Repository navigation
feat(PhyslibAlpha): formalize the Hohmann transfer and its two-impulse optimality - #2077
hkjtsgmc79-boop wants to merge 4 commits into
Conversation
Co-authored-by: Codex <codex@openai.com>
Co-authored-by: Codex <codex@openai.com>
|
Thank you for this pull-request (PR). If this is your first PR, welcome to the community! While you wait for a review, note that you can review PRs too; reviewing is not restricted to maintainers. Our review guidelines explain what to look for. Over time, contributions such as reviews, maintenance and documentation are recorded on contributions.physlib.io. Three main points:
|
Assisted-by: Codex <codex@openai.com>
Assisted-by: Codex <codex@openai.com>
Summary
This PR adds to
PhyslibAlphaa formalization of the Hohmann transfer between coplanar, co-oriented circular orbits and of its global optimality in the two-impulse comparison class. It includes the explicit transfer ellipse and its two circular coasts, the closed-form burn magnitudes and half-period, a global lower bound over all non-colliding Newtonian two-impulse coasts, attainment by the Hohmann coast, and a concrete three-impulse bi-elliptic counterexample showing why the two-impulse restriction is essential.The mathematical statements are classical (Hohmann; Prussing); the contribution is the Lean formalization and its API.
Main results
exists_hohmann_transferandhohmann_transferconstruct the outward transfer for arbitrary positivemuandr₁ < r₂. The statements expose actualHasDerivAt/derivcertificates, the half-periodτ = π * sqrt (a³ / mu)witha = (r₁ + r₂) / 2, the two tangential velocity jumps, the closed-formΔv₁andΔv₂, and the constant specific energy and angular momentum of the coast.exists_hohmann_transferdoes not use a project-defined orbit predicate: all geometric and motion properties are conclusions.global_two_impulse_endpoint_optimalityproves the endpoint inequality and its equality case for all feasible endpoint invariants.global_two_impulse_newton_lower_boundlifts this to every non-colliding Newtonian coast whose endpoints have radiir₁andr₂, with free departure phase and flight time.h1_global_two_impulse_minimumandh1_attains_bound_in_kepler_classshow that the explicit Hohmann cost is at most every competitor in that class, and that the Hohmann half-ellipse is itself packaged as aKeplerArcattaining the bound.finite_three_impulse_counterexampleandBiellipticWitness.physical_three_impulse_counterexample: for normalized radii1 → 64 → 16, a three-impulse bi-elliptic transfer is cheaper than the direct Hohmann transfer by more than9/1000of the inner circular speed. The physical witness uses two actual finite-time Newtonian coasts and three velocity jumps, not only an algebraic cost comparison.Fuelconnects theΔvbound to the ideal rocket equation at fixed positive effective exhaust speed: the propellant fraction is1 - exp (-Δv / vₑ), andnewton_arc_ideal_fuel_lower_boundtransfers the two-impulse bound to this model.Comparison class and scope
The optimality theorem is for outward (
r₁ < r₂), coplanar, prograde/co-oriented circular terminal orbits, two impulses, free phase and flight time, in a single inverse-square field. The competitor coast may have radial velocity, negative angular momentum, and any energy sign; no tangency, ellipse, energy-sign, or positive-angular-momentum assumption is made. A third burn is not allowed, and no optimality over all impulse counts is claimed — the bi-elliptic counterexample lies outside the two-impulse class. The inward case is not formalized. No stability, time-fixed, thrust-limited, or plane-change claims are made.Proof structure
The proof normalizes by the inner radius and circular speed, proves the endpoint inequality over a relaxation of the energy/angular-momentum constraints, and characterizes equality: zero radial velocity together with the prograde tangential Hohmann values. Conservation of energy and angular momentum for an actual Newtonian coast is derived from the ODE in
KeplerConservation, not assumed; physical units are restored inPhysicalUnits, and the explicit Hohmann arc is bridged back inH1Bridge.On the dynamics side, the development proves the Kepler-equation/eccentric-anomaly parameterization, the half-period, the tangential impulses, and the closed forms. The endpoint equality characterization is at the level of endpoint velocities; no uniqueness of flight time or number of revolutions is claimed.
Verification
Checked with Lean 4.35.0 and mathlib
56a005855ea17e0b0958f73ec801ae140b0a1679:propext,Classical.choice, andQuot.sound;sorry,axiom,native_decide,unsafe, oropaqueoccurs in the new files.The full local build, required lint checks, and axiom audit passed. GitHub CI is pending.
References
The global optimality statement is the one proved by Prussing, Simple proof of the global optimality of the Hohmann transfer, Journal of Guidance, Control, and Dynamics 15(4) (1992), 1037–1038, DOI
10.2514/3.20941. The Lean proof is self-contained and does not claim to reproduce the paper’s proof steps.The Hohmann geometry and half-period follow the standard treatment in JPL, Fundamentals of Orbital Mechanics, chapter 7, §7.5.2.1. Weber’s Bi-elliptic Hohmann Transfer supplies the bi-elliptic geometry and finite-radius comparison background; the exact
1 → 64 → 16witness is the Lean computation in this PR. NASA Glenn’s Ideal Rocket Equation supplies the propellant-fraction form used inFuel.AI assistance and authorship
The implementation was produced with AI assistance; Codex is recorded as a co-author on the commit. The human author is responsible for the statements, proofs, and references, and has reviewed the formalization. The work was produced at Chengyuan AI Lab.
Review notes
This is a large PR (23 new modules, ~3k lines). It is one coherent development, but I am happy to split it if reviewers prefer, for example into the Hohmann dynamics, the two-impulse optimality proof, and the bi-elliptic/fuel consequences. This draft is for early feedback.