Scribd Logo
Upload

Welcome to Scribd!

  • Volterra y Fredholm

    Uploaded by

    Daniel Rodriguez Luis
    14 views13 pages
    Volterra

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document
    Volterra

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    14 views13 pages

    Volterra y Fredholm

    Uploaded by

    Daniel Rodriguez Luis
    Volterra

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    of 13
    INTRODUCTORY REAL ANALYSIS A, N. KOLMOGOROV S. V. FOMIN Revised English Edition Translated and Edited by Richard A. Silverman PRENTICE-HALL, INC, Englewood Cliffs, Nt. 66 METRIC spaces cuap. 2 Problem 6. Give an example of a complete me sequence {4,} of closed subsets of R such that, na, at Reconcile this example with Problem 4. space R and a nested 2. Problem 7. Prove that a subspace of a complete metric space R is com- plete if and only if it is closed. Problem 8. Prove that the real line equipped with the distance o(x, y) = lare tan x — are tan p| is an incomplete metric space. Problem 9. Give an example of a complete metric space homeomorphic to an incomplete metric space. Hint. Consider the example on p. 44. Comment. Thus homeomorphic metric spaces can have different “metric properties,” Problem 10. Carry out the program discussed in the last sentence of the example on p. 65. Hint. UW {x,} and {y,} are Cauchy sequences of rational numbers serving as “reptesentatives” of real numbers x* and y*, respectively, define x* + y* as the real number with representative {x, + y,} Contraction Mappings 8.1. Definition of a contraction mapping. The fixed point theorem. Let 4 be a mapping of a metric space R into itself. Then x is called a fixed point of 4 if Ax = x, ic., if A carries » into itself. Suppose there exists a number a <1 such that e(Ax, Ay) < aplx, y) 0) for every pair of points x, y © R Then A is said to be @ contraction mapping Every contraction mapping is automatically continuous, since it follows from the “contraction condition” (1) that Ax,,~> Ax whenever x,,— x. Tarorem 1 (Fixed point theorent!). Every contraction mapping A defined on a complete metric space R has a unique fixed point, ¥ Often called the method of successive approximations (sce the remark following ‘Theorem 1) of the principle of contraction mappings. sec. 8 CONTRACTION MAPPINGS 67 Proof. Given an arbitrary point x5 R, let! y= Ax, Xa= Ax = Ay Xn Aka = Ay 2) Then the sequence {x,} is fundamental, In fact, assuming to be explicit that n < n', we have P(X ps Xa) = OCA A” Xo) < A" Xoy Xan) < a"LeC%o0 1) + ans M2) + + Oe ea Xan] L 1). Then a | A has a unique fixed point we Proof. Choosing any point xp € R, let x =limA™ry, pars ‘Then, by the continuity of 4, A. lim AA*"xg. \ kee : 7 But A” is a contraction mapping, and hence (AM ANG, Pp) < 1p(AP MAN, AMM xg) << ++ «Sate AXs Xo) where # <1, It follows that e(Ax, x) = Tim p( A.A xg, AP'x9) kee ie., Ax = so that xis a fixed point of A. To prove the uniqueness of-x, SEC. & CONTRACTION MAPPINGS 71 we merely note that if 4 has more than one fixed point, then so does 4", which is impossible, by Theorem 1, since A" is a contraction mapping. § 8.2. Contraction mappings and differential equations. The most interesting applications of Theorems 1 and 1’ arise when the space R is a function space. We can then use these theorems to prove a number of existence and uniqueness theorems for differential and integral equations, as shown in this section and the next. ‘TutorEM 2 (Picard). Given a function f(x, y) defined and continuous on a plane domain G containing the point (2, yo)-* suppose f satisfies a Lipschitz condition of the form If.) —£@ Di < Mly— FI in the variable y. Then there is an interval |x — Xq| < 3 in which the differential equation (10) has a unique solution satisfying the initial condition P%0) = Yor ay Proof. Together the differential equation (10) and the initial condition (1) are equivalent to the integral equation a(x) = vot J Ft (0) at (12) By the continuity of f, we have YV@yl< K (13) in some domain G’ < G containing the point (x9, yo)-* Choose 3 > 0 such that 1) & eG if la 2) MB<1, — xol < 8, 1p — pol < KS; and let C* be the space of continuous functions 9 defined on the interval 18 By an n-dimensional domain we mean an open connected set in Euclidean m-space RX (connectedness is defined in Problem 12, p. $5). 2 In fact, fis bounded on [G'] if [G'] © G (cf. Theorem 2, p. 110). 72 METRIC SPACES cHap, 2 ix — x9] < Sand such that |p(x) — yol < K3, equipped with the metric DI. eC, §) = max |9(x) — The space C* is complete, since it is a closed subspace of the space of all continuous functions on [xy — 3, x» + 8]. Consider the mapping § = Ag defined by the integral equation HO= n+ ff ear (Inds. Clearly 4 is a contraction mapping carrying C* into itself. In fact, if gC, |x — xo < 8 then wo wl= |. fe ead < [1st toate < Kx — ad < KB by (13), and hence = A@ also belongs to C*. Moreover, (UC) = BOIL < [PLP (0) — Fee, YI at < MB max Jol) — BOIL, and hence i 2b, D) < M3p(9, 3) after maximizing with respect to x. But M3 <1, so that A is a con- traction mapping, It follows from Theorem 1 that the equation 9 ~ 4¢, i.e., the integral equation (12), has a unique solution in the space C*. Theorem 2 can easily be generalized to the case of systems of differential equations: THEOREM 2’. Given n functions f(x, Yr...» Yq) defined and continuous on an (a +- 1)-dimensional domain 6 containing the point (0 Yous + + =» Yom)s suppose each f, satisfies a Lipschitz condition of the form LAG Yass +5 Yn) flO Fry +s nF) SM max [yy — Fil in the variables y,,... , Yq. Then there is an interval|x — xl < 8in which the system of differential equations seaYa) GSA yn) (14) has a unique solution Y= AB) ++ In = Pn) SEC. 8 CONTRACTION MaPPINGs 73 satisfying the initial conditions @r0%o) = Yous +++ » Pa(%o) = Yon (as) Proof, Together the differential equations (14) and the initial con- ditions (15) are equivalent to the system of integral equations AO) = yt [MEAD Ode GH. 16) By the continuity of the functions f;, we have AG Yo ye K G seam) ay in some domain G' © G containing the point (x, Yous - + +» Yon): Choose 8 > 0 such that 1) (Yar eee Ya) EG iff [x — Xo) < 8, Lye — You! < KB for all i= Teen ne 2) MB<1. This time let C* be the space of ordered n-tuples P= (G+ +> Pad of continuous functions 9... , 9n defined on the interval Jz — x9] < 3 such that }@(x) — yor < KS for all i= 1,..-,, equipped with the metric 0%» @) = max | p(x) — GG). Clearly C* is complete. Moreover, the mapping } = A¥ defined by the system of integral equations WO) = Yost PAE G0, --5 9a at (x — xo < 8,8 = 1.040) isa contraction mapping carrying C* into itself, In fact, if P=(Prer NEC, [xml <3, then 14) — sol =| flees eslO)ae] < KB = Bo) by (17), so that } = (i, --- 5%, Ag also belongs to C*. Moreover, 1G) — WO = FE GD. Gal — A WD, 5 BADIA < MB imax 4,9 — |@,(01, 74 MERIC SPACES cHar. 2 and hence o(b, §) < MBe(o, $) after maximizing with respect to x and i. But M3 <1, so that 4 isa contraction mapping. It follows from Theorem 1 that the equation = Ag, ie, the system of integral equations (16), has a unique solution in the space C*. 8.3. Contraction mappings and integral equations. We now show how the method of successive approximations can be used to prove the existence and uniqueness of solutions of integral equations. Example 1. By a Fredholm equation (of the second kind) is meant an integral equation of the form f0) = 2 Ke 970) 4 + 900. (8) involving two given functions K and 9, an unknown function f and an arbitrary parameter 4. The function K is called the kernel of the equation, and the equation is said to be homogeneous if @ = 0 (but otherwise non- homogeneous). Suppose K(x, ») and 9(x) are continuous on the square a-< x fy» +++ to this solution are given by 1) KG Ms dy +9) H=h2-.d, FAX) SEC. 8 CONTRACTION MAPPINGS 75 where any function continuous on [a, b] can be chosen as f. Note that the method of successive approximations can be applied to the equation (18) only for sufficiently small |). Example 2. Next consider the Volterra equation oo) = 21° Ke 9 f0) dy + 20), 20) which differs from the Fredholm equation (18) by having the variable x rather than the fixed number b as the upper limit of integration." It is easy to sce that the method of successive approximations can be applied to the Volterra equation (20) for arbitrary %, not just for sufficiently small |A as in the case of the Fredholm equation (18). In fact, let A be the mapping of C,,a) into itself defined by AFG) = 2]. KC, VFO) dy + ol), and let fy, fs © Craay Then [4fiG) — AAC) K(x, NAO) — AON dy < AM(x — a) max | f(x) — f(0)|s where : M = max | K(x, y)| It follows that [AA(8) — AYRO)| < 22M max |) — HO [= a) ae ae =e =O max Lf) ~All, and in general, nay” [A%ACe) — AXf(X)] < A"M ee IAC) —ACIL