Paper: arXiv 2410.05524

Authors: Ashley Davey, Harry Zheng

Abstract

We consider the portfolio optimisation problem where the terminal function is an S-shaped utility applied at the difference between the wealth and a random benchmark process. We develop several numerical methods for solving the problem using deep learning and duality methods. We use deep learning methods to solve the associated Hamilton-Jacobi-Bellman equation for both the primal and dual problems, and the adjoint equation arising from the stochastic maximum principle. We compare the solution of this non-concave problem to that of concavified utility, a random function depending on the benchmark, in both complete and incomplete markets. We give some numerical results for power and log utilities to show the accuracy of the suggested algorithms.

Complexity vs Empirical Score

  • Math Complexity: 9.5/10
  • Empirical Rigor: 3.0/10
  • Quadrant: Lab Rats — high math complexity, limited empirical testing

Why this score: The paper is highly theoretical, focusing on advanced mathematical techniques like stochastic maximum principle, Malliavin calculus, concavification, and deep learning PDE solvers, which is dense and advanced. Empirical rigor is low as it only presents numerical results for algorithm accuracy with power/log utilities, lacking backtesting on real market data or trading strategy implementation.

Research Flowchart

  flowchart TD
  A["Research Goal:<br>Solve S-shaped Utility Maximization<br>with Random Reference Point"] --> B
  B["Key Inputs:<br>Portfolio Constraints, <br>Random Benchmark Process, <br>S-shaped Utility Function"] --> C
  subgraph C ["Methodology & Computation"]
      direction TB
      C1["Deep Learning: <br>HJB Equation (Primal/Dual)"]
      C2["Stochastic Maximum Principle: <br>Adjoint Equation"]
      C3["Concavified Utility: <br>Comparison Baseline"]
  end
  C --> D["Computational Execution:<br>ML & Numerical Algorithms<br>in Complete/Incomplete Markets"]
  D --> E["Key Outcomes:<br>1. Accurate Solutions via DL<br>2. Comparison to Concave Methods<br>3. Numerical Results (Power/Log Utilities)"]

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 several deep learning based numerical methods, including Physics Informed Neural Networks and deep BSDE/SMP solvers, to solve the S‑shaped utility maximisation problem with a random benchmark in both complete and incomplete markets.

Source sentence
We develop several numerical methods for solving the problem using deep learning and duality methods. We use deep learning methods to solve the associated Hamilton-Jacobi-Bellman equation for both the primal and dual problems, and the adjoint equation arising from the stochastic maximum principle.

Assumes 6

  • The investment horizon is a fixed interval [0,T] with a risk‑free asset following dS0_t = αS0_t dt and a risky asset following dS_t = μS_t dt + σS_t dW_t. model
    Source sentence
    PROBLEM FORMULATION Assume (Ω,F,{Ft},P) is a filtered probability space with {Ft} the natural filtration gen-erated by two standard Brownian motions W,W R with correlation ρ ∈[−1,1], augmented with all P-null sets. Assume the investment period is [0,T] and the risk-free asset price S0 and the risky asset price S satisfy the following stochastic differential equations (SDEs), for t ∈[0,T], dS0 t = αS0 t dt, dSt =
  • The two Brownian motions driving the assets have correlation ρ ∈ [−1,1]. model
    Source sentence
    PROBLEM FORMULATION Assume (Ω,F,{Ft},P) is a filtered probability space with {Ft} the natural filtration gen-erated by two standard Brownian motions W,W R with correlation ρ ∈[−1,1], augmented with all P-null sets.
  • The benchmark is a non‑negative random variable (or process) that may be non‑replicable in incomplete markets. other
    Source sentence
    INTRODUCTION In this paper we apply the deep learning method to solve utility maximisation problem with a benchmark that is a non-negative random variable.
  • For the SMP formulation a complete market is assumed so that the random reference R is replicable. market
    Source sentence
    To apply the SMP as it is given in the literature, for example [21], we assume a complete market here, so RT is replicable.
Show 2 more
  • The control variable π (or Π) is admissible, i.e., the resulting wealth process remains non‑negative where required. model
    Source sentence
    We therefore require Π to be in the admissible control set, defined by, for (t,x) ∈[0,T]×[0,∞), A (t,x) =  Π ∈H 2 t (R) Xt,x,Π s ≥0 a.s. for all s ∈[t,T] , where Xt,x,Π satisfies (4.3) with Xt,x,Π t = x.
  • The utility function is S‑shaped: concave on (0,∞) and convex on (−∞,0). model
    Source sentence
    The agent wishes to maximise an S-shaped utility, a function which is concave on (0,∞) and convex on (∞,0), applied at the difference between their terminal wealth and the benchmark.

Data

  • Synthetic grid data for PINN training synthetic unknown
    Source sentence
    We choose a grid size X ..= (Xmin,Xmax)d ∈(0,∞)d, then uniformly sample three batches: the collocation data within the time horizon, given by (t j,Xj) ∈[0,T)×X for j = 1,...,Mc, the boundary data (T,Xb j ) ∈{T} × X for j = 1,...,Mb and the zero state data (ts j,Xs j) ∈ S SHAPED UTILITY MAXIMISATION WITH A RANDOM REFERENCE POINT 9 [0,T]×∂(Xmin,Xmax)d for j = 1,...,Ms. Using these sample data we define the loss

Costs

Not applicable The paper reports no trading or execution result.

Finds 5

  • Numerical experiments for power and log utilities demonstrate that the deep learning algorithms achieve high accuracy compared to reference solutions. performance
    Source sentence
    We give some numerical results for power and log utilities to show the accuracy of the suggested algorithms. Results show the PINN-based algorithms that reduce dimensionality of the problem, solving the scaled problem with z = x r, then rescaling.
  • The PINN‑based methods reduce dimensionality and successfully solve the scaled problem, with results matching those of the dual and primal formulations. performance
    Source sentence
    Results show the PINN-based algorithms that reduce dimensionality of the problem, solving the scaled problem with z = x r, then rescaling. (x, r) Primal General Primal Non-Conc General Dual General SMP Solution (0.5, 1.0) -0.0188 -0.0577 -0.0050 -0.0450 -0.0149 (1.0, 1.0) 0.3793 0.3340 0.3866 0.3735 0.3872 (5.0, 1.0) 2.0546 2.0631 2.0787 2.0678 2.0653 (1.0, 0.5) 0.7335 0.7429 0.7320 0.7361 0.7353 (1.0, 5.0) -0.6758
  • In incomplete markets, the primal and dual value approximations agree, suggesting that the concavification principle still holds despite lack of analytical solutions. theoretical
    Source sentence
    However, the concave primal, and dual PINN methods appear to agree, suggesting the concavification principle still occurs. As the primal and dual val-ues coincide, up to numerical errors, we conclude that the payoff XT = −∂y ˜U(YT,RT) has been replicated by the primal algorithm, where YT is the optimal payoff for the dual problem.
  • The SMP method computes solutions in about one minute and produces consistent average results over multiple runs. performance
    Source sentence
    The SMP method takes about 1 minute. We run the algorithms 10 times and present the average results.
  • Overall, the deep neural network approach is shown to be viable and robust for solving nonlinear HJB equations and adjoint BSDEs in this non‑concave utility setting. theoretical
    Source sentence
    These numerical results show robustness of our methods outside of problem classes that are well understood. FIGURE 2. We present numerical results for different solution methods (HJB equations for primal, concavified primal, dual value functions, SMP for adjoint BSDE, reduced and full problems) and show that the DNN method is viable and robust in solving nonlinear PDEs and BSDEs.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 33 related papers). Coverage is partial, so this is neither replication nor refutation.

Cited by 1 paper in the archive (1 by independent authors): Portfolio selection with exogenous and endogenous … (uses its method or data)

Code

https://github.com/Ashley-Davey/ML-for-S-Utility — the authors' own repository

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.

Cited References (21)

  1. Deep Learning for Constrained Utility Maximisation (2020) — link
  2. Optimal investment with S-shaped utility and trading and Value at Risk constraints: An application to defined contribution pension plan (2020) — link
  3. Speculative trading, prospect theory and transaction costs (2019) — link
  4. Non-Concave Utility Maximization without the Concavification Principle (2019) — link
  5. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations (2019) — link
  6. Utility Maximization Under Trading Constraints with Discontinuous Utility (2019) — link
  7. A proof that artificial neural networks overcome the curse of dimensionality in the numerical approximation of Black-Scholes partial differential equations (2018) — link
  8. Solving high-dimensional partial differential equations using deep learning (2017) — link
and 13 more
  1. Deep Learning Approximation for Stochastic Control Problems (2016) — link
  2. On the dual problem of utility maximization in incomplete markets (2015) — link
  3. Utility maximization with a given pricing measure when the utility is not necessarily concave (2013) — link
  4. A mean-field stochastic maximum principle via Malliavin calculus (2012) — link
  5. Continuous-time stochastic control and optimization with financial applications / Huyen Pham (2009) — link
  6. Utility maximization in incomplete markets with random endowment (2001) — link
  7. Does Option Compensation Increase Managerial Risk Appetite (2000) — link
  8. The asymptotic elasticity of utility functions and optimal investment in incomplete markets (1999) — link
  9. Stochastic Controls: Hamiltonian Systems and HJB Equations (1999)
  10. Original Contribution: Multilayer feedforward networks with a nonpolynomial activation function can approximate any function (1993) — link
  11. Convex Duality in Constrained Portfolio Optimization (1992) — link
  12. Martingale and duality methods for utility maximization in a incomplete market (1991) — link
  13. Portfolio optimization for piecewise concave criteria functions (2010)

0 of 21 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.