Appendix IV, pp. 358–359 — optimum phase times of a two-phase fixed-time signal: minimum cycle if D > 0, (x₀, η) if D ≤ 0
ProvedWardropTraffic.Signal.optimum_phase_timesLet , , and . Let
on the feasible region , and put
Then:
- if , the minimum cycle minimizes over : for all ;
- if , the point lies in and minimizes over : for all .
Here and are the effective red times of the two phases as fractions of the cycle, and their largest values compatible with the flows, and the flow-weighted average delay (16). The theorem answers the question of Appendix IV: the shortest cycle (Adams' formula) minimizes the average delay exactly in the first case, and otherwise the optimum keeps phase 2 saturated and lengthens the cycle to .
Formalization Note The theorem is stated in the reduced parameters alone, which are all the argument uses; their expression through the flows is the separate theorem original_quantities. The hypothesis is the page's labelling of the phases. means ; is printed; means , without which the feasible region is empty. The final "if this condition is not satisfied" includes ; then , so the formula gives the minimum cycle.
import Mathlib import Definitions.Def_WardropTraffic_Signal_Setting
namespace WardropTraffic.Signal
theorem optimum_phase_times (lam mu xi eta : ℝ) (hlam : 0 < lam) (hmu : 0 < mu)
(hlm : mu ≤ lam) (hxi0 : 0 < xi) (hxi1 : xi < 1) (heta0 : 0 < eta) (heta1 : eta < 1)
(hsum : 1 < xi + eta) :
(0 < cornerD lam mu xi eta →
∀ z ∈ feasibleXY xi eta, delayT lam mu xi eta ≤ delayT lam mu z.1 z.2) ∧
(cornerD lam mu xi eta ≤ 0 →
(edgeRoot lam mu eta, eta) ∈ feasibleXY xi eta ∧
∀ z ∈ feasibleXY xi eta,
delayT lam mu (edgeRoot lam mu eta) eta ≤ delayT lam mu z.1 z.2) := by sorry
end WardropTraffic.Signal
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.