chaintron and A. Diez,Propagation of chaos: a review of models, methods and applica- tions. I. Models and methods, Kinet. Relat. Models, 15 (2022), pp. 895–1015

2022
[14]

Fornasier, H

M. Fornasier, H. Huang, L. Pareschi, and P. Sünnen,Consensus-based optimization on hypersurfaces: Well-posedness and mean-field limit, Mathematical Models and Methods in Applied Sciences, 30 (2020), pp. 2725–2751

2020
[15]

Fornasier, H

M. Fornasier, H. Huang, L. Pareschi, and P. Sünnen,Anisotropic diffusion in consensus- based optimization on the sphere, SIAM J. Optim., 32 (2022), pp. 1984–2012

2022
[16]

Fornasier, T

M. Fornasier, T. Klock, and K. Riedl,Convergence of anisotropic consensus-based optimiza- tion in mean-field law, in International conference on the applications of evolutionary computation (part of evostar), Springer, 2022, pp. 738–754

2022
[17]

2973–3004

,Consensus-based optimization methods converge globally, SIAM Journal on Optimization, 34 (2024), pp. 2973–3004

2024
[18]

Fornasier, L

M. Fornasier, L. Pareschi, H. Huang, and P. Sünnen,Consensus-based optimization on the sphere: Convergence to global minimizers and machine learning, Journal of Machine Learning Research, 22 (2021), pp. 1–55

2021
[19]

Fornasier and L

M. Fornasier and L. Sun,A PDE framework of consensus-based optimization for objectives with multiple global minimizers, Comm. Partial Differential Equations, 50 (2025), pp. 493–541

2025
[21]

Gilbarg and N

D. Gilbarg and N. S. Trudinger,Elliptic partial differential equations of second order, Classics in Mathematics, Springer-Verlag, Berlin, 2001. Reprint of the 1998 edition

2001
[22]

Huang and J

H. Huang and J. Qiu,On the mean-field limit for the consensus-based optimization, Math. Methods Appl. Sci., 45 (2022), pp. 7814–7831

2022
[23]

Huang, J

H. Huang, J. Qiu, and K. Riedl,Consensus-based optimization for saddle point problems, SIAM J. Control Optim., 62 (2024), pp. 1093–1121

2024
[24]

Huang and J

H. Huang and J. Warnett,Well-posedness and mean-field limit estimate of a consensus-based algorithm for multiplayer games, Communications on Pure and Applied Analysis, (2025)

2025
[25]

Kalise, A

D. Kalise, A. Sharma, and M. V. Tretyakov,Consensus-based optimization via jump-diffusion stochastic differential equations, Mathematical Models and Methods in Applied Sciences, 33 (2023), pp. 289–339

2023
[26]

Klamroth, M

K. Klamroth, M. Stiglmayr, and C. Totzeck,Consensus-based optimization for multi- objective problems: a multi-swarm approach, Journal of Global Optimization, 89 (2024), pp. 745– 776

2024
[27]

H. P. McKean, Jr.,Propagation of chaos for a class of non-linear parabolic equations, in Stochastic Differential Equations (Lecture Series in Differential Equations, Session 7, Catholic Univ., 1967), vol. Session 7 of Lecture Series in Differential Equations, Air Force Office ...

1967
[28]

S. Méléard,Asymptotic behaviour of some interacting particle systems; McKean-Vlasov and Boltzmann models, in Probabilistic models for nonlinear partial differential equations (Montecatini Terme, 1995), vol. 1627 of Lecture Notes in Math., Springer, Berlin, 1996, pp. 42–95

1995
[29]

Pinnau, C

R. Pinnau, C. Totzeck, O. Tse, and S. Martin,A consensus-based model for global optimiza- tion and its mean-field limit, Math. Models Methods Appl. Sci., 27 (2017), pp. 183–204

2017
[30]

Santambrogio,Optimal transport for applied mathematicians, vol

F. Santambrogio,Optimal transport for applied mathematicians, vol. 87 of Progress in Nonlinear Differential Equations and their Applications, Birkhäuser/Springer, Cham, 2015. Calculus of variations, PDEs, and modeling

2015
[31]

D. W. Stroock and S. R. S. Varadhan,Multidimensional diffusion processes, Classics in Mathematics, Springer-Verlag, Berlin, 2006. Reprint of the 1997 edition

2006
[32]

Sznitman,Topics in propagation of chaos, in École d’Été de Probabilités de Saint-Flour XIX—1989, vol

A.-S. Sznitman,Topics in propagation of chaos, in École d’Été de Probabilités de Saint-Flour XIX—1989, vol. 1464 of Lecture Notes in Math., Springer, Berlin, 1991, pp. 165–251

1989
[33]

Totzeck,Trends in consensus-based optimization, in Active particles

C. Totzeck,Trends in consensus-based optimization, in Active particles. Vol. 3. Advances in theory, models, and applications, Model. Simul. Sci. Eng. Technol., Birkhäuser/Springer, Cham, 2022, pp. 201–226. AN ALTERNATIVE APPROACH TO WELL-POSEDNESS OF MEAN-FIELD CBO 17

2022
[34]

G. L. Lagrange

C. Villani,Optimal Transport: Old and New, vol. 338 of Grundlehren der mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences], Springer-Verlag, Berlin, 2009. (Alessandro Baldi)Dipartimento di Scienze Matematiche “G. L. Lagrange”, Politecnico di Torino,...

2009

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.

[1] [20]

N. J. Gerber, F. Hoffmann, and U. Vaes,Mean-field limits for Consensus-Based Optimization and Sampling, ESAIM Control Optim. Calc. Var., 31 (2025), p. Paper No. 74

2025

[2] [1]

S. Almi, A. Baldi, M. Morandotti, and F. Solombrino,A general perspective on CBO methods with stochastic rate of information, preprint arXiv:2507.20029, 2025

arXiv 2025

[3] [2]

Ambrosio, E

L. Ambrosio, E. Brué, and D. Semola,Lectures on optimal transport, vol. 130 of Unitext, Springer, Cham, 2021

2021

[4] [3]

Ambrosio, N

L. Ambrosio, N. Gigli, and G. Savaré,Gradient flows in metric spaces and in the space of probability measures, Lectures in Mathematics ETH Zürich, Birkhäuser Verlag, Basel, second ed., 2008

2008

[5] [4]

Baldi,Stochastic calculus

P. Baldi,Stochastic calculus. An introduction through theory and exercises, Universitext, Springer, Cham, 2017

2017

[6] [5]

Bellman,Introduction to matrix analysis, vol

R. Bellman,Introduction to matrix analysis, vol. 19 of Classics in Applied Mathematics, Society for Industrial and Applied Mathematics (SIAM), Philadelphia, PA, 1997. Reprint of the second (1970) edition, With a foreword by Gene Golub

1997

[7] [6]

Borghi and M

G. Borghi and M. Herty,Model predictive control strategies using consensus-based optimization, Mathematical Control and Related Fields, 15 (2025), pp. 876–894

2025

[8] [7]

Borghi, M

G. Borghi, M. Herty, and L. Pareschi,An adaptive consensus based method for multi-objective optimization with uniform Pareto front approximation, Appl. Math. Optim., 88 (2023), pp. Paper No. 58, 43

2023