Pith. sign in

REVIEW 3 major objections 5 minor 51 references

For a robust leader who fears the worst follower equilibrium, the value is characterized, in the limit of penalization and vanishing noise, by an infinite-dimensional HJB equation on the Wasserstein space.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-01 06:25 UTC pith:NSFXMZ2E

load-bearing objection A genuinely novel robust Stackelberg formulation and a clean convergence theorem, but the headline Wasserstein HJB characterization is explicitly heuristic for the original problem—referee it, but expect major revision. the 3 major comments →

arxiv 2607.21890 v1 pith:NSFXMZ2E submitted 2026-07-24 math.OC

Stackelberg Games with a Robust Leader

Juan Li , Jianfeng Zhang , Zimu Zhu This is my paper
classification math.OC MSC 91A1891A1591B4349L2035Q89
keywords Stackelberg gamesrobust leaderNash equilibriumzero-sum gamesopen loop controlsmean field controlWasserstein spaceprincipal-agent problems
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper studies a Stackelberg game in which one leader faces several followers who may settle on any of multiple Nash equilibria, and the leader is robust: she assumes the followers will pick the equilibrium worst for her. The authors reformulate this as an unconstrained sup-inf problem—a zero-sum game with open-loop controls—by penalizing the gap between equilibrium and best-response conditions, and prove that the penalized values converge to the robust value as the penalty grows. They then show that, under strong boundedness and smoothness assumptions, the value of this open-loop zero-sum game can be characterized by an HJB equation on the Wasserstein space of probability laws, with the state lifted to include the distribution of the auxiliary processes. The same mechanism applies to principal-agent problems with multiple agents, where individual-rationality constraints only alter the terminal condition. The authors explicitly flag that the application back to the original Stackelberg setup is heuristic because the unbounded follower control violates the required regularity.

Core claim

The paper's central discovery is a chain of equivalences and approximations. Starting from a robust Stackelberg game—where the leader evaluates each of her controls by the worst Nash equilibrium the followers might implement—the authors use a characterization of ε-Nash equilibria by a penalty functional I(α0, α, y, β) measuring how far a control profile is from equilibrium. The leader's problem becomes Vλ = sup_{α0} inf_{α,y,β} [J0 + λI], and Theorem 2.4 shows Vλ ↑ V0 as λ→∞. This sup-inf problem is a zero-sum game with open-loop controls, which is time-inconsistent even under the Isaacs condition (Example 7.4). The paper's main positive theorem (Theorem 6.9) states that, for the general mea

What carries the argument

The central mechanism is a three-step reduction. First, the equilibrium condition for the followers is rewritten through a BSDE solution (Y,Z), then converted into a forward SDE with controls y=Y0 and β=Z; the penalty functional I(α0, α, y, β) encodes both the running Hamiltonian mismatch and the terminal misalignment (equations (2.6)-(2.10)). Second, the resulting sup-inf problem is lifted to the Wasserstein space of probability measures by treating the law of the auxiliary state as part of the state; this turns the time-inconsistent mean-variance-type objective into a dynamic value Uλ,κ(t, μ) that satisfies an HJB equation whose Hamiltonian is built from the linear functional derivative of

Load-bearing premise

The whole chain rests on strong boundedness and regularity assumptions for the dynamic system, which the authors admit fail for the original Stackelberg model because the followers' auxiliary controls are unbounded (Section 6.1).

What would settle it

Take an explicit open-loop zero-sum game satisfying Assumptions 6.1-6.7 (e.g., the solvable example in Section 7.3) and compute both V0 and the limit of sup_y U_{λ,κ}(0, δ_{(0,x,y)}) as λ→∞, κ→0; a nonzero gap would falsify Theorem 6.9. Alternatively, exhibit a system satisfying the assumptions for which the FBSDE (6.14) has multiple solutions or the HJB equation (6.23) lacks a classical solution, breaking the chain from (6.16) to (6.23).

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • If the Wasserstein HJB characterization holds, the robust leader's value can be approximated by solving an infinite-dimensional PDE and then taking λ→∞ and κ→0, giving a principled alternative to direct equilibrium search.
  • For principal-agent problems with multiple agents, the same HJB equation applies with the individual-rationality constraint and the lump-sum payment optimization appearing only in the terminal condition, so the generator of the PDE is unchanged.
  • The explicit formulas (4.10), (6.26)-(6.31) provide a candidate maximizer for the principal's contract; the authors conjecture it is approximately optimal even beyond their technical assumptions.
  • Under the Isaacs condition, the open-loop zero-sum game still lacks a saddle point and its value differs from the classical Bellman-Isaacs PDE solution; the Wasserstein HJB equation is the right object instead (Example 7.4).
  • The ε-equilibrium relaxation via the penalty functional I makes the value robust to small perturbations of the followers' controls, avoiding the discontinuity of the exact equilibrium set.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The paper's own 'heuristic' label for the original Stackelberg application indicates that the rigorous theorem is proved one level up—for the general open-loop zero-sum game (6.1)—and that a missing regularity bridge would be needed to apply it to the motivating model; this is an inference about the scope of the result, not a claim the paper makes.
  • A concrete testable extension would be to verify, on the explicit multi-agent example, whether the contract constructed from (6.31) remains approximately optimal when the strong assumptions are relaxed; the paper conjectures yes but does not demonstrate it.
  • The mechanism suggests a numerical route: approximate the Wasserstein HJB equation on a finite-dimensional subspace of probability laws and measure the convergence in λ and κ; the paper does not provide such an experiment.
  • One could also fold model ambiguity into the worst-case operator, since both the equilibrium-selection worst case and the measure-perturbation worst case are inf operations; the Hamiltonian in (6.23) would then contain an additional inf, which may be handled by the same machinery.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper studies a continuous-time Stackelberg game with one leader and multiple followers when the leader is robust/pessimistic: given her control, the followers play a Nash game, and the leader evaluates the worst Nash equilibrium for her. Because true equilibria are fragile, the authors follow Feinstein–Rudloff–Zhang [23] and use approximate ε-equilibria, defining the robust value V0 as the limit of Vε. They then introduce a penalized value Vλ, transform the BSDE characterization of ε-equilibria into an unconstrained sup-inf problem, and argue that this reduces to an open-loop zero-sum game with mean-field terms. The paper's central advertised claim is that this value is characterized by an HJB equation on the Wasserstein space of probability measures. Sections 5 and 6 develop general results for zero-sum games with open-loop controls: a non-McKean–Vlasov special case is handled through BSPDEs and FBSDEs (Theorems 5.8–5.9), and a mean-field version is treated by injecting small independent noises and deriving a Wasserstein HJB equation (Theorems 6.8–6.9). Applications to the original Stackelberg problem and to principal-agent problems are given in Sections 6.1–6.2, but these are explicitly called heuristic. Several examples in Section 7 illustrate the distinction between optimistic and pessimistic equilibria, the gap between V0 and the set value v0, and the failure of the Isaacs-equation characterization for open-loop games.

Significance. If the advertised Wasserstein HJB characterization were rigorously established for the original robust Stackelberg problem, this would be a substantial contribution: it would provide a PDE-type characterization for a pessimistic multi-follower Stackelberg game in continuous time, address equilibrium selection through approximate equilibria, and expose the time-inconsistency induced by open-loop controls. The paper's reformulation steps are interesting and largely coherent; Theorem 2.4 is a clean consequence of the external ε-equilibrium characterization in [23], and the explicit examples (7.1–7.4) are useful and correctly computed. The authors are also refreshingly transparent about the strong and often unverified assumptions needed for the later theorems, and they explicitly flag Section 6.1 as heuristic. However, the gap between the conditional general framework and the original motivating problem is load-bearing: the abstract claims a characterization of the robust leader's value, but the paper proves such a characterization only under strong regularity hypotheses that are acknowledged to be unlikely to hold for the Stackelberg system. I found no circularity: the penalized value

major comments (3)
  1. [Section 6.1 (Application to (3.2))] The abstract states that the zero-sum game induced by the robust Stackelberg problem (3.2) is characterized by an HJB equation on the Wasserstein space. However, Section 6.1 says explicitly that the derivation is "just heuristic" and "requires very strong technical conditions, which are unlikely satisfied by the system for (3.2)". Concretely, the control β in (2.6) is unbounded, so the induced coefficient σ in (6.1) is unbounded, and Assumption 6.1 (boundedness and uniform Lipschitz continuity in (μ,x)) fails. Thus Theorem 6.9 does not apply to the original Stackelberg problem, and the advertised characterization of V0 or Vλ for that problem is not established. A rigorous approximation/regularity argument, or a substantial weakening of Assumption 6.1 that covers unbounded volatility controls, is required before the central claim of the paper is supported.
  2. [Theorems 6.8 and 6.9; Assumptions 6.4 and 6.7] The convergence lim_{λ→∞} V_{λ,κ} = Vκ and the final Wasserstein HJB characterization rest on Assumptions 6.4 and 6.7. In particular, 6.7(iii) requires that the penalized problem (6.18) has 1/λ-optimal controls with α^{0,λ,κ} ∈ A0 and uniformly bounded derivatives of the associated U^{κ,α0}, uniformly in (λ,κ), and 6.4 requires global invertibility of q ↦ q + γ ∂_{γ0}H. No example or nontrivial class of coefficients is shown to satisfy these assumptions; for the Stackelberg system of Section 2, the unbounded β makes them highly questionable. Since these assumptions are not verified, Theorem 6.9 is a conditional structural theorem for a general class, not a characterization of the value of the games introduced in Sections 2–3.
  3. [Theorem 5.8 and Assumption 5.7] In the special case of Section 5, part (i) gives only V0 ≤ Vλ for all λ, while the claimed convergence lim_{λ→∞} Vλ = V0 in part (ii) depends entirely on Assumption 5.7, which postulates the existence of 1/λ-optimal controls with uniformly bounded derivatives and classical BSPDE solutions. The paper does not provide verifiable sufficient conditions for Assumption 5.7, nor an example in this section where it holds. Thus the penalization approximation of V0 in the non-mean-field case is also conditional. This is a load-bearing point because the penalization method is the paper's main methodological tool.
minor comments (5)
  1. [Equation (5.1)] In the definition of J(α̂), the running-cost integral is written as ∫_0^t f_s ds; it should almost certainly be ∫_0^T f_s ds, consistently with the rest of the paper.
  2. [Equations (6.6), (6.14), and (6.16)] The indexing is inconsistent: the state dimension is d in (6.1)–(6.2), but several sums and components are written with n (e.g., "for i=1,…,n" around (6.6) and the sum in (6.14)). Please unify to d.
  3. [p.16 (Remark 5.4)] "Duprie's derivatives" should be "Dupire's derivatives".
  4. [Remark 3.3(ii)] The notation Vλ, V1, V1, Vλ is confusing at first reading; a sentence defining each symbol in words would improve readability.
  5. [References] The paper relies on [48] and [51] for viscosity solutions of master equations and HJB equations on Wasserstein spaces, but these references are cited without specifying the exact theorem used. Since Theorem 6.9 assumes classical solvability, the exact comparison/uniqueness result from [51] should be stated or quoted precisely.

Circularity Check

0 steps flagged

No significant circularity: the penalized-value/HJB chain is a conditional verification argument, and the paper's own Section 6.1 caveat about the heuristic application is a scope limitation, not a circular reduction.

full rationale

I walked the derivation chain V0 (1.7) -> Vλ (2.11) -> reduced zero-sum problem (3.2) -> Wasserstein HJB (6.23) -> Theorem 6.9. No step reduces to its own input by construction. Theorem 2.4 bridges Vλ and V0 through the external ε-equilibrium characterization in [23], which is a published, standalone result rather than a fitted parameter or an identity; the fact that one of the present authors coauthored [23] does not make the bridge definitional. The elimination of y in (3.1) is an explicit minimization, and the final identity V_{λ,κ} = sup_y U(0, δ_{(0,x,y)}) in Theorem 6.9 is the natural statement that the value with initial data (0,x,y) is U, not a rearrangement that assumes the conclusion. The HJB equations (5.17) and (6.23) are derived by Itô calculus and dynamic programming under stated regularity assumptions, and no fitted constants or renamed existing results appear. The paper's own limitation—Section 6.1 states that applying the general theorem to the original Stackelberg system (3.2) is 'just heuristic' because the β in (2.6) is unbounded and hence Assumption 6.1 is unlikely to hold—is a genuine gap between the abstract's broad claim and the proved conditional theorem, but it is a correctness/scope issue, not circularity. Therefore the circularity score is 0.

Axiom & Free-Parameter Ledger

2 free parameters · 6 axioms · 1 invented entities

The paper's central claim rests on external BSDE/game-theoretic characterizations and on several strong regularity and well-posedness assumptions that are stated but not verified for the motivating Stackelberg system. The constants λ and κ are analytical regularization parameters, not fitted physical constants.

free parameters (2)
  • λ (penalty weight) = λ → ∞ (limit)
    Introduced in (2.11), (5.14), and (6.18) as a penalization weight to relax ε-equilibrium and terminal constraints; the main theorems require taking λ→∞ under uniform regularity assumptions.
  • κ (noise-injection amplitude) = κ → 0 (limit)
    Added in (6.2) as independent Brownian noise κB^i to make the coupled FBSDE well posed; Lemma 6.2 gives approximation error Cκ, and the final characterization is after κ→0.
axioms (6)
  • standard math Weak formulation / Girsanov representation (2.2)-(2.3) with Assumption 2.1 boundedness
    Used throughout to express players' utilities as expectations under P^αhat; requires bounded and uniformly continuous coefficients.
  • standard math Theorem 4 of Feinstein-Rudloff-Zhang [23] characterizing ε-equilibria via the functional I
    Basis of Proposition 2.3 and therefore of Theorem 2.4; not proved in this paper, but a published standalone result.
  • domain assumption E_ε(α0) nonempty for every α0 and ε (stated in Theorem 2.4 and Theorem 4.3)
    Needed for the penalization limit Vλ ↑ V0; the paper notes the true equilibrium set can be empty in general.
  • ad hoc to paper Assumptions 5.5 and 6.7(ii): A0 is dense in A and 1/λ-optimal controls exist with uniformly bounded derivatives
    Needed for convergence Vλ → V0 and for the HJB representation; unverified and likely hard for the motivating Stackelberg system.
  • ad hoc to paper Assumption 6.4: invertibility of q ↦ q + γ ∂_{γ0} H
    Needed to solve for ∂_{x0} v_i and close the FBSDE in Section 6; no evidence it holds for the Stackelberg system.
  • ad hoc to paper Classical solvability of BSPDEs (5.5), (5.17), and Wasserstein HJB equation (6.23)
    The final characterizations in Theorems 5.9 and 6.9 are conditional on classical solutions with bounded derivatives.
invented entities (1)
  • Independent Brownian noises B^1,...,B^d with small amplitude κ no independent evidence
    purpose: Regularization of the forward-backward system in Section 6; the value Vκ approximates V0 with error Cκ.
    Pure analytical regularization device; no physical or empirical content, and no falsifiable prediction outside the paper.

pith-pipeline@v1.3.0-alltime-deepseek · 31247 in / 19109 out tokens · 177943 ms · 2026-08-01T06:25:23.402517+00:00 · methodology

0 comments
read the original abstract

In this paper we study a Stackelberg game with one leader and multiple followers. Given the leader's control, the followers solve a Nash game with possibly multiple equilibria. We consider a robust leader who considers the worst scenario, namely the followers would select the equilibrium worst for the leader. By using the weak formulation, the problem induces a zero sum game problem with open loop controls, which is time inconsistent. We shall characterize the last problem through an HJB equation on the Wasserstein space of probability measures. The principal-agent problem with one principal and multiple agents can be viewed as a special case of our problem, but with certain constraints.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

51 extracted references · 3 linked inside Pith

  1. [1]

    and Svensson, A., 2020

    Aussel, D. and Svensson, A., 2020. A short state of the art on multi-leader-follower games, in Bilevel optimization: Advances and next challenges, pp.53-76

    2020

  2. [2]

    and Olsder, G.J., 1995

    Ba¸ sar, T. and Olsder, G.J., 1995. Dynamic noncooperative game theory. Society for Indus- trial and Applied Mathematics

    1995

  3. [3]

    Optimal Replication of Contingent Claims under Portfolio Constraints, Rev

    Broadie, M., Cvitani´ c, J., and Soner, H.M., 1998. Optimal Replication of Contingent Claims under Portfolio Constraints, Rev. Financial Studies, 11, 59-79

    1998

  4. [4]

    and Li, J., 2008

    Buckdahn, R. and Li, J., 2008. Stochastic differential games and viscosity solutions of Hamil- ton–Jacobi–Bellman–Isaacs equations. SIAM Journal on Control and Optimization, 47(1), pp.444-475

    2008

  5. [5]

    and Rainer, C., 2009

    Cardaliaguet, P. and Rainer, C., 2009. Stochastic differential games with asymmetric infor- mation. Applied Mathematics and Optimization, 59, pp.1-36. 32

    2009

  6. [6]

    stochastic differential games with asymmetric information

    Cardaliaguet, P. and Rainer, C., 2013. Pathwise strategies for stochastic differential games with an erratum to “stochastic differential games with asymmetric information”. Applied Mathematics and Optimization, 68, pp.75-84

    2013

  7. [7]

    and Delarue, F., 2018

    Carmona, R. and Delarue, F., 2018. Probabilistic theory of mean field games with appli- cations I - Mean field FBSDEs, control, and games, Probability Theory and Stochastic Modeling, 83. Springer, Cham

    2018

  8. [8]

    and Delarue, F., 2018

    Carmona, R. and Delarue, F., 2018. Probabilistic theory of mean field games with applica- tions II - Mean field games with common noise and master equations, Probability Theory and Stochastic Modeling, 84. Springer, Cham

    2018

  9. [9]

    and Fourni´ e, D.A., 2013

    Cont, R. and Fourni´ e, D.A., 2013. Functional Itˆ o calculus and stochastic integral represen- tation of martingales, Annals of Probability, 41(1), pp. 109-133

    2013

  10. [10]

    and Pham, H., 2019

    Cosso, A. and Pham, H., 2019. Zero-sum stochastic differential games of generalized McK- ean–Vlasov type. Journal de Math´ ematiques Pures et Appliqu´ ees, 129, pp.180-212

    2019

  11. [11]

    and Touzi, N., 2017

    Cvitani´ c, J., Possama ¨ ı, D. and Touzi, N., 2017. Moral hazard in dynamic risk management. Management Science, 63(10), pp.3328-3346

    2017

  12. [12]

    and Touzi, N., 2018

    Cvitani´ c, J., Possama ¨ ı, D. and Touzi, N., 2018. Dynamic programming approach to princi- pal–agent problems. Finance and Stochastics, 22(1), pp.1-37

    2018

  13. [13]

    Methods of variational analysis in pessimistic bilevel programming

    Dassanayaka, S.M., 2010. Methods of variational analysis in pessimistic bilevel programming. Ph.D. thesis, Department of Mathematics, Wayne State University, Detroit

    2010

  14. [14]

    Foundations of bilevel programming

    Dempe, S., 2002. Foundations of bilevel programming. Boston, MA: Springer US

    2002

  15. [15]

    and Mordukhovich, B.S., 2007

    Dempe, S., Dutta, J. and Mordukhovich, B.S., 2007. New necessary optimality conditions in optimistic bilevel programming. Optimization, 56(5-6), pp.577-604

    2007

  16. [16]

    and Zemkoho, A.B., 2014

    Dempe, S., Mordukhovich, B.S. and Zemkoho, A.B., 2014. Necessary optimality conditions in pessimistic bilevel programming. Optimization, 63(4), pp.505-533

    2014

  17. [17]

    Stackelberg mean field games: convergence and existence results to the problem of principal with multiple agents in competition

    Djete, M.F., 2023. Stackelberg mean field games: convergence and existence results to the problem of principal with multiple agents in competition. arXiv preprint arXiv:2309.00640

    Pith/arXiv arXiv 2023

  18. [18]

    Functional Itˆ o calculus

    Dupire, B., 2019. Functional Itˆ o calculus. Quantitative Finance, 19(5), pp.721-729

    2019

  19. [19]

    and Jentzen, A., 2017

    E, W., Han, J. and Jentzen, A., 2017. Deep learning-based numerical methods for high- dimensional parabolic partial differential equations and backward stochastic differential equations. Communications in mathematics and statistics, 5, pp.349-380. 33

    2017

  20. [20]

    and Possama ¨ ı, D., 2019

    Elie, R., Mastrolia, T. and Possama ¨ ı, D., 2019. A tale of a principal and many, many agents. Mathematics of Operations Research, 44(2), pp.440-467

    2019

  21. [21]

    and Possama ¨ ı, D., 2019

    Elie, R. and Possama ¨ ı, D., 2019. Contracting theory with competitive interacting agents. SIAM Journal on Control and Optimization, 57(2), pp.1157-1188

    2019

  22. [22]

    and Huang, J., 2024

    Feng, X., Hu, Y. and Huang, J., 2024. Linear-quadratic two-person differential game: Nash game versus Stackelberg game, local information versus global information. ESAIM: Control, Optimisation and Calculus of Variations, 30(47), pp. 1-41

    2024

  23. [23]

    and Zhang, J., 2022

    Feinstein, Z., Rudloff, B. and Zhang, J., 2022. Dynamic set values for nonzero-sum games with multiple equilibriums. Mathematics of Operations Research, 47(1), pp.616-642

    2022

  24. [24]

    and Souganidis, P.E., 1989

    Fleming, W.H. and Souganidis, P.E., 1989. On the existence of value functions of two- player, zero-sum stochastic differential games. Indiana University Mathematics Journal, 38(2), pp.293-314

    1989

  25. [25]

    and Tirole, J., 1991

    Fudenberg, D. and Tirole, J., 1991. Game theory. MIT press

    1991

  26. [26]

    and Stokey, N.L., 1983

    Green, J.R. and Stokey, N.L., 1983. A comparison of tournaments and contracts. Journal of Political Economy, 91(3), pp.349-364

    1983

  27. [27]

    and Lepeltier, J.P., 1995

    Hamadene, S. and Lepeltier, J.P., 1995. Zero-sum stochastic differential games and backward equations. Systems and Control Letters, 24(4), pp.259-263

    1995

  28. [28]

    and Peng, S., 1997

    Hamadene, S., Lepeltier, J.P. and Peng, S., 1997. BSDEs with continuous coefficients and stochastic differential games. In Backward Stochastic Differential Equations (El Karoui N and Mazliak. L eds.) Pitman res notes math ser, 364. Longman, Harlow, pp 115–128

    1997

  29. [29]

    and Selten, R., 1988

    Harsanyi, J.C. and Selten, R., 1988. A general theory of equilibrium selection in games. MIT Press Books

    1988

  30. [30]

    and Wu, Z., 2021

    Huang, J., Wang, S. and Wu, Z., 2021. Robust Stackelberg differential game with model uncertainty. IEEE Transactions on Automatic Control, 67(7), pp.3363-3380

    2021

  31. [31]

    and Sung, J., 2008

    Keun Koo, H., Shim, G. and Sung, J., 2008. Optimal Multi-Agent Performance Measures for Team Contracts. Mathematical Finance, 18(4), pp.649-667

    2008

  32. [32]

    and Zheng, Y., 2018

    Liu, J., Fan, Y., Chen, Z. and Zheng, Y., 2018. Pessimistic bilevel optimization: a survey. International Journal of Computational Intelligence Systems, 11(1), pp.725-736. 34

    2018

  33. [33]

    and Zheng, Y., 2020

    Liu, J., Fan, Y., Chen, Z. and Zheng, Y., 2020. Methods for pessimistic bilevel optimiza- tion. In Bilevel optimization: Advances and next challenges (pp. 403-420). Cham: Springer International Publishing

    2020

  34. [34]

    and Yong, J., 1994

    Ma, J., Protter, P. and Yong, J., 1994. Solving forward-backward stochastic differential equations explicitly - a four step scheme, Probab. Theory Relat. Fields., 98(3), pp. 339-359

    1994

  35. [35]

    Solvability of forward-backward SDEs and the nodal set of Hamilton- Jacobi-Bellman equations

    Ma, J., Yong, J., 1995. Solvability of forward-backward SDEs and the nodal set of Hamilton- Jacobi-Bellman equations. Chin. Ann. Math. Ser. B 16(3), pp. 279–298

    1995

  36. [36]

    Optimal incentive schemes with many agents

    Mookherjee, D., 1984. Optimal incentive schemes with many agents. The Review of Eco- nomic Studies, 51(3), pp.433-446

    1984

  37. [37]

    and Yong, J., 2006

    Mou, L. and Yong, J., 2006. Two-person zero-sum linear quadratic stochastic differential games by a Hilbert space method. J. Ind. Manag. Optim, 2(1), pp.93-115

    2006

  38. [38]

    Stochastic Hamilton–Jacobi–Bellman equations

    Peng, S., 1992. Stochastic Hamilton–Jacobi–Bellman equations. SIAM Journal on Control and Optimization, 30(2), pp.284-304

    1992

  39. [39]

    and Zhang, J., 2014

    Pham, T. and Zhang, J., 2014. Two person zero-sum game in weak formulation and path dependent Bellman–Isaacs equation. SIAM Journal on Control and Optimization, 52(4), pp.2090-2121

    2014

  40. [40]

    and Zhang, J., 2020

    Possama ¨ ı, D., Touzi, N. and Zhang, J., 2020. Zero-sum path-dependent stochastic differential games in weak formulation. Annals of Applied Probability, 30, pp.1415-1457

    2020

  41. [41]

    A continuous-time version of the principal-agent problem

    Sannikov, Y., 2008. A continuous-time version of the principal-agent problem. The Review of Economic Studies, 75(3), pp.957-984

    2008

  42. [42]

    and Bard, J.F., 1997

    Shimizu, K., Ishizuka, Y. and Bard, J.F., 1997. Nondifferentiable and Two-level Mathemat- ical Programming, Kluwer Academic Publishers, Dordrecht

    1997

  43. [43]

    Stochastic Perron’s method and elementary strategies for zero-sum differ- ential games

    Sirbu, M., 2014. Stochastic Perron’s method and elementary strategies for zero-sum differ- ential games. SIAM Journal on Control and Optimization, 52(3), pp.1693-1711

    2014

  44. [44]

    Asymptotic perron’s method and simple Markov strategies in stochastic games and control

    S ˆ ırbu, M., 2015. Asymptotic perron’s method and simple Markov strategies in stochastic games and control. SIAM Journal on Control and Optimization, 53(4), pp.1713-1733

    2015

  45. [45]

    and Touzi, N., 2000

    Soner, H.M. and Touzi, N., 2000. Super-Replication under Gamma Constraints, SIAM J.Control Opt. 39(1), 73–96

    2000

  46. [46]

    Two-person zero-sum stochastic linear-quadratic differential games

    Sun, J., 2021. Two-person zero-sum stochastic linear-quadratic differential games. SIAM Journal on Control and Optimization, 59(3), pp.1804-1829. 35

    2021

  47. [47]

    and Wen, J., 2023

    Sun, J., Wang, H. and Wen, J., 2023. Zero-sum Stackelberg stochastic linear-quadratic differential games. SIAM Journal on Control and Optimization, 61(1), pp.250-282

    2023

  48. [48]

    and Zhang, J., 2020

    Wu, C. and Zhang, J., 2020. Viscosity solutions to parabolic master equations and McK- ean–Vlasov SDEs with closed-loop controls. Annals of Applied Probability, 30, 936-986

    2020

  49. [49]

    Backward stochastic differential equations

    Zhang, J., 2017. Backward stochastic differential equations. In Backward Stochastic Differ- ential Equations: From Linear to Fully Nonlinear Theory. New York, NY: Springer New York

    2017

  50. [50]

    Instability and efficiency of non-cooperative games

    Zhang, J., 2024. Instability and efficiency of non-cooperative games. Mathematical Finance, Special Issue in Honour of Prof. Xunyu Zhou, accepted, arXiv:2405.17196

    Pith/arXiv arXiv 2024

  51. [51]

    and Zhang, J., 2024

    Zhou, J., Touzi, N. and Zhang, J., 2024. Viscosity solutions for HJB equations on the process space: Application to mean field control with common noise. preprint, arXiv:2401.04920. 36

    Pith/arXiv arXiv 2024