Paper: arXiv 2407.00266

Authors: Igor Cialenco, Gabriela Kováčová

Abstract

We study a dynamic stochastic control problem subject to Knightian uncertainty with multi-objective (vector-valued) criteria. Assuming the preferences across expected multi-loss vectors are represented by a given, yet general, preorder, we address the model uncertainty by adopting a robust or minimax perspective, minimizing expected loss across the worst-case model. For loss functions taking real (or scalar) values, there is no ambiguity in interpreting supremum and infimum. In contrast to the scalar case, major challenges for multi-loss control problems include properly defining and interpreting the notions of supremum and infimum, and addressing the non-uniqueness of these suprema and infima. To deal with these, we employ the notion of an ideal point vector-valued supremum for the robust part of the problem, while we view the control part as a multi-objective (or vector) optimization problem. Using a set-valued framework, we derive both a weak and strong version of the dynamic programming principle (DPP) or Bellman equations by taking the value function as the collection of all worst expected losses across all feasible actions. The weak version of Bellman’s principle is proved under minimal assumptions. To establish a stronger version of DPP, we introduce the rectangularity property with respect to a general preorder. We also further study a particular, but important, case of component-wise partial order of vectors, for which we additionally derive DPP under a different set-valued notion for the value function, the so-called upper image of the multi-objective problem. Finally, we provide illustrative examples motivated by financial problems. These results will serve as a foundation for addressing time-inconsistent problems subject to model uncertainty through the lens of a set-valued framework, as well as for studying multi-portfolio allocation problems under model uncertainty.

Complexity vs Empirical Score

  • Math Complexity: 9.2/10
  • Empirical Rigor: 2.1/10
  • Quadrant: Lab Rats — high math complexity, limited empirical testing

Why this score: The paper is highly theoretical, focusing on novel extensions of dynamic programming principles and set-valued supremum definitions under Knightian uncertainty for vector-valued criteria, with minimal computational or empirical implementation details.

Research Flowchart

  flowchart TD
  A["Research Goal<br>Multi-Objective Robust Control<br>Under Knightian Uncertainty"] --> B["Methodology Setup<br>Preorder on Vector Loss &<br>Minimax Worst-Case Perspective"]
  B --> C["Core Challenge<br>Define Sup/Inf for Vectors &<br>Handle Non-Uniqueness"]
  C --> D{"Solution Framework"}
  D --> E["General Case<br>Set-Valued Value Function"]
  D --> F["Specific Case<br>Component-Wise Partial Order"]
  E --> G["Computational Process<br>Dynamic Programming Principles (DPP)"]
  F --> G
  G --> H["Outcomes<br>Weak/Strong Bellman Equations<br>Foundations for Time-Inconsistent & Portfolio Problems"]

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 a vector‑valued robust stochastic control framework that defines a set‑valued value function using an ideal‑point vector supremum and derives both weak and strong versions of the dynamic programming principle for multi‑objective criteria under model uncertainty.

Source sentence
To deal with these, we employ the notion of an ideal point vector-valued supremum for the robust part of the problem, while we view the control part as a multi-objective (or vector) optimization problem. Using a set-valued framework, we derive both a weak and strong version of the dynamic programming principle (DPP) or Bellman equations by taking the value function as the collection of all worst expected losses

Assumes 5

  • Preferences over expected multi‑loss vectors are represented by a given preorder. other
    Source sentence
    Assuming the preferences across expected multi-loss vectors are represented by a given, yet general, preorder, we address the model uncertainty by adopting a robust or minimax perspective, minimizing expected loss across the worst-case model.
  • The weak version of Bellman’s principle holds under minimal assumptions. other
    Source sentence
    The weak version of Bellman's principle is proved under minimal assumptions.
  • A rectangularity property with respect to the preorder is assumed to obtain the strong DPP. model
    Source sentence
    To establish a stronger version of DPP, we introduce the rectangularity property with respect to a general preorder. The third key contribution of our study is the introduction of rectangularity property with respect to a general preorder.
  • The component‑wise partial order on vectors is considered as a special preorder case. other
    Source sentence
    We also further study a particular, but important, case of component-wise partial order of vectors, for which we additionally derive DPP under a different set-valued notion for the value function, the so-called upper image of the multi-objective problem. Fourth, we further study a particular, but important, case of component-wise partial order of vectors for which we additionally derive DPP under a second popular
Show 1 more
  • The set‑valued supremum (ideal‑point) exists and may be non‑unique, but the Bellman principle is invariant to the choice of supremum. other
    Source sentence
    Using a series of results, we show that while the vector-valued supremum may not be unique, the Bellman’s principle is invariant with respect to chosen supremum.

Data

No dataset named in the text.

Costs

Not applicable The paper reports no trading or execution result.

Finds 4

  • A weak dynamic programming principle is derived for the vector‑valued robust control problem without additional assumptions. theoretical
    Source sentence
    Using a set-valued framework, we derive both a weak and strong version of the dynamic programming principle (DPP) or Bellman equations by taking the value function as the collection of all worst expected losses across all feasible actions. The weak version of Bellman's principle is proved under minimal assumptions.
  • Introducing the rectangularity property yields a strong version of the DPP (Theorems 3.7 and 3.8). theoretical
    Source sentence
    To establish a stronger version of DPP, we introduce the rectangularity property with respect to a general preorder. Leveraging this we prove a stronger version of Bellman’s principle of optimality; see Theorem 3.7 and Theorem 3.8.
  • For the component‑wise partial order, the DPP can also be expressed via the upper‑image (upper set) of the multi‑objective value function. theoretical
    Source sentence
    We also further study a particular, but important, case of component-wise partial order of vectors, for which we additionally derive DPP under a different set-valued notion for the value function, the so-called upper image of the multi-objective problem. Fourth, we further study a particular, but important, case of component-wise partial order of vectors for which we additionally derive DPP under a second popular
  • Illustrative examples (including a binomial tree model) demonstrate the importance of rectangularity and the equivalence of weak and strong set‑valued Bellman principles under rectangular families of models. theoretical
    Source sentence
    Finally, we provide illustrative examples motivated by financial problems. Finally, we provide some illustrative examples.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 14 related papers). Coverage is partial, so this is neither replication nor refutation.

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.