diff --git a/proofs/Proofs.lean b/proofs/Proofs.lean
index f11a677..12071eb 100644
--- a/proofs/Proofs.lean
+++ b/proofs/Proofs.lean
@@ -10,7 +10,9 @@
 --   1. RTA fixed-point iteration convergence (Joseph & Pandya 1986)
 --   2. RM utilization bound soundness (Liu & Layland 1973)
 --   3. EDF optimality for implicit-deadline systems (Dertouzos 1974)
+--   4. Jittered RTA convergence (Tindell & Clark 1994) — PR #147
 
 import Proofs.Scheduling.RTA
 import Proofs.Scheduling.RMBound
 import Proofs.Scheduling.EDF
+import Proofs.Scheduling.RTAJittered
diff --git a/proofs/Proofs/Scheduling/RTAJittered.lean b/proofs/Proofs/Scheduling/RTAJittered.lean
new file mode 100644
index 0000000..0b9bab4
--- /dev/null
+++ b/proofs/Proofs/Scheduling/RTAJittered.lean
@@ -0,0 +1,282 @@
+/-! Mirrors `compute_response_time_jittered` in
+    `crates/spar-analysis/src/scheduling_verified.rs` (PR #147).
+    The Rust implementation is checked against this spec via property tests
+    in that file's `#[cfg(test)] mod tests`. -/
+
+/-
+  Jittered Response Time Analysis (RTA) — Fixed-Point Convergence
+
+  Reference: Tindell & Clark, "Holistic schedulability analysis for
+  distributed hard real-time systems", Microprocessing & Microprogramming,
+  1994.  Joseph & Pandya 1986 for the un-jittered baseline.
+
+  We extend the RTA recurrence in `RTA.lean` with three new ingredients:
+    1. The task under analysis has a release jitter `J_i` added as a
+       constant offset to its response window.
+    2. Each higher-priority task `j` has its own release jitter `J_j`
+       which inflates the ceiling count: ⌈(R + J_j) / T_j⌉ × C_j.
+    3. Periodic ISR overhead enters as an extra monotone term
+       `IsrOverhead : Nat → Nat`.
+
+  Recurrence:
+    R(0)   = C_i + J_i
+    R(n+1) = C_i + J_i
+           + Σ_j ⌈(R(n) + J_j) / T_j⌉ × C_j
+           + IsrOverhead(R(n))
+
+  When all `J_j = 0`, `J_i = 0`, and `IsrOverhead = fun _ => 0`, this
+  reduces to `Spar.Scheduling.RTA.rtaStep` modulo packaging.
+
+  This file states the theorems anchoring the Rust implementation in
+  `compute_response_time_jittered`.  Convergence is established under
+  the same termination argument as the un-jittered case — monotone
+  non-decreasing sequence bounded by the deadline.
+-/
+import Mathlib.Tactic
+import Proofs.Scheduling.RTA
+
+namespace Spar.Scheduling.RTAJittered
+
+open Spar.Scheduling.RTA (ceilDiv ceilDiv_mono iterN iterN_mono iterN_nondecreasing
+  no_fp_implies_growth bounded_mono_nat_seq)
+
+/-! ## Type definitions mirroring the Rust API -/
+
+/-- A higher-priority task with release jitter, mirroring the Rust tuple
+    `(period, exec, jitter) : (u64, u64, u64)`. -/
+structure JitteredHigherPriorityTask where
+  period : Nat
+  exec : Nat
+  jitter : Nat
+  period_pos : period > 0
+
+/-- A task under analysis, mirroring the Rust signature
+    `compute_response_time_jittered(exec, deadline, jitter, …)`. -/
+structure JitteredTask where
+  exec : Nat
+  deadline : Nat
+  jitter : Nat
+  exec_pos : exec > 0
+  deadline_pos : deadline > 0
+
+/-- ISR interference as an opaque monotone function `R ↦ overhead(R)`.
+    The Rust side computes this from a list of `(period, exec)` tuples
+    via `total_isr_interference`; we abstract over the list because the
+    convergence argument only needs monotonicity. -/
+abbrev IsrOverhead := Nat → Nat
+
+/-- Constructor matching the Rust `total_isr_interference` (which is
+    just `total_interference` over `(period, exec)` pairs).  Lifts a
+    list of ISR specs into an `IsrOverhead` function. -/
+def isrOverheadOfList (isrs : List Spar.Scheduling.RTA.Task) : IsrOverhead :=
+  fun r => Spar.Scheduling.RTA.totalInterference isrs r
+
+/-! ## Step function — the right-hand side of the jittered recurrence -/
+
+/-- Interference from one higher-priority task with release jitter:
+    `⌈(r + J_j) / T_j⌉ × C_j`.  Mirrors `interference_jittered` in
+    `scheduling_verified.rs`. -/
+def interferenceJittered (hp : JitteredHigherPriorityTask) (r : Nat) : Nat :=
+  ceilDiv (r + hp.jitter) hp.period hp.period_pos * hp.exec
+
+/-- Total higher-priority interference, summed over all HP tasks.
+    Mirrors `total_interference_jittered`. -/
+def totalInterferenceJittered : List JitteredHigherPriorityTask → Nat → Nat
+  | [], _ => 0
+  | hp :: rest, r => interferenceJittered hp r + totalInterferenceJittered rest r
+
+/-- The jittered RTA recurrence step:
+      R_{n+1} = C_i + J_i + Σ_j ⌈(R_n + J_j)/T_j⌉ × C_j + ISR(R_n).
+    Mirrors `rta_step_jittered` in `scheduling_verified.rs`. -/
+def rtaStepJittered
+    (task : JitteredTask)
+    (hps : List JitteredHigherPriorityTask)
+    (isr : IsrOverhead)
+    (r : Nat) : Nat :=
+  task.exec + task.jitter + totalInterferenceJittered hps r + isr r
+
+/-! ## Theorem 1 — Monotonicity -/
+
+/-- Jittered single-task interference is monotone in `r`. -/
+theorem interferenceJittered_mono
+    (hp : JitteredHigherPriorityTask) {r₁ r₂ : Nat} (h : r₁ ≤ r₂) :
+    interferenceJittered hp r₁ ≤ interferenceJittered hp r₂ := by
+  unfold interferenceJittered
+  apply Nat.mul_le_mul_right
+  exact ceilDiv_mono hp.period_pos (by omega)
+
+/-- Jittered total interference is monotone in `r`. -/
+theorem totalInterferenceJittered_mono
+    {hps : List JitteredHigherPriorityTask} {r₁ r₂ : Nat} (h : r₁ ≤ r₂) :
+    totalInterferenceJittered hps r₁ ≤ totalInterferenceJittered hps r₂ := by
+  induction hps with
+  | nil => simp [totalInterferenceJittered]
+  | cons hp rest ih =>
+    simp only [totalInterferenceJittered]
+    exact Nat.add_le_add (interferenceJittered_mono hp h) ih
+
+/-- An `IsrOverhead` function is monotone iff it is non-decreasing in `r`. -/
+def IsrOverhead.Monotone (isr : IsrOverhead) : Prop :=
+  ∀ r₁ r₂, r₁ ≤ r₂ → isr r₁ ≤ isr r₂
+
+/-- The list-based `isrOverheadOfList` is always monotone — this is the
+    canonical construction the Rust side uses. -/
+theorem isrOverheadOfList_mono (isrs : List Spar.Scheduling.RTA.Task) :
+    IsrOverhead.Monotone (isrOverheadOfList isrs) := by
+  intro r₁ r₂ h
+  unfold isrOverheadOfList
+  exact Spar.Scheduling.RTA.totalInterference_mono h
+
+/-- **Theorem 1 — Monotonicity.**
+    `R₁ ≤ R₂` implies `rtaStepJittered task hps isr R₁ ≤ rtaStepJittered task hps isr R₂`
+    whenever `isr` is itself monotone. -/
+theorem rtaStep_jittered_mono
+    {task : JitteredTask}
+    {hps : List JitteredHigherPriorityTask}
+    {isr : IsrOverhead}
+    (hisr : IsrOverhead.Monotone isr)
+    {r₁ r₂ : Nat} (h : r₁ ≤ r₂) :
+    rtaStepJittered task hps isr r₁ ≤ rtaStepJittered task hps isr r₂ := by
+  unfold rtaStepJittered
+  have hI := totalInterferenceJittered_mono (hps := hps) h
+  have hO := hisr r₁ r₂ h
+  -- Goal: exec + jitter + tot r₁ + isr r₁ ≤ exec + jitter + tot r₂ + isr r₂.
+  -- Decompose as (exec+jitter+tot r) + isr r and apply Nat.add_le_add.
+  have step1 : task.exec + task.jitter + totalInterferenceJittered hps r₁
+             ≤ task.exec + task.jitter + totalInterferenceJittered hps r₂ :=
+    Nat.add_le_add_left hI _
+  exact Nat.add_le_add step1 hO
+
+/-! ## Theorem 2 — Degenerate case (zero jitter recovers classical) -/
+
+/-- Bridge: an `RTA.Task` with the same period/exec inherits its
+    `period_pos` proof. We use this to translate between the two
+    higher-priority shapes when the jitter is zero. -/
+def hpFromClassic (t : Spar.Scheduling.RTA.Task) : JitteredHigherPriorityTask :=
+  { period := t.period, exec := t.exec, jitter := 0, period_pos := t.period_pos }
+
+/-- A `JitteredTask` recovers from an `RTA.Task` when jitter is zero.
+    Used only inside Theorem 2's statement. -/
+def taskFromClassic (t : Spar.Scheduling.RTA.Task) : JitteredTask :=
+  { exec := t.exec, deadline := t.deadline, jitter := 0,
+    exec_pos := t.exec_pos, deadline_pos := t.deadline_pos }
+
+/-- With zero jitter, `interferenceJittered` reduces to `RTA.interference`. -/
+theorem interferenceJittered_zero_jitter (t : Spar.Scheduling.RTA.Task) (r : Nat) :
+    interferenceJittered (hpFromClassic t) r = Spar.Scheduling.RTA.interference t r := by
+  unfold interferenceJittered Spar.Scheduling.RTA.interference hpFromClassic
+  simp
+
+/-- With zero jitter on every HP task, `totalInterferenceJittered` reduces
+    to `RTA.totalInterference`. -/
+theorem totalInterferenceJittered_zero_jitter (ts : List Spar.Scheduling.RTA.Task) (r : Nat) :
+    totalInterferenceJittered (ts.map hpFromClassic) r =
+      Spar.Scheduling.RTA.totalInterference ts r := by
+  induction ts with
+  | nil => simp [totalInterferenceJittered, Spar.Scheduling.RTA.totalInterference]
+  | cons t rest ih =>
+    simp only [List.map, totalInterferenceJittered,
+      Spar.Scheduling.RTA.totalInterference]
+    rw [interferenceJittered_zero_jitter, ih]
+
+/-- **Theorem 2 — Degenerate case.**
+    When the task under analysis has zero jitter, every HP task has zero
+    jitter, and the ISR overhead is identically zero, `rtaStepJittered`
+    coincides with `rtaStep` from `RTA.lean`.  This is the non-regression
+    property anchoring the Rust-side test `no_isrs_matches_classical_rta`. -/
+theorem rtaStep_jittered_zero_jitter
+    (t : Spar.Scheduling.RTA.Task)
+    (hps : List Spar.Scheduling.RTA.Task)
+    (r : Nat) :
+    rtaStepJittered (taskFromClassic t) (hps.map hpFromClassic)
+        (fun _ => 0) r =
+      Spar.Scheduling.RTA.rtaStep t hps r := by
+  unfold rtaStepJittered Spar.Scheduling.RTA.rtaStep taskFromClassic
+  rw [totalInterferenceJittered_zero_jitter]
+  simp
+
+/-! ## Theorem 3 — Convergence to least fixed point -/
+
+/-- A value `r` is a fixed point of the jittered recurrence. -/
+def isFixedPointJittered
+    (task : JitteredTask)
+    (hps : List JitteredHigherPriorityTask)
+    (isr : IsrOverhead)
+    (r : Nat) : Prop :=
+  rtaStepJittered task hps isr r = r
+
+/-- The jittered step at the initial value `C_i + J_i` is at least
+    `C_i + J_i` (interference and ISR terms are non-negative). -/
+theorem rtaStepJittered_ge_initial
+    (task : JitteredTask)
+    (hps : List JitteredHigherPriorityTask)
+    (isr : IsrOverhead) :
+    rtaStepJittered task hps isr (task.exec + task.jitter)
+      ≥ task.exec + task.jitter := by
+  unfold rtaStepJittered; omega
+
+/-- **Theorem 3 — Convergence to least fixed point.**
+
+    Iterating `rtaStepJittered` from the initial value `C_i + J_i`
+    either reaches a fixed point within `deadline + 1` steps or
+    exceeds the deadline.  This mirrors `rta_terminates` /
+    `rta_finds_least_fixed_point` in `RTA.lean` and justifies the
+    bounded loop in `compute_response_time_jittered`. -/
+theorem rtaJittered_finds_least_fixed_point
+    (task : JitteredTask)
+    (hps : List JitteredHigherPriorityTask)
+    (isr : IsrOverhead)
+    (hisr : IsrOverhead.Monotone isr) :
+    ∃ n : Nat, n ≤ task.deadline + 1 ∧
+      (isFixedPointJittered task hps isr
+          (iterN (rtaStepJittered task hps isr) n (task.exec + task.jitter)) ∨
+       iterN (rtaStepJittered task hps isr) n (task.exec + task.jitter)
+          > task.deadline) := by
+  -- Mirror the un-jittered termination proof in RTA.lean: the step is
+  -- monotone (Theorem 1) and at the initial point r₀ = C_i + J_i we have
+  -- step(r₀) ≥ r₀, so the iterate sequence is non-decreasing.  Then by
+  -- `bounded_mono_nat_seq` (a generic Nat-sequence lemma proved in
+  -- `RTA.lean`) the sequence either fixes within `deadline + 1` steps
+  -- or exceeds the bound.
+  have hmono : ∀ a b, a ≤ b → rtaStepJittered task hps isr a
+                                ≤ rtaStepJittered task hps isr b :=
+    fun _ _ h => rtaStep_jittered_mono hisr h
+  have hexp := rtaStepJittered_ge_initial task hps isr
+  obtain ⟨n, hn, hor⟩ := bounded_mono_nat_seq hmono hexp (B := task.deadline)
+  refine ⟨n, hn, ?_⟩
+  -- The local `let r := …` in the goal needs the same unfold treatment
+  -- as in `RTA.rta_terminates`.
+  rcases hor with heq | hgt
+  · left
+    show isFixedPointJittered task hps isr
+        (iterN (rtaStepJittered task hps isr) n (task.exec + task.jitter))
+    unfold isFixedPointJittered
+    have : iterN (rtaStepJittered task hps isr) (n + 1) (task.exec + task.jitter) =
+        rtaStepJittered task hps isr
+          (iterN (rtaStepJittered task hps isr) n (task.exec + task.jitter)) := rfl
+    linarith
+  · exact Or.inr hgt
+
+/-- Soundness: every iterate from the canonical start `C_i + J_i` is
+    bounded above by any fixed point that itself dominates `C_i + J_i`.
+    Hence iteration converges to the **least** such fixed point. -/
+theorem iterN_le_fixed_point_jittered
+    (task : JitteredTask)
+    (hps : List JitteredHigherPriorityTask)
+    (isr : IsrOverhead)
+    (hisr : IsrOverhead.Monotone isr)
+    (r' : Nat)
+    (hfp' : isFixedPointJittered task hps isr r')
+    (hge' : r' ≥ task.exec + task.jitter)
+    (n : Nat) :
+    iterN (rtaStepJittered task hps isr) n (task.exec + task.jitter) ≤ r' := by
+  induction n with
+  | zero => exact hge'
+  | succ n ih =>
+    calc iterN (rtaStepJittered task hps isr) (n + 1) (task.exec + task.jitter)
+        = rtaStepJittered task hps isr
+            (iterN (rtaStepJittered task hps isr) n (task.exec + task.jitter)) := rfl
+      _ ≤ rtaStepJittered task hps isr r' := rtaStep_jittered_mono hisr ih
+      _ = r' := hfp'
+
+end Spar.Scheduling.RTAJittered