4.6.8
Intuition Badi picture (WHY yeh theorem exist karti hai)Jab tum y ′ = f ( x , y ) y ( x 0 ) = y 0 maan lete ho ki solution exist karti hai aur wo sirf ek hi hai. Lekin aisa kyun hona chahiye? Picard–Lindelöf entry ka ticket deta hai: agar f x 0 exists karti hai aur unique hoti hai.
Key idea yeh hai: IVP ka solution ek integral operator ke fixed point ke barabar hai. "Zyada wild nahi" (Lipschitz) us operator ko ek contraction banata hai (bas interval kaafi chhota ho), aur contractions ka exactly ek fixed point hota hai. Yahi theorem ki poori rooh hai.
Definition Initial Value Problem (IVP)Ek IVP yeh pair hai:
y ′ ( x ) = f ( x , y ) , y ( x 0 ) = y 0 . y ( x ) x 0
y f ( x , y ) R y L ≥ 0 Lipschitz constant hai) exist kare aisa ki
∣ f ( x , y 1 ) − f ( x , y 2 ) ∣ ≤ L ∣ y 1 − y 2 ∣ for all ( x , y 1 ) , ( x , y 2 ) ∈ R . Matlab: jab y f linear rate se zyada tez change nahi kar sakta. y L
Dhyan do ki dono hypotheses complementary roles play karti hain: continuity slope ko bound karti hai (M Lipschitz condition iteration ko converge karaati hai aur ek akela solution pin down karti hai. (Sirf continuity se bhi existence milti hai — yeh kamzor Peano theorem hai — lekin usme uniqueness kho sakti hai.)
Intuition Imaandar bookkeeping: proof mein har hypothesis asal mein kya karti haiEk common textbook slogan kehta hai "continuity ⟹ existence, Lipschitz ⟹ uniqueness." Yeh ek useful pehli picture hai, lekin Picard-iteration proof mein Lipschitz estimate dono ke liye use hoti hai — convergence (existence) aur uniqueness ke liye. "Sirf continuity ⟹ existence" wala clean split ek alag proof (Peano, Arzelà–Ascoli ke zariye) ka hai, Picard iteration ka nahi.
h M ( x 0 , y 0 ) M Δ x M Δ x b f
Solution ± M ∣ y ( x ) − y 0 ∣ ≤ M ∣ x − x 0 ∣. ∣ y − y 0 ∣ ≤ b M ∣ x − x 0 ∣ ≤ b ∣ x − x 0 ∣ ≤ b / M ∣ x − x 0 ∣ ≤ a ∣ x − x 0 ∣ ≤ min ( a , b / M ) = h Min kyun? Jo tighter constraint hai woh jeetega.
Intuition ODE ko integral equation mein badloDifferentiation fragile hai; integration robust hai. y ′ = f ( x , y ) x 0 x y ( x 0 ) = y 0 y ( x ) = y 0 + ∫ x 0 x f ( t , y ( t )) d t . ⟺ T [ y ] ( x ) = y 0 + ∫ x 0 x f ( t , y ( t )) d t badlati nahi . Hume ek fixed point T [ y ] = y
Picard iterates ki sequence banao:
y 0 ( x ) = y 0 , y n + 1 ( x ) = y 0 + ∫ x 0 x f ( t , y n ( t ) ) d t .
Step A — T ∣ x − x 0 ∣ ≤ h ∣ f ∣ ≤ M ∣ y n + 1 ( x ) − y 0 ∣ ≤ ∫ x 0 x ∣ f ( t , y n ) ∣ d t ≤ M h ≤ M ⋅ M b = b , R f (Yahaan M h ≤ b / M
Step B — Yeh converge kyun karta hai (Lipschitz kaam aata hai). Tab
∣ y n + 1 ( x ) − y n ( x ) ∣ ≤ ∫ x 0 x ∣ f ( t , y n ) − f ( t , y n − 1 ) ∣ d t ≤ L ∫ x 0 x ∣ y n − y n − 1 ∣ d t . ∣ y n + 1 ( x ) − y n ( x ) ∣ ≤ M ( n + 1 )! L n ∣ x − x 0 ∣ n + 1 , L M ( e L h − 1 ) < ∞ uniformly converge karte hain (Weierstrass M-test) ek limit y existence establish karta hai, aur dhyan do yeh pehle se Lipschitz use kar raha hai — Lipschitz sirf "uniqueness only" nahi hai.
T Kyun sahi lagta hai: estimate sup ∣ T [ u ] − T [ v ] ∣ ≤ L h sup ∣ u − v ∣ Fix: T K = L h < 1 interval ko h < 1/ L . h < 1/ L
h h ≤ α / L α < 1 K = L h < 1 T [ x 0 − h , x 0 + h ] h = min ( a , b / M ) Weighted norm / factorial bound (bina shrinking ke): upar wala bound use karo (ya equivalent Bielecki weighted sup-norm ∥ u ∥ = sup e − 2 L ∣ x − x 0 ∣ ∣ u ( x ) ∣ T puri interval ∣ x − x 0 ∣ ≤ h h L h ≥ 1
Uniqueness. Agar y y ~ ϕ ( x ) = ∣ y − y ~ ∣ ϕ ( x ) ≤ L ∫ x 0 x ϕ ( t ) d t , Grönwall's inequality ϕ ≡ 0 L h < 1
y ′ = x 2 + y 2 y ( 0 ) = 0 h a = 1 , b = 1 R = { ∣ x ∣ ≤ 1 , ∣ y ∣ ≤ 1 }
M R ∣ f ∣ = ∣ x 2 + y 2 ∣ ≤ 1 + 1 = 2 = M Kyun? Dono squares zyada se zyada 1 Lipschitz L ∂ f / ∂ y = 2 y ∣2 y ∣ ≤ 2 R L = 2 Yeh Lipschitz kyun deta hai? Convex set par bounded y L = sup ∣ ∂ y f ∣ Interval: h = min ( a , b / M ) = min ( 1 , 2 1 ) = 2 1 Contraction check: L h = 2 ⋅ 2 1 = 1 nahi < 1 ∣ x ∣ ≤ 2 1
Worked example 2. Uniqueness KAHAN fail kar sakti hai:y ′ = y 2/3 y ( 0 ) = 0
f = y 2/3 continuous hai ⟹ Peano existence deta hai. ✔Lekin ∂ f / ∂ y = 3 2 y − 1/3 → ∞ y → 0 y = 0 Lipschitz nahi .
Wakai do solutions exist karti hain: y ≡ 0 y = ( 3 x ) 3 Yeh kyun matter karta hai: Lipschitz hato, uniqueness jao — exactly wahi jo theorem warn karta hai.
Worked example 3. Ek clean linear case:y ′ = y y ( 0 ) = 1
f = y ∣ ∂ y f ∣ = 1 L = 1 R 2 Picard iterates: y 0 = 1 y 1 = 1 + ∫ 0 x 1 d t = 1 + x y 2 = 1 + ∫ 0 x ( 1 + t ) d t = 1 + x + 2 x 2
Kyun beautiful hai: iterates exactly e x y = e x L h < 1 h < 1 x
f Kyun sahi lagta hai: beginning calculus mein, continuous functions acchi tarah behave karti hain, to tum expect karte ho "ek aur sirf ek." Fix: continuity sirf existence deti hai (Peano). Example 2 (y 2/3 Lipschitz condition chahiye — extra control ki f y
Common mistake "Solution poore rectangle∣ x − x 0 ∣ ≤ a Kyun sahi lagta hai: tumne f R Fix: solution box se bahar nikal sakti hai x a h = min ( a , b / M ) ∣ y − y 0 ∣ ≤ b local hai.
x Kyun sahi lagta hai: x Fix: operator y ∣ f ( t , y 1 ) − f ( t , y 2 ) ∣ y x x
L h < 1 Kyun sahi lagta hai: slick Banach proof ko K = L h < 1 Fix: L h < 1 ek proof route ki convenience hai; tum hamesha h Stated theorem ko sirf continuity + Lipschitz chahiye; h = min ( a , b / M )
Recall Feynman: 12-saal ke bacche ko samjhaoSocho tum ek video-game car khel rahe ho jo zaroor ek jagah se start hoti hai, aur har point par ek rule (ek arrow) batata hai aage kaunsi taraf jaana hai. Existence kehti hai: tumhari jagah se ek road actually shuru hoti hai — tum kahin ja sakte ho. Uniqueness kehti hai: sirf ek HI road hai, tum suddenly do cars mein split nahi ho sakte. Catch yeh hai: arrows upar/neeche move karne par bahut achanak change nahi hone chahiye (yahi "Lipschitz" rule hai). Agar arrows crazy-steep ho jaate hain (jaise y 2/3
Mnemonic Roles yaad rakho"Continuous gets you Started, Lipschitz gets you Settled (converged + unique)."
Aur interval ke liye: "h m in hai, kyunki m ountain (M b ox dono tumhe box in karte hain" → h = min ( a , b / M ) L h < 1
Picard–Lindelöf ko f f M f y
Picard-iteration proof mein continuity kya deta hai aur Lipschitz kya deta hai? Continuity ∣ f ∣ ≤ M
Kya "continuity ⟹ existence, Lipschitz ⟹ uniqueness" Picard iteration ke liye exactly sahi hai? Nahi — Picard ke proof mein Lipschitz existence ke liye bhi use hota hai. Clean continuity-only existence ek alag proof (Peano) hai.
y ∣ f ( x , y 1 ) − f ( x , y 2 ) ∣ ≤ L ∣ y 1 − y 2 ∣ ( x , y i )
Solution kis interval par guaranteed hai? ∣ x − x 0 ∣ ≤ h h = min ( a , b / M )
Interval b / M Solution slope M ∣ y − y 0 ∣ ≤ M ∣ x − x 0 ∣ b ∣ x − x 0 ∣ ≤ b / M
Operator T Sirf jab K = L h < 1 h 1/ L h
IVP kis fixed-point problem ke equivalent hai? y = T [ y ] T [ y ] ( x ) = y 0 + ∫ x 0 x f ( t , y ( t )) d t
Picard iterates define karo. y 0 ( x ) = y 0 y n + 1 ( x ) = y 0 + ∫ x 0 x f ( t , y n ( t )) d t
Uniqueness clean tarike se prove karne ke liye L h < 1 Grönwall: ϕ ( x ) ≤ L ∫ x 0 x ϕ d t ϕ = ∣ y − y ~ ∣ ≥ 0 ϕ ≡ 0
Aisa IVP do jisme existence ho lekin uniqueness NAH ho, aur kyun. y ′ = y 2/3 , y ( 0 ) = 0 f y = 0 y ≡ 0 y = ( x /3 ) 3
Practice mein Lipschitz condition verify kaise karte hain? Dikhao ki ∂ f / ∂ y L = sup ∣ ∂ y f ∣
y ′ = x 2 + y 2 , y ( 0 ) = 0 ∣ x ∣ , ∣ y ∣ ≤ 1 M L h L h < 1 M = 2 L = 2 h = 1/2 L h = 1 < 1
Lipschitz condition — Picard ke proof mein convergence AUR uniqueness drive karta haiBanach fixed-point theorem (contraction mapping) — K = L h < 1 Bielecki weighted norm — kisi bhi h T Grönwall's inequality — L h < 1 Peano existence theorem — sirf continuity se existence (alag proof, uniqueness nahi)Picard iteration (method of successive approximations) Linear ODEs first order — global Lipschitz ⟹ global unique solutionsMaximal interval of existence — local solutions ko patch karna / continuation
alone gives existence via
