betaFeedback

WorksMathematics

Singular Tracking of Fractional Brownian Motion under Proportional Costs

Abstract

We study the tracking of fractional Brownian motion by adapted controls of finite variation, with quadratic loss and proportional adjustment costs. Let BHB^H be standard fractional Brownian motion with Hurst parameter H∈(0,1)H\in(0,1). We write Vλ(T;a,b,σ)V_λ(T;a,b,σ) for the optimal expected cost of tracking σBHσB^H over a horizon T>0T>0, where σ>0σ>0 is the target amplitude, a>0a>0 weights the quadratic loss, and λbλb is the cost per unit of adjustment, with λ,b>0λ,b>0. We prove Vλ(T;a,b,σ)∼λ→0κHTa(1−H)/(1+H)(λb)2H/(1+H)σ2/(1+H).V_λ(T;a,b,σ)\underset{λ\to0}{\sim}κ_HT a^{(1-H)/(1+H)}(λb)^{2H/(1+H)}σ^{2/(1+H)}. Here κH>0κ_H>0 is the long-run optimal cost per unit time in the normalized problem a=b=σ=λ=1a=b=σ=λ=1. Observing the entire past does not change κHκ_H. For the normalized problem with full past information, a horizon L>0L>0 and a terminal adjustment to the target, the difference between its value and κHLκ_H L is bounded uniformly in LL. This estimate also yields continuity of κHκ_H in HH. In the Brownian case, κ1/2=12(32)2/3κ_{1/2}=\frac12(\frac32)^{2/3}. We extend the asymptotic formula to nonstationary targets under a coupling condition on local paths and their observation filtrations. We verify this condition for Gaussian Volterra targets under Brownian observation. All asymptotically optimal controls have the same leading-order split between tracking and adjustment costs. We also give primal and dual bounds on κHκ_H. For H<1/2H<1/2, we construct approximations based on exponential sums with positive weights and quantify their error as the horizon grows.

Introduction

Fix a time horizon T>0T>0 and an observation filtration F=(Ft)0≤t≤T\mathbf{F}=(\mathcal{F}_t)_{0\le t\le T}, where Ft\mathcal{F}_t represents the information available at time tt. A controller tracks a continuous adapted target Θ=(Θt)0≤t≤T\Theta=(\Theta_t)_{0\le t\le T} by choosing a control ϕ=(ϕt)0≤t≤T\phi=(\phi_t)_{0\le t\le T}. Adaptedness means that the target and control at time tt are measurable with respect to Ft\mathcal{F}_t. The positive deterministic weights ata_t and btb_t specify the relative importance of tracking error and adjustments. The parameter λ>0\lambda>0 sets the overall adjustment-cost level. Writing E\mathbb{E} for expectation, we minimize

Jλ(ϕ;Θ):=E[12∫0Tat(ϕt−Θt)2 dt+λ∫[0,T]bt d∣ϕ∣t].(1)J_\lambda(\phi;\Theta):=\mathbb{E}\left[\frac{1}{2}\int_0^T a_t(\phi_t-\Theta_t)^2\,\mathrm{d}t+\lambda\int_{[0,T]}b_t\,\mathrm{d}|\phi|_t\right]. \tag*{(1)}

We minimize over adapted, càdlàg controls of finite variation and write

Vλ(Θ):=inf⁡ϕJλ(ϕ;Θ).V_\lambda(\Theta):=\inf_{\phi}J_\lambda(\phi;\Theta).

Here càdlàg means right-continuous with left limits, and finite variation means that the total amount of adjustment is finite. The notation d∣ϕ∣t\mathrm{d}|\phi|_t denotes the total-variation measure of the control increments. It includes the initial adjustment from the prescribed pre-initial position ϕ0−=Θ0\phi_{0-}=\Theta_0, with cost λb0∣ϕ0−−Θ0∣\lambda b_0|\phi_{0-}-\Theta_0|. § 2 gives the precise admissible class, integrability requirements and coefficient assumptions. Suppose that target increments over a small time interval of length h>0h>0 have typical size hHh^H, where H∈(0,1)H\in(0,1) describes their scaling. An error tolerance δ>0\delta>0 then suggests an adjustment interval h≍δ1/Hh\asymp\delta^{1/H}. We use ≍\asymp for comparison up to positive multiplicative constants independent of the small parameter. The tracking cost per unit time is then of order δ2\delta^2, whereas the variation-cost rate satisfies

λδh≍λδ1−1/H.\lambda\frac{\delta}{h}\asymp\lambda\delta^{1-1/H}.

Balancing these two rates suggests a tracking-error scale δλ\delta_\lambda and a value of order

δλ≍λH/(1+H),Vλ(Θ)≍λ2H/(1+H).(2)\delta_\lambda\asymp\lambda^{H/(1+H)}, \qquad V_\lambda(\Theta) \asymp\lambda^{2H/(1+H)}. \tag*{(2)}

An increment–energy inequality proves the lower bound even for noncausal controls, which have access to future information. Foresight therefore leaves the exponent unchanged. For the fractional Brownian target, let BHB^H be standard fractional Brownian motion with Hurst parameter HH, normalized by B0H=0B^H_0=0 and

E[(Bt+hH−BtH)2]=h2H,t≥0, h>0.\mathbb{E}\left[(B^H_{t+h}-B^H_t)^2\right]=h^{2H}, \qquad t\geq0,\ h>0.

The constant σ>0\sigma>0 scales the target amplitude. For Θ=σBH\Theta=\sigma B^H and constant weights at=a>0a_t=a>0 and bt=b>0b_t=b>0, we write

Vλ(T;a,b,σ):=Vλ(σBH),V_\lambda(T;a,b,\sigma):=V_\lambda(\sigma B^H),

using the usual natural filtration of BHB^H, so that controls use only observations of the target from time zero up to the current time. The natural time scale is

ℓλ:=(λbaσ)1/(1+H).(3)\ell_\lambda:=\left(\frac{\lambda b}{a\sigma}\right)^{1/(1+H)}. \tag*{(3)}

Rescaling time, the target, the control and the cost transforms the small-cost problem into a normalized tracking problem, called the cell problem, with a=b=σ=λ=1a=b=\sigma=\lambda=1. Its horizon is L:=T/ℓλL:=T/\ell_\lambda, which tends to infinity as λ→0\lambda\to0. The limiting optimal cost per unit time is denoted by κH\kappa_H. A terminal reset is a final adjustment of the control to the target. Adding its cost makes the cell value subadditive, and the extra cost is sublinear in LL. Consequently,

Vλ(T;a,b,σ)∼λ→0κHTa(1−H)/(1+H)(λb)2H/(1+H)σ2/(1+H),κH∈(0,∞).(4)V_\lambda(T;a,b,\sigma)\mathop{\sim}_{\lambda\to0}\kappa_H T a^{(1-H)/(1+H)}(\lambda b)^{2H/(1+H)}\sigma^{2/(1+H)}, \qquad\kappa_H\in(0,\infty). \tag*{(4)}

Full past information means that the controller also observes the target at all negative times, using a two-sided fractional Brownian motion. This additional information can change finite-horizon values but not κH\kappa_H. For any normalized horizon L>0L>0, let uHpast(L)u_H^{\mathrm{past}}(L) denote the optimal expected cost of the normalized cell problem with full past information and an additional terminal reset cost. Averaging over shifted partitions gives

0≤uHpast(L)−κHL≤22κH,0\leq u_H^{\mathrm{past}}(L)-\kappa_H L\leq2\sqrt{2\kappa_H},

which also yields continuity in HH when fractional Brownian motions with different Hurst parameters are constructed from the same Brownian motion. For nonstationary targets, let ρ:[0,T]→(0,∞)\rho:[0,T]\to(0,\infty) be the continuous local amplitude: near time tt, target increments over a small interval of length hh have characteristic scale ρ(t)hH\rho(t)h^H. Thus ρ(t)\rho(t) plays the local role of σ\sigma in the homogeneous model. The filtered tangent condition in assumption 6.1 couples the normalized local paths with standard fractional Brownian motion and matches their observation filtrations. Under this condition, for continuous positive weights aa and bb,

λ−2H/(1+H)Vλ(Θ)⟶λ→0κH∫0Tat(1−H)/(1+H)bt2H/(1+H)ρ(t)2/(1+H) dt.\lambda^{-2H/(1+H)}V_\lambda(\Theta)\mathop{\longrightarrow}_{\lambda\to0}\kappa_H\int_0^T a_t^{(1-H)/(1+H)}b_t^{2H/(1+H)}\rho(t)^{2/(1+H)}\,\mathrm{d}t.

We verify the condition for Gaussian Volterra targets when their driving Brownian motion is observed. Perturbing the coefficients then identifies the leading tracking and variation costs for every asymptotically optimal family of controls, meaning one whose cost divided by Vλ(Θ)V_\lambda(\Theta) tends to one.

Contributions. The asymptotic formula for fractional Brownian motion in theorem 3.4 extends to local Volterra targets through theorem 6.2 and proposition 6.8. The proof uses the information and reset estimates in theorems 5.3 and 5.4. Prediction states account for the past information carried by fractional targets. To estimate the cell constant, we derive primal and dual bounds and approximations based on exponential sums with positive weights in § 7 and § 8.

Related work. Our Brownian benchmark builds on singular-control results [1, 2, 3] and small-cost tracking of continuous Itô targets [4, 5]. Fractional linear–quadratic control with absolutely continuous controls is treated in [6, 7, 8, 9, 10]. Schelling [11] links rough-target tracking to an ergodic fractional cell under quadratic trading costs. In our setting, the penalty is linear in variation and permits singular controls. One-sided fractional singular control is studied in [12, 13]. Truncated variation and hard-tube approximation [14, 15, 16, 17] impose pathwise constraints. Related level-crossing asymptotics for fractional Brownian motion are studied in [18]. Prediction states and Markovian fractional or Volterra representations appear in [19, 20, 21, 22, 23, 24], while positive exponential approximations are studied in [25, 26, 27, 28]. We approximate control values under a fixed observation filtration, with stationary-past and growing-horizon error bounds. The hedging application in section J gives a penalized mean-square identity. For self-financing models with proportional transaction costs in fractional finance, see [29].

Organization. We first establish the rate and asymptotic formula in § 2 and § 3, before turning to Brownian reflection and past information in § 4 and § 5. These results lead to the local limit theorem and its Volterra application in § 6. We then develop the bounds and factor approximations in § 7 and § 8. The appendices give the filtration and path-space arguments, the prediction-state construction, and the technical proofs used in the main text. They also contain the multi-window bounds and the hedging identity in § 18.§ 2 and J.

Model and rate bounds

Notation and admissibility

Fix a complete filtered probability space (Ω,F,F,P)(\Omega,\mathcal{F},\mathbf{F},\mathbb{P}), where F=(Ft)t≥0\mathbf{F}=(\mathcal{F}_t)_{t\geq0}, and a horizon T>0T>0. All filtrations are complete and right-continuous. Expectations, variances, and covariances are denoted by E\mathbb{E}, Var⁡\operatorname{Var}, and Cov⁡\operatorname{Cov}. A bar denotes completion under the indicated probability measure. For a standard Borel space EE, write B(E)\mathcal{B}(E) for its Borel sigma-field and P(E)\mathcal{P}(E) for the set of probability measures on EE. We use L0(Ω,H,P;E)L^0(\Omega,\mathcal{H},\mathbb{P};E) for the H\mathcal{H}-measurable EE-valued random variables, identified almost surely. For a separable Banach space (E,∥⋅∥E)(E,\|\cdot\|_E) and 1≤p<∞1\leq p<\infty, the subspace Lp(Ω,H,P;E)L^p(\Omega,\mathcal{H},\mathbb{P};E) of L0(Ω,H,P;E)L^0(\Omega,\mathcal{H},\mathbb{P};E) consists of the variables with E[∥X∥Ep]<∞\mathbb{E}[\|X\|_E^p]<\infty and carries the norm ∥X∥Lp:=(E[∥X∥Ep])1/p\|X\|_{L^p}:=(\mathbb{E}[\|X\|_E^p])^{1/p}. Likewise, L∞L^\infty uses the essential-supremum norm, while deterministic Lp(I;E)L^p(I;E) uses Lebesgue measure on II. Omitted sigma-fields and measures are those of the ambient space. An omitted scalar codomain is R\mathbb{R}. Eμ\mathbb{E}_\mu denotes expectation under a specified law μ\mu. On Rd\mathbb{R}^d, x⋅y=x⊤yx\cdot y=x^\top y and ∥x∥=(x⊤x)1/2\|x\|=(x^\top x)^{1/2} are the Euclidean inner product and norm. Inner products and norms with subscript 2 refer to the stated L2L^2 space. We write L(X)\mathcal{L}(X) for the law of XX, =d\overset{d}{=} for equality in law, and U⊥ ⁣ ⁣ ⁣⊥H∣YU\perp\!\!\!\perp\mathcal{H}\mid Y for conditional independence given YY. Conditional essential infima are taken in the almost-sure order. C(I;E)\mathcal{C}(I;E) and D(I;E)\mathcal{D}(I;E) denote continuous and càdlàg paths, with the uniform and Skorokhod J1J_1 topologies on compact II. Cloc\mathcal{C}_{\mathrm{loc}} uses uniform convergence on compact intervals. For Hölder estimates, use

[f]α:=sup⁡s≠t∣f(t)−f(s)∣∣t−s∣α,0<α≤1.[f]_\alpha:=\sup_{s\ne t}\frac{|f(t)-f(s)|}{|t-s|^\alpha},\qquad0<\alpha\leq1.

For 0<α<10<\alpha<1, Cα\mathcal{C}^\alpha is the space of continuous functions with finite [f]α[f]_\alpha. When derivatives are specified, Ck\mathcal{C}^k denotes kk continuous derivatives. Set CT:=C([0,T];R)\mathcal{C}_T:=\mathcal{C}([0,T];\mathbb{R}) and DT:=D([0,T];R)\mathcal{D}_T:=\mathcal{D}([0,T];\mathbb{R}). For x∈DTx\in\mathcal{D}_T and 0≤r<s≤T0\leq r<s\leq T, define

References

  1. [1]I. Karatzas. “A Class of Singular Stochastic Control Problems”. In: Advances in Applied Probability 15.2 (1983), pp. 225–254. DOI: 10.2307/1426435.DOI
  2. [2]J. M. Harrison and M. I. Taksar. “Instantaneous Control of Brownian Motion”. In: Mathematics of Operations Research 8.3 (1983), pp. 439–453. DOI: 10.1287/moor.8.3.439.DOI
  3. [3]M. I. Taksar. “Average Optimal Singular Control and a Related Stopping Problem”. In: Mathematics of Operations Research 10.1 (1985), pp. 63–81. DOI: 10.1287/moor.10.1.63.DOI
  4. [4]J. Cai, M. Rosenbaum, and P. Tankov. “Asymptotic Lower Bounds for Optimal Tracking: A Linear Programming Approach”. In: The Annals of Applied Probability 27.4 (2017), pp. 2455–2514. DOI: 10.1214/16-AAP1264.DOI
  5. [5]J. Cai, M. Rosenbaum, and P. Tankov. “Asymptotic Optimal Tracking: Feedback Strategies”. In: Stochastics 89.6–7 (2017), pp. 943–966. DOI: 10.1080/17442508.2017.1285304.DOI
  6. [6]Y. Hu and X. Y. Zhou. “Stochastic Control for Linear Systems Driven by Fractional Noises”. In: SIAM Journal on Control and Optimization 43.6 (2005), pp. 2245–2277. DOI: 10.1137/S0363012903426045.DOI
  7. [7]M. L. Kleptsyna, A. Le Breton, and M. Viot. “About the Linear–Quadratic Regulator Problem under a Fractional Brownian Perturbation”. In: ESAIM: Probability and Statistics 7 (2003), pp. 161–170. DOI: 10.1051/ps:2003007.DOI
  8. [8]M. L. Kleptsyna, A. Le Breton, and M. Viot. “On the Infinite Time Horizon Linear–Quadratic Regulator Problem under a Fractional Brownian Perturbation”. In: ESAIM: Probability and Statistics 9 (2005), pp. 185–205. DOI: 10.1051/ps:2005008.DOI
  9. [9]T. E. Duncan, B. Maslowski, and B. Pasik-Duncan. “Linear–Quadratic Control for Stochastic Equations in a Hilbert Space with Fractional Brownian Motions”. In: SIAM Journal on Control and Optimization 50.1 (2012), pp. 507–531. DOI: 10.1137/110831416.DOI
  10. [10]T. E. Duncan and B. Pasik-Duncan. “Linear–Quadratic Fractional Gaussian Control”. In: SIAM Journal on Control and Optimization 51.6 (2013), pp. 4504–4519. DOI: 10.1137/120877283.DOI
  11. [11]D. M. Schelling. “Rough Volatility and Portfolio Optimisation under Small Transaction Costs”. PhD thesis. London School of Economics and Political Science, 2019. URL: https://researchonline.lse.ac.uk/id/eprint/134983/.
  12. [12]A. P. Ghosh, A. Roitershtein, and A. Weerasinghe. “Optimal Control of a Stochastic Processing System Driven by a Fractional Brownian Motion Input”. In: Advances in Applied Probability 42.1 (2010), pp. 183–209. DOI: 10.1239/aap/1269611149.
  13. [13]C. Lee and A. Weerasinghe. “Stationarity and Control of a Tandem Fluid Network with Fractional Brownian Motion Input”. In: Advances in Applied Probability 43.3 (2011), pp. 847–874. DOI: 10.1239/aap/1316792672.
  14. [14]W. M. Bednorz and R. M. Łochowski. “Integrability and Concentration of the Truncated Variation for the Sample Paths of Fractional Brownian Motions, Diffusions and Lévy Processes”. In: Bernoulli 21.1 (2015), pp. 437–464. DOI: 10.3150/13-BEJ574.
  15. [15]R. M. Łochowski. “On a Generalisation of the Hahn–Jordan Decomposition for Real Càdlàg Functions”. In: Colloquium Mathematicum 132.1 (2013), pp. 121–138. DOI: 10.4064/cm132-1-10.
  16. [16]R. M. Łochowski. “On Pathwise Stochastic Integration with Respect to Semimartingales”. In: Probability and Mathematical Statistics 34.1 (2014), pp. 23–43. arXiv: 1211.3868.
  17. [17]R. Ghomrasni and R. M. Łochowski. “The Play Operator, the Truncated Variation and the Generalisation of the Jordan Decomposition”. In: Mathematical Methods in the Applied Sciences 38.3 (2015), pp. 403–419. DOI: 10.1002/mma.3077.
  18. [18]P. Das et al. “Level Crossings of Fractional Brownian Motion”. In: The Annals of Probability 53.6 (2025), pp. 2354–2393. DOI: 10.1214/25-AOP1764.DOI
  19. [19]P. Carmona and L. Coutin. “Fractional Brownian Motion and the Markov Property”. In: Electronic Communications in Probability 3 (1998), pp. 95–107. DOI: 10.1214/ECP.v3-998.
  20. [20]P. Carmona, L. Coutin, and G. Montseny. “Approximation of Some Gaussian Processes”. In: Statistical Inference for Stochastic Processes 3.1–2 (2000), pp. 161–171. DOI: 10.1023/A:100999518898.
  21. [21]M. Hairer. “Ergodicity of Stochastic Differential Equations Driven by Fractional Brownian Motion”. In: The Annals of Probability 33.2 (2005), pp. 703–758. DOI: 10.1214/009117904000000892.
  22. [22]F. Viens and J. Zhang. “A Martingale Approach for Fractional Brownian Motions and Related Path Dependent PDEs”. In: The Annals of Applied Probability 29.6 (2019), pp. 3489–3540. DOI: 10.1214/19-AAP1486.
  23. [23]C. Cuchiero and J. Teichmann. “Generalized Feller Processes and Markovian Lifts of Stochastic Volterra Processes: The Affine Case”. In: Journal of Evolution Equations 20.4 (2020), pp. 1301–1348. DOI: 10.1007/s00028-020-00557-2.
  24. [24]K. Kirchner and J. Willems. “Multiple and Weak Markov Properties in Hilbert Spaces with Applications to Fractional Stochastic Evolution Equations”. In: Stochastic Processes and their Applications 186 (2025), p. 104639. DOI: 10.1016/j.spa.2025.104639.
  25. [25]E. Abi Jaber and O. El Euch. “Multifactor Approximation of Rough Volatility Models”. In: SIAM Journal on Financial Mathematics 10.2 (2019), pp. 309–349. DOI: 10.1137/18M1170236.
  26. [26]P. Harms. “Strong Convergence Rates for Markovian Representations of Fractional Processes”. In: Discrete and Continuous Dynamical Systems - B 26.10 (2021), pp. 5567–5579. DOI: 10.3934/dcdsb.2020367.
  27. [27]C. Bayer and S. Breneis. “Markovian Approximations of Stochastic Volterra Equations with the Fractional Kernel”. In: Quantitative Finance 23.1 (2023), pp. 53–70. DOI: 10.1080/14697688.2022.2139193.
  28. [28]A. Alfonsi and A. Kebaier. “Approximation of Stochastic Volterra Equations with Kernels of Completely Monotone Type”. In: Mathematics of Computation 93.346 (2024), pp. 643–677. DOI: 10.1090/mcom/3911.
  29. [29]P. Guasoni. “No Arbitrage under Transaction Costs, with Fractional Brownian Motion and Beyond”. In: Mathematical Finance 16.3 (2006), pp. 569–582. DOI: 10.1111/j.1467-9965.2006.00283.x.
  30. [30]P. E. Protter. Stochastic Integration and Differential Equations. 2nd ed. Vol. 21. Stochastic Modelling and Applied Probability. Berlin: Springer, 2005. DOI: 10.1007/978-3-662-10061-5.DOI
  31. [31]L. Decreusefond and A. S. Üstünel. “Stochastic Analysis of the Fractional Brownian Motion”. In: Potential Analysis 10.2 (1999), pp. 177–214. DOI: 10.1023/A:1008634027843.
  32. [32]O. Kallenberg. Foundations of Modern Probability. 3rd ed. Cham: Springer, 2021. DOI: 10.1007/978-3-030-61871-1.DOI
  33. [33]B. B. Mandelbrot and J. W. Van Ness. “Fractional Brownian Motions, Fractional Noises and Applications”. In: SIAM Review 10.4 (1968), pp. 422–437. DOI: 10.1137/1010093.
  34. [34]V. Pipiras and M. S. Taqqu. “Deconvolution of Fractional Brownian Motion”. In: Journal of Time Series Analysis 23.4 (2002), pp. 487–501. DOI: 10.1111/1467-9892.00274.DOI
  35. [35]P. Billingsley. Convergence of Probability Measures. 2nd ed. Wiley Series in Probability and Statistics. John Wiley & Sons, 1999. DOI: 10.1002/9780470316962.DOI

Paper details

vlnc:W_bspf5u5k
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents