Zermelo Navigation Problem

• Zermelo Navigation Problem is a time-optimal control challenge that determines the shortest-time paths amid drift fields while respecting maximum speed constraints.
• It is modeled on Riemannian manifolds using Randers and Kropina Finsler metrics, facilitating the derivation of geodesics and optimal trajectories through methods like PMP and HJB.
• Its applications span ship routing, spacecraft re-entry, and quantum control, showcasing significant cross-disciplinary impact in physics and engineering.

The Zermelo Navigation Problem is a prototypical time-optimal control problem of steering a system in a medium subject to drift (or perturbation), seeking the shortest-time connection between prescribed endpoints, given constraints on the control input (“speed”). Its importance spans classical control, Finsler geometry, quantum dynamics, differential games, and engineering applications such as ship routing, spacecraft re-entry, and turbulent flow navigation.

1. Mathematical Formulation and Finsler-Geometric Structure

In the classical setting, the state is $x \in M$, often a Riemannian manifold $(M, h)$. The dynamics are

$\dot{x}(t) = u(t) + W(x(t)),$

where $u(t)$ is a control vector subject to $|u(t)|_h \leq u_{\max}(x) \leq 1$, and $W$ is the “wind” or drift field. The aim is to reach a target point in minimal time $T$, i.e., minimize

$T = \int_0^T dt,$

subject to the initial and terminal constraints.

Under the “mild wind” assumption, $|W(x)|_h < u_{\max}(x) \leq 1$ for all $x$, attainable velocities $(M, h)$0 at $(M, h)$1 fill a strongly convex body, and the time-optimal trajectories coincide with the geodesics of a Randers-type Finsler metric: $(M, h)$2 This formulation is both necessary and sufficient for strongly convex Randers metrics (Kopacz, 2016, Alves et al., 2023). The geodesics of $(M, h)$3 are solutions to the second-order ODE

$(M, h)$4

where $(M, h)$5 are Finsler spray coefficients derived from $(M, h)$6 (Fathia et al., 2023).

Generalizations: Kropina and Non-Convex Regimes

When $(M, h)$7, the strong drift regime, the Finsler functional degenerates to Kropina type: $(M, h)$8 defined on the conic domain where $(M, h)$9 (Yoshikawa et al., 2012, Kopacz, 2016). Geodesics and optimality properties survive in this setting under standard convexity and regularity assumptions.

If the control set $\dot{x}(t) = u(t) + W(x(t)),$0 is merely strictly convex but not necessarily ball-shaped (e.g., elliptic or asymmetric propulsion), Zermelo's equation extends to a non-parametric form dependent on the support function of $\dot{x}(t) = u(t) + W(x(t)),$1 (Rossa et al., 15 Oct 2025). For non-convex indicatrices (multi-convex or piecewise-smooth speed constraints), optimal paths can involve “tacking”—concatenations of segments in different patches, with discontinuous heading (Markvorsen et al., 10 Aug 2025).

2. Control and Hamiltonian Structures

The time-optimal problem is equivalent to minimizing the Finsler length between endpoints or solving the value function $\dot{x}(t) = u(t) + W(x(t)),$2 from the Hamilton–Jacobi–Bellman (HJB) equation: $\dot{x}(t) = u(t) + W(x(t)),$3 with $\dot{x}(t) = u(t) + W(x(t)),$4. The minimum is achieved when $\dot{x}(t) = u(t) + W(x(t)),$5 is chosen colinear with $\dot{x}(t) = u(t) + W(x(t)),$6 at maximal allowed norm.

Application of the Pontryagin Maximum Principle (PMP) yields coupled Hamiltonian equations,

$\dot{x}(t) = u(t) + W(x(t)),$7

with optimal $\dot{x}(t) = u(t) + W(x(t)),$8 (Kopacz, 2016, Rossa et al., 15 Oct 2025). For strictly convex $\dot{x}(t) = u(t) + W(x(t)),$9, optimal controls are analytic.

In the normal case, the trajectory is a regular time-minimizer; abnormal (or exceptional) controls—arising at the boundaries of controllability, e.g., near vortex singularities—lead to extremals that may separate time-minimal and time-maximal arcs and can instigate jump discontinuities in the value function (Bonnard et al., 2022, Bonnard et al., 2019).

3. Global Geometry, Singularities, and Extensions

Finsler Geodesic Viewpoint

Geodesics of Randers or Kropina metrics encode all optimal-control information for Zermelo-type problems with convex indicatrix (Kopacz, 2016, Fathia et al., 2023, Kopacz, 2016). When the underlying manifold has boundary or obstacles, the problem extends naturally to constrained geodesics.

Singularities: Vortex, Tacking, and Cut Locus

Non-smooth drift fields (e.g., point vortices, as in $u(t)$0) generate local “holes” and singular extremal structure in the cut locus: the ball around the base point can develop holes or cusps—topologically encoded by wavefront self-intersections or conjugate points—requiring micro-local analysis and homotopy techniques (Bonnard et al., 2019, Bonnard et al., 2022).

For multi-convex speed constraints, optimality enforces tacking: the trajectory alternates between segments on different convex patches, with explicit switching conditions derived from PMP (Markvorsen et al., 10 Aug 2025).

4. Computational Methods and Algorithms

Extensive algorithmic development targets both ODE-based and field-theoretic formulations:

Shooting, Hybrid, and Variational Smoothing

• Initial-Value/Shooting Methods: Integrate Zermelo's ODE above for the heading, with shooting-angle search to meet the terminal constraint. For spherical geometry, use great-circle bearings and adapt heading dynamics to the sphere (Precioso et al., 2023, Kopacz, 2016).
• Hybrid Search Algorithms: Combine systematic search within a heading-cone (exploration), recursive refinement, and piecewise smoothing via discrete variational principles (e.g., Ferraro–Mart de Diego–Sato method) to yield feasible and smooth paths, even in the presence of obstacles or spherical geometry (Precioso et al., 2023).
• Dynamic Programming and Quantum Annealing: Discrete formulations together with Landau–Zener analysis translate to efficient $u(t)$1 classical algorithms or adiabatic quantum methods, highlighting underlying structure in the combinatorial space of feasible trajectories (Selstø et al., 8 Jun 2026).

Multi-Fluid and HJB-Based PDE Schemes

• Fluid-Dynamics Mapping: The value function $u(t)$2 is computed on large domains by recasting HJB as a coupled multi-fluid system: the optimal path density and velocity are advanced alongside the background flow using standard CFD tools (e.g., OpenFOAM). The “Bellman fluid” encodes the cost-to-go and trajectory ensemble (Pratt et al., 20 Oct 2025).
• RL and Data-Driven Methods: Reinforcement learning (Q-learning, actor-critic) produces robust navigation policies even in time-dependent, turbulent, or incompletely known flows, outperforming deterministic optimal controls in realistic scenarios (Biferale et al., 2019, Parfenyev, 9 Dec 2025).

5. Quantum Zermelo Navigation: Unitary Evolution

The quantum Zermelo navigation problem seeks a time-optimal control Hamiltonian $u(t)$3 steering the system from $u(t)$4 to $u(t)$5 in the presence of a fixed background $u(t)$6: $u(t)$7 The optimal protocol is characterized by

$u(t)$8

yielding a propagator factorization

$u(t)$9

where $|u(t)|_h \leq u_{\max}(x) \leq 1$0 (Clausen, 2015, Brody et al., 2014). The time-optimal curves are geodesics for the Randers–Finsler structure on $|u(t)|_h \leq u_{\max}(x) \leq 1$1 induced by the background Hamiltonian.

If the system-bath interaction is to a harmonic oscillator bath with frequencies $|u(t)|_h \leq u_{\max}(x) \leq 1$2, there is a strict lower bound on control time, $|u(t)|_h \leq u_{\max}(x) \leq 1$3 (Clausen, 2015).

Arbitrary unitaries are reached by concatenating building blocks of the form $|u(t)|_h \leq u_{\max}(x) \leq 1$4; if the set $|u(t)|_h \leq u_{\max}(x) \leq 1$5 generates $|u(t)|_h \leq u_{\max}(x) \leq 1$6, the protocol is universal.

6. Connections to Physics and Advanced Geometric Structures

The Zermelo navigation problem informs the understanding of Randers and Kropina Finsler geometry, spacetime frame-dragging phenomena in general relativity, and the generation of new (static or rotating) metric solutions via the “Zermelo–Randers–spacetime triangle” (Li et al., 2024, Chanda, 2024). For example:

• In stationary spacetime, drift (frame dragging) corresponds exactly to the “wind” in Zermelo, and geodesics in the spatial Randers metric project to null geodesics of the full spacetime (Chanda, 2024).
• Choosing particular winds transforms flat space into Schwarzschild, Kerr, Rindler, or de Sitter metrics (and more), unifying navigation and spacetime construction (Li et al., 2024).

7. Applications, Computational Results, and Practical Impact

The Zermelo navigation framework underpins optimal ship routing, aircraft trajectory planning, spacecraft orbital insertion, sailboat/tacking protocols, and non-Markovian quantum gate synthesis.

Empirical findings include:

• HS algorithms yield travel-time savings up to $|u(t)|_h \leq u_{\max}(x) \leq 1$7 in idealized four-vortex test fields and up to $|u(t)|_h \leq u_{\max}(x) \leq 1$8 on real oceanic data, with moderate path-length increase and nontrivial fuel reductions (Precioso et al., 2023).
• RL policies in 2D turbulence consistently achieve high success (>99% in regular flows and >95% in moderate turbulence) and transfer robustly from coarse-grained (large-scale-only) fields to full turbulence (Parfenyev, 9 Dec 2025).
• In the quantum setting, the Zermelo protocol attains the quantum speed limit and constructs all time-optimal nonlocal unitaries mediated by a shared bath (Clausen, 2015).

Summary Table: Metric Types, Geometric Regimes, and Optimal Paths

Drift/Control Setting	Finsler Metric	Geodesics / Optimal Paths
Mild wind ($|u(t)|_h \leq u_{\max}(x) \leq 1$9)	Randers ($W$0)	Strongly regular, minimizing geodesics
Strong wind ($W$1)	Kropina ($W$2)	Conic Finsler, optimal paths on specific cone
Strictly convex control $W$3	Non-parametric	Smooth, C$W$4 controls, extended ODEs
Multi-convex/Non-convex $W$5	Multi-patch piecewise	Zig-zag (tacking) concatenations
Quantum/Background drift	Group Randers	Co-adjoint orbits in control Lie group

• (Clausen, 2015) Time-optimal bath-induced unitaries by Zermelo navigation
• (Kopacz, 2016) On generalization of Zermelo navigation problem on Riemannian manifolds
• (Bonnard et al., 2022) Abnormal Geodesics in 2D-Zermelo Navigation Problems...
• (Selstø et al., 8 Jun 2026) Zermelo's navigation problem through the lens of quantum annealing...
• (Bonnard et al., 2019) A Zermelo navigation problem with a vortex singularity
• (Fathia et al., 2023) On the geometry of Zermelo's optimal control trajectories
• (Biferale et al., 2019) Zermelo's problem: Optimal point-to-point navigation in 2D turbulent flows using Reinforcement Learning
• (Precioso et al., 2023) Hybrid Search method for Zermelo's navigation problem
• (Rossa et al., 15 Oct 2025) A non-parametric Zermelo navigation equation for strictly convex control sets
• (Chanda, 2024) Zermelo Wind: a geometrization of the frame dragging effect
• (Yoshikawa et al., 2012) Kropina metrics and Zermelo navigation on Riemannian manifolds
• (Kopacz, 2016) A note on generalization of Zermelo navigation problem on Riemannian manifolds with strong perturbation
• (Pratt et al., 20 Oct 2025) Optimal Control from a Fluid Dynamics Perspective
• (Li et al., 2024) Generating New Spacetimes through Zermelo Navigation
• (Aldea et al., 2016) Generalized Zermelo navigation on Hermitian manifolds under mild wind
• (Alves et al., 2023) Isoparametric functions and mean curvature in manifolds with Zermelo navigation
• (Kopacz, 2016) A note on time-optimal paths on perturbed spheroid
• (Markvorsen et al., 10 Aug 2025) Time-dependent Zermelo navigation with tacking
• (Parfenyev, 9 Dec 2025) Optimal navigation in two-dimensional regular and turbulent flows
• (Brody et al., 2014) Solution to the quantum Zermelo navigation problem

This body of literature establishes the Zermelo navigation problem as a unifying geometric and algorithmic framework for time-optimal steering under drift, yielding profound links between Finsler geometry, optimal control theory, relativity, and quantum information.