Paper: arXiv 2407.17151
Authors: Aurélien Alfonsi, Edoardo Lombardo
Abstract
We present weak approximations schemes of any order for the Heston model that are obtained by using the method developed by Alfonsi and Bally (2021). This method consists in combining approximation schemes calculated on different random grids to increase the order of convergence. We apply this method with either the Ninomiya-Victoir scheme (2008) or a second-order scheme that samples exactly the volatility component, and we show rigorously that we can achieve then any order of convergence. We give numerical illustrations on financial examples that validate the theoretical order of convergence. We also present promising numerical results for the multifactor/rough Heston model and hint at applications to other models, including the Bates model and the double Heston model.
Complexity vs Empirical Score
- Math Complexity: 9.5/10
- Empirical Rigor: 4.0/10
- Quadrant: Lab Rats — high math complexity, limited empirical testing
Why this score: The paper is intensely theoretical, focusing on high-order weak approximations, random grid methods, and rigorous convergence proofs for the log-Heston SDE, demonstrating advanced stochastic calculus and functional analysis. However, its empirical validation is limited to numerical illustrations of theoretical convergence order, with no backtesting or real-world trading strategies presented.
Research Flowchart
flowchart TD
A["Research Goal"] --> B["Methodology: Combined Approximation Schemes"]
B --> C["Base Schemes: Ninomiya-Victoir &<br>Exact Volatility Sampling"]
C --> D["Simulation on Random Grids"]
D --> E["Weak Approximation of<br>Log-Heston Process"]
E --> F["Outcomes"]
subgraph F ["Key Findings"]
F1["Proved rigorous convergence<br>to any desired order"]
F2["Numerical validation of<br>theoretical convergence rates"]
F3["Extensions to Rough Heston,<br>Bates, and Double Heston models"]
end
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes weak approximation schemes of arbitrary order for the Heston model using random grid boosting combined with the Ninomiya‑Victoir scheme or an exact volatility sampler.
Source sentence
We present weak approximations schemes of any order for the Heston model that are obtained by using the method developed by Alfonsi and Bally (2021).
Assumes 4
- The Ninomiya‑Victoir scheme is applied only when σ² ≤ 4a. model
Source sentence
consider either the Ninomiya‑Victoir scheme for σ²
- Exact CIR simulation of the volatility component is available for any σ > 0. model
Source sentence
or the exact CIR simulation for any σ > 0
- The test functions f must belong to a high‑order Sobolev‑type space C^{12ν}_{pol}. statistical
Source sentence
∀f ∈C^{12ν}_{pol}(R × R+), ∃C > 0, L ∈N, ∥ˆP_{ν,n}f −P_T f∥_{0,L+3ν} ≤C∥f∥_{12ν,L} (T/n)^{2ν}
- Regularity assumption on f is strong but required for the analysis. other
Source sentence
The regularity assumption on f required by Theorem 2.1 is rather strong, but needed with our analysis.
Data
Synthetic data only.
Costs
Not applicable
Finds 4
- The random‑grid boosting method can achieve any prescribed order of weak convergence for the log‑Heston process. theoretical
Source sentence
we show rigorously that we can achieve then any order of convergence.
- Numerical experiments on European and Asian options confirm the theoretical orders (2 and 4) of convergence. performance
Source sentence
the estimated slopes are in line with the theoretical order of convergence (2 and 4)
- Using the “one‑step” coupling reduces the variance of the correction term compared to other couplings. statistical
Source sentence
The “one‑step” coupling is even much better. It gives a variance
- The scheme based on the Strang composition with Bernoulli random variable has higher variance and requires an extra random variable. statistical
Source sentence
the scheme using a Bernoulli random variable has a correcting term of higher variance. Besides, it requires to simulate one more random variable.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 31 related papers). Coverage is partial, so this is neither replication nor refutation.
3 papers in the archive
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.