The Lie symmetry algebra of the Longstaff-Schwartz model

Basic Information

• Paper ID: 2501.00155
• Title: The Lie symmetry algebra of the Longstaff-Schwartz model
• Author: OUKNINE Anas (Univ Rouen Normandie, CNRS, Normandie Univ, LMRS UMR 6085, F-76000 Rouen France)
• Classification: math.RA math.AP
• Publication Date: January 3, 2025
• Paper Link: https://arxiv.org/abs/2501.00155

Abstract

This study applies Lie group theory to compute the symmetry group of a class of partial differential equations parameterized by four constants: $u_t = -((a-bx)u_x + (d-ey)u_y + x^2u_{xx} + y^2u_{yy})$. Under various conditions on the constants $a, b, d, e$, the author derives the maximal and minimal Lie algebra symmetries and determines the structures of these algebras.

Research Background and Motivation

Problem Background

1. Financial Mathematics Context: The Longstaff-Schwartz model is a financial pricing model developed in 1992 by Francis Longstaff and Eduardo Schwartz for valuing American options. The model is based on a two-factor Cox-Ingersoll-Ross (CIR) process: $dX_t = (a-bX_t)dt + \sqrt{X_t}dB_t^1$$dY_t = (d-eY_t)dt + \sqrt{Y_t}dB_t^2$
2. Mathematical Problem: Investigation of Lie symmetries of the corresponding Kolmogorov backward equation: $u_t = -Lu$ where $Lu = (a-bx)u_x + (d-ey)u_y + x^2u_{xx} + y^2u_{yy}$

Research Motivation

1. Software Tool Limitations: Existing software such as Maple, Mathematica, and Python cannot distinguish between the four constant parameters in the drift function or impose necessary conditions to identify the symmetry group
2. Theoretical Completeness: Systematic determination of all symmetry groups under all parameter conditions and comparison with results from software such as Maple
3. Application Value: Lie symmetries are important for solving PDEs, reducing order, and finding exact solutions

Core Contributions

1. Complete Classification: Systematic computation of all possible Lie symmetry algebras for the PDE corresponding to the Longstaff-Schwartz model under 16 different combinations of parameters $(a,b,d,e)$
2. Structural Analysis: Determination of the specific structure of Lie algebras in various cases, including isomorphism relations and dimensions
3. Explicit Symmetries: Provision of infinitesimal generators and corresponding one-parameter group transformations for all cases
4. Theoretical Framework: Establishment of a systematic theoretical framework from finite-dimensional symmetry algebras $L_{a,b,d,e}$ to complete symmetry algebras $V_{a,b,d,e} = L_{a,b,d,e} \ltimes H$

Methodology Details

Problem Formulation

Investigation of Lie point symmetries of the partial differential equation $\Delta(x,y,t,u^{(2)}) = (a-bx)u_x + (d-ey)u_y + x^2u_{xx} + y^2u_{yy} + u_t = 0$, where $(x,y,t)$ are independent variables and $u$ is the dependent variable.

Theoretical Foundation

Based on Olver's Lie group theory method, seeking vector fields of the form $v = \xi(x,y,t,u)\frac{\partial}{\partial x} + \gamma(x,y,t,u)\frac{\partial}{\partial y} + \tau(x,y,t,u)\frac{\partial}{\partial t} + \phi(x,y,t,u)\frac{\partial}{\partial u}$ such that the corresponding one-parameter group $\exp(\epsilon v)$ is a symmetry group of the equation.

Key Technical Steps

1. Prolongation Computation: Calculation of the second-order prolongation $pr^{(2)}v$ of the vector field $v$
2. Application of Lie Conditions: According to Lie's theorem, $v$ generates a symmetry group if and only if $pr^{(2)}v[\Delta] = 0$ holds when $\Delta = 0$
3. Determining System Solution: This leads to a polynomial system of equations in $\xi, \gamma, \tau, \phi$ and their partial derivatives
4. Parameter Classification: Systematic classification according to different values of constants $a,b,d,e$

Principal Findings

Through solving the determining system, it is found that:

• $\tau$ depends only on $t$
• $\xi, \gamma$ are independent of $u$
• $\phi$ is affine in $u$: $\phi(x,y,t,u) = \lambda(x,y,t)u + \mu(x,y,t)$
• Both $\lambda$ and $\mu$ satisfy the original PDE

Experimental Setup

Classification Scheme

The problem is divided into 16 subcases based on parameter conditions:

• Conditions on $b$: $(b = e, b \neq 0)$, $(b = -e, b \neq 0)$, $(b = e = 0)$, $(b \neq \pm e)$
• Conditions on $a,d$: $(a = \frac{1}{4}, d = \frac{1}{4})$, $(a = \frac{1}{4}, d \neq \frac{1}{4})$, $(a \neq \frac{1}{4}, d = \frac{1}{4})$, $(a \neq \frac{1}{4}, d \neq \frac{1}{4})$

Computational Method

1. For each case, solve the determining system equations (3.39a)-(3.39f)
2. Compute the corresponding infinitesimal generators
3. Construct Lie bracket tables and analyze algebraic structure
4. Determine one-parameter group transformations and solution transformation formulas

Experimental Results

Main Results

Dimension Analysis

• Maximal Lie algebra: Case 1.4 $(b = e = 0, a = d = \frac{1}{4})$, with 9 vector fields
• Minimal Lie algebra: Case 2.1 $(b \neq \pm e, a \neq \frac{1}{4}, d \neq \frac{1}{4})$, with only 2 vector fields

Structural Classification

Based on analysis, the structure of finite-dimensional Lie algebras $L_{a,b,d,e}$ is:

Specific Symmetries

Taking Case 1.4 as an example, the 9 fundamental symmetries are: $v_1 = \frac{\partial}{\partial t}, \quad v_2 = u\frac{\partial}{\partial u}$$v_3 = 2tx\frac{\partial}{\partial x} + 2ty\frac{\partial}{\partial y} + t^2\frac{\partial}{\partial t} + (2(x+y) - \frac{t}{2})u\frac{\partial}{\partial u}$ and so on.

Isomorphism Relations

Important isomorphism relations are discovered:

• $L_1^1 \cong L_2^1 \cong L_3^1 \cong L_1^2 \cong L_2^2 \cong L_3^2 \cong sl_2 \ltimes h_3$
• $L_0^1 \cong L_0^2 \cong iso_2$
• $L_1^0 \cong L_2^0 \cong L_3^0 \cong \mathbb{R} \times sl_2$

Related Work

Lie Symmetry Theory

1. Olver Method: Based on Olver's prolongation method for computing Lie symmetries
2. Equivariant Vector Method: Harrison and Estabrook's equivariant vector method as an alternative approach
3. Financial Applications: Research by Lescot and Zambrini on symmetries of Hamilton-Jacobi-Bellman equations and Black-Scholes equations

Heat Equation Comparison

The author uses the heat equation $u_{xx} = u_t$ as an example to demonstrate the method, which has a 6-dimensional Lie algebra: $v_1 = \frac{\partial}{\partial x}, \quad v_2 = \frac{\partial}{\partial t}, \quad v_3 = u\frac{\partial}{\partial u}$$v_4 = x\frac{\partial}{\partial x} + 2t\frac{\partial}{\partial t} - \frac{u}{2}\frac{\partial}{\partial u}$ and so on.

Conclusions and Discussion

Main Conclusions

1. Complete Classification: Successful complete classification of Lie symmetries for all 16 parameter cases of the Longstaff-Schwartz model
2. Structure Determination: Determination of the precise structure of Lie algebras in various cases, ranging from 2-dimensional Abelian algebras to 9-dimensional composite algebras
3. Application Value: Provision of powerful tools for solving corresponding PDEs and stochastic differential equations

Theoretical Significance

• Establishment of symmetry theory for important models in financial mathematics
• Provision of a systematic analysis method for similar multi-parameter PDE systems
• Connection between Lie group theory and financial mathematics

Practical Applications

Through symmetries, one can:

1. Construct new solutions: If $u(x,y,t)$ is a solution, then $u_{\epsilon,v_i}$ is also a solution
2. Reduce order for solving: Simplify equations using symmetries
3. Find invariant solutions: Seek special solutions invariant under symmetry transformations

In-Depth Evaluation

Strengths

1. Strong Systematicity: Complete analysis of all possible parameter combinations without omission
2. Theoretical Rigor: Strict adherence to standard procedures of Lie group theory with detailed and reliable calculations
3. Rich Results: Not only provides symmetries but also analyzes algebraic structures and isomorphism relations
4. Practical Value: Provides explicit group transformations and solution transformation formulas
5. Methodological Innovation: Overcomes limitations of existing software tools through manual derivation of complete results

Limitations

1. Computational Complexity: The calculation process for 16 cases is lengthy, with potential risk of computational errors
2. Limited Application Demonstration: Although theoretical results are provided, specific financial application cases are limited
3. Absence of Numerical Verification: Lack of numerical method verification of theoretical results
4. Insufficient Software Comparison: Although comparison with Maple is mentioned, detailed comparison results are not shown

Impact

1. Theoretical Contribution: Significant contribution to symmetry theory in financial mathematics
2. Methodological Value: Provides an exemplary approach for symmetry analysis of multi-parameter PDE systems
3. Application Prospects: Provides new mathematical tools for option pricing, risk management, and other financial problems

Applicable Scenarios

1. Financial Engineering: American option pricing, interest rate model analysis
2. Mathematical Physics: Solution of PDEs with similar structures
3. Theoretical Research: Applications of Lie groups in financial mathematics

References

The paper cites 41 related references, primarily including:

• Olver's classical works on applications of Lie groups
• Original financial model papers by Longstaff and Schwartz
• Research by Lescot and Zambrini on symmetries of financial equations
• Foundational literature on symmetry theory by Bluman and Kumei

Overall Assessment: This is a theoretically rigorous and computationally detailed mathematical paper that systematically addresses the Lie symmetry problem of the Longstaff-Schwartz model. Although the calculation process is complex, the results possess significant theoretical value and potential practical application prospects. The paper's main contribution lies in providing complete classification and structural analysis, laying a foundation for further research in related fields.