Chapter 7 · Linear Integral Equations

7.2 Eigenvalues, Eigenfunctions and the Resolvent Kernel

Characteristic numbers and eigenfunctions of a kernel, bounds on them, iterated kernels, the Neumann series and the resolvent kernel $R(x,t;\lambda)$, entire resolvents of Volterra kernels, and the spectral theory of symmetric kernels: real characteristic numbers, orthogonal eigenfunctions, the Hilbert–Schmidt and Mercer theorems.

In §7.1 we solved Fredholm equations with separable kernels by linear algebra and found that everything hinges on the numbers $\lambda$ at which the homogeneous equation $y=\lambda\int_a^bK(x,t)y(t)\,dt$ has nonzero solutions. For a general continuous kernel there is no finite matrix to diagonalise, but the same objects survive: characteristic numbers and eigenfunctions, the integral-equation analogues of the eigenvalues and eigenvectors of a matrix. Away from them we would like a formula for the solution valid for every right-hand side at once, $y=f+\lambda\int_a^bR(x,t;\lambda)f(t)\,dt$, just as $(I-\lambda A)^{-1}$ solves every linear system. The kernel $R$ is the resolvent kernel, and the most direct way to build it is to iterate: $y=f+\lambda\mathcal{K}f+\lambda^2\mathcal{K}^2f+\cdots$, the Neumann series, which is the geometric series $(I-\lambda\mathcal{K})^{-1}$ $=\sum\lambda^n\mathcal{K}^n$ for operators.

The section answers the natural questions. How small must $|\lambda|$ be for the Neumann series to converge, and what is its true radius of convergence? Why does the resolvent of a Volterra kernel exist for every $\lambda$? And what extra structure appears when the kernel is symmetric, $K(x,t)=K(t,x)$ — the case of every Green's function of a self-adjoint boundary value problem? The answer to the last question, due to Hilbert and Schmidt, is a complete analogue of the spectral theorem for symmetric matrices: real characteristic numbers, orthogonal eigenfunctions, and expansions of the kernel and of the solution in eigenfunctions. CSIR NET asks for resolvents of simple kernels, for the range of $\lambda$ in which the Neumann series converges, for characteristic numbers and eigenfunctions, and for true/false statements about symmetric kernels.

Characteristic Numbers and Eigenfunctions

Throughout, $K$ is continuous on $Q=[a,b]^2$ (possibly complex-valued), and we write $$(\mathcal{K}y)(x)=\int_a^bK(x,t)\,y(t)\,dt,\qquad (f,g)=\int_a^bf(x)\,\overline{g(x)}\,dx,\qquad\|f\|_2=(f,f)^{1/2}.$$ Two sizes of a kernel will appear constantly: the maximum $M=\max_Q|K|$ and the Hilbert–Schmidt norm $$B=\|K\|_2=\left(\int_a^b\!\!\int_a^b|K(x,t)|^2\,dx\,dt\right)^{1/2}\le M(b-a).$$

📘 Definition 7.4 — Eigenvalues of the integral operator, multiplicity
Recall (§7.1) that $\lambda$ is a characteristic number of $K$ if $y=\lambda\mathcal{K}y$ has a continuous solution $y\not\equiv0$, an eigenfunction. Equivalently $\mu=1/\lambda$ is a nonzero eigenvalue of the operator $\mathcal{K}$: $\mathcal{K}y=\mu y$. The eigenfunctions belonging to $\lambda$, together with $0$, form a subspace of $C[a,b]$, the eigenspace of $\lambda$; its dimension is the multiplicity (also called the rank or index) of $\lambda$. A list $\lambda_1,\lambda_2,\dots$ of characteristic numbers "repeated according to multiplicity" contains each $\lambda$ as many times as its multiplicity.
📐 Theorem 7.10 — Bounds on characteristic numbers and multiplicities
Let $K$ be continuous on $[a,b]^2$ with Hilbert–Schmidt norm $B$ and maximum $M$.
  1. Every characteristic number satisfies $$|\lambda|\ge\frac{1}{B}\ge\frac{1}{M(b-a)}.$$ In particular, for $|\lambda|B\lt1$ the homogeneous equation has only the trivial solution.
  2. The multiplicity $m$ of a characteristic number $\lambda$ is finite, and $m\le|\lambda|^2B^2$.

Proof. (1) Let $y=\lambda\mathcal{K}y$, $y\not\equiv0$. By the Cauchy–Schwarz inequality, for each $x$ $$|y(x)|^2=|\lambda|^2\left|\int_a^bK(x,t)y(t)\,dt\right|^2\le|\lambda|^2\int_a^b|K(x,t)|^2dt\;\|y\|_2^2.$$ Integrating over $x$ gives $\|y\|_2^2$ $\le|\lambda|^2B^2\|y\|_2^2$. Since $y$ is continuous and not identically zero, $\|y\|_2\gt0$, so $|\lambda|B\ge1$. Finally $B^2\le M^2(b-a)^2$.

(2) Let $\varphi_1,\dots,\varphi_m$ be linearly independent eigenfunctions for $\lambda$; by Gram–Schmidt we may take them orthonormal. Fix $x$ and apply Bessel's inequality to the function $g_x(t)=\overline{K(x,t)}$ and the orthonormal set $\varphi_1,\dots,\varphi_m$: since $(\varphi_i,g_x)$ $=\int K(x,t)\varphi_i(t)\,dt$ $=\varphi_i(x)/\lambda$, $$\sum_{i=1}^m\frac{|\varphi_i(x)|^2}{|\lambda|^2}\le\int_a^b|K(x,t)|^2\,dt.$$ Integrating over $x$ and using $\|\varphi_i\|_2=1$ gives $m/|\lambda|^2\le B^2$. ∎

Both bounds can be sharp. For $K=xt$ on $[0,1]$, $B=\tfrac13$ and the only characteristic number is $3=1/B$. For $\cos(x-t)$ on $[0,2\pi]$, $B^2=2\pi^2$ and the characteristic number $\frac1\pi$ has multiplicity $2=\frac{1}{\pi^2}\cdot2\pi^2$.

For a separable kernel the characteristic numbers are the zeros of the polynomial $D(\lambda)=\det(I-\lambda A)$ (the Fredholm alternative for separable kernels in §7.1). Fredholm showed that the whole theory survives for every continuous kernel, with $D$ replaced by an entire function.

📐 Theorem 7.11 — Fredholm's theorems for continuous kernels (statement, with proof sketch)
Let $K$ be continuous on $[a,b]^2$. There is an entire function $D(\lambda)$, the Fredholm determinant, with $D(0)=1$, such that:
  1. the characteristic numbers of $K$ are exactly the zeros of $D$; they form a finite or countably infinite set with no finite accumulation point, and each has finite multiplicity;
  2. $\lambda$ is a characteristic number of $K$ if and only if it is one of $K^{T}$, with the same multiplicity;
  3. if $D(\lambda)\ne0$, then $y=f+\lambda\mathcal{K}y$ has exactly one solution for every $f\in C[a,b]$, namely $y=f+\lambda\int_a^bR(x,t;\lambda)f(t)\,dt$ with $R=D(x,t;\lambda)/D(\lambda)$ for a kernel $D(x,t;\lambda)$ entire in $\lambda$;
  4. if $D(\lambda)=0$, the equation is solvable if and only if $(f,\bar z)=\int_a^bf\,z\,dx=0$ for every solution $z$ of the transposed homogeneous equation $z=\lambda\int_a^bK(t,x)z(t)\,dt$.
For a separable kernel, $D(\lambda)=\det(I-\lambda A)$, and in general $$D(\lambda)=\sum_{n=0}^{\infty}\frac{(-\lambda)^n}{n!}\int_a^b\!\!\cdots\!\int_a^b\det\big[K(x_i,x_j)\big]_{i,j=1}^n\,dx_1\cdots dx_n.$$

Proof (sketch). Fix $r\gt0$; we prove (1)–(4) for $|\lambda|\lt r$. By the Weierstrass approximation theorem in two variables there is a polynomial $P(x,t)$ with $\max|K-P|\lt\frac{1}{r(b-a)}$. Write $K=P+H$. For $|\lambda|\lt r$, $|\lambda|\max|H|(b-a)\lt1$, so by Theorem 7.12 below the operator $I-\lambda\mathcal{H}$ is invertible, with inverse $I+\lambda\mathcal{R}_H$ given by a Neumann series. The equation $y-\lambda\mathcal{P}y-\lambda\mathcal{H}y=f$ is then equivalent to $$y-\lambda(I+\lambda\mathcal{R}_H)\mathcal{P}y=(I+\lambda\mathcal{R}_H)f,$$ and $(I+\lambda\mathcal{R}_H)\mathcal{P}$ is an integral operator with a separable kernel (because $P$ is). The Fredholm alternative for separable kernels (§7.1) now gives the alternative for $|\lambda|\lt r$, with a determinant that is analytic in $\lambda$; the transposed statement follows from the same reduction applied to $K^T$, and the determinants can be matched to Fredholm's series. Since $r$ was arbitrary, the zeros of an analytic function that is not identically zero are isolated, and finite-dimensionality is Theorem 7.10(2). ∎

✏️ Example 1 — Eigenfunctions of $x+t$ and the bound $|\lambda|\ge1/B$
In Example 5 of §7.1 we found the characteristic numbers $\lambda_{1,2}=-6\pm4\sqrt3$ of $K=x+t$ on $[0,1]$. Find the eigenfunctions, show they are orthogonal, and compare $|\lambda_1|$ with the bounds of Theorem 7.10.
Null vectors With $a=(x,1)$, $b=(1,t)$, the first row of $(I-\lambda A)c=0$ is $(1-\tfrac\lambda2)c_1-\lambda c_2=0$, so $c_2/c_1=\tfrac1\lambda-\tfrac12$. For $\lambda_1=-6+4\sqrt3$: $\tfrac1{\lambda_1}$ $=\tfrac{6+4\sqrt3}{12}$ $=\tfrac12+\tfrac{\sqrt3}{3}$, so $c_2/c_1=\tfrac1{\sqrt3}$. For $\lambda_2=-6-4\sqrt3$: $c_2/c_1=-\tfrac1{\sqrt3}$.
Eigenfunctions $y_1=\sqrt3\,x+1$ (for $\lambda_1\approx0.928$) and $y_2=\sqrt3\,x-1$ (for $\lambda_2\approx-12.93$). Check: $\lambda_1\int_0^1(x+t)(\sqrt3t+1)\,dt$ $=\lambda_1\left[\left(\tfrac{\sqrt3}{2}+1\right)x+\tfrac{\sqrt3}3+\tfrac12\right]$, and $\lambda_1\left(\tfrac{\sqrt3}2+1\right)$ $=\sqrt3$ (multiply out), so the result is $\sqrt3x+1$.
Orthogonal $\int_0^1(\sqrt3x+1)(\sqrt3x-1)\,dx$ $=\int_0^1(3x^2-1)\,dx=0$ — as Theorem 7.15 predicts, because the kernel is symmetric.
Bounds $M=\max(x+t)=2$, so $|\lambda|\ge\tfrac12$. $B^2=\int_0^1\!\int_0^1(x+t)^2$ $=\tfrac13+\tfrac12+\tfrac13$ $=\tfrac76$, so $|\lambda|$ $\ge\sqrt{6/7}\approx0.9258$. The true value $|\lambda_1|\approx0.9282$ is remarkably close to the Hilbert–Schmidt bound.
✏️ Example 2 — The kernel $\min(x,t)$ on $[0,1]$
Find all characteristic numbers and eigenfunctions of $K(x,t)=\min(x,t)$ on $[0,1]$.
To an ODE $y(x)=\lambda\left[\int_0^xt\,y(t)\,dt+x\int_x^1y(t)\,dt\right]$. Differentiating (the boundary terms $\pm\lambda xy(x)$ cancel): $y'(x)=\lambda\int_x^1y(t)\,dt$, and $y''=-\lambda y$. Putting $x=0$ in the equation and $x=1$ in the formula for $y'$: $y(0)=0$, $y'(1)=0$. (This is Example 3 of §7.1 read backwards.)
Sign of $\lambda$ If $\lambda=-\mu^2\lt0$, then $y=c\sinh\mu x$ and $y'(1)=c\mu\cosh\mu\ne0$ unless $c=0$; $\lambda=0$ is never characteristic. So $\lambda\gt0$.
Solve $y=c\sin(\sqrt\lambda x)$ and $y'(1)=c\sqrt\lambda\cos\sqrt\lambda=0$ give $\sqrt\lambda=\tfrac{(2n-1)\pi}{2}$: $$\lambda_n=\frac{(2n-1)^2\pi^2}{4},\qquad\varphi_n(x)=\sqrt2\,\sin\frac{(2n-1)\pi x}{2}\qquad(n\in\mathbb{N}),$$ normalised so that $\int_0^1\varphi_n^2=1$. Each has multiplicity one (the ODE solution with $y(0)=0$ is unique up to a factor).
Check the bound $B^2=\int\!\!\int\min(x,t)^2$ $=2\int_0^1\!\int_0^xt^2\,dt\,dx$ $=\tfrac16$, so Theorem 7.10 gives $\lambda\ge\sqrt6\approx2.449$; the smallest characteristic number is $\tfrac{\pi^2}4\approx2.467$.

Iterated Kernels, the Neumann Series and the Resolvent Kernel

Successive approximations $y_0=f$, $y_{n+1}=f+\lambda\mathcal{K}y_n$ give $y_n=f+\lambda\mathcal{K}f+\dots+\lambda^n\mathcal{K}^nf$. The powers $\mathcal{K}^n$ are again integral operators, whose kernels are obtained by "matrix multiplication" of kernels.

📘 Definition 7.5 — Iterated kernels and the resolvent kernel
The iterated kernels of $K$ are $K_1=K$ and $$K_{n+1}(x,t)=\int_a^bK(x,s)\,K_n(s,t)\,ds\qquad(n\ge1).$$ Then $(\mathcal{K}^nf)(x)$ $=\int_a^bK_n(x,t)f(t)\,dt$, and $K_{m+n}(x,t)$ $=\int_a^bK_m(x,s)K_n(s,t)\,ds$ for all $m,n\ge1$. The series $$R(x,t;\lambda)=\sum_{n=1}^{\infty}\lambda^{n-1}K_n(x,t)=K+\lambda K_2+\lambda^2K_3+\cdots,$$ where it converges, is the resolvent kernel (reciprocal kernel) of $K$; the corresponding series for the solution, $y=f+\sum_{n\ge1}\lambda^n\mathcal{K}^nf$, is the Neumann series.

The identity $K_{m+n}=\int K_mK_n$ is associativity of the operator product $\mathcal{K}^{m+n}$ $=\mathcal{K}^m\mathcal{K}^n$ (Fubini's theorem). For a Volterra kernel ($K=0$ for $t\gt x$) all iterated kernels are again Volterra kernels: $K_{n+1}(x,t)$ $=\int_t^xK(x,s)K_n(s,t)\,ds$.

📐 Theorem 7.12 — Convergence of the Neumann series
Let $K$ be continuous on $[a,b]^2$, with $M$ and $B$ as above, and let $k(x)=\left(\int_a^b|K(x,s)|^2ds\right)^{1/2}$, $\tilde k(t)$ $=\left(\int_a^b|K(s,t)|^2ds\right)^{1/2}$.
  1. $|K_n(x,t)|\le M^n(b-a)^{n-1}$, $\ \|K_n\|_2\le B^n$, and $|K_{n+2}(x,t)|$ $\le k(x)\,B^n\,\tilde k(t)$.
  2. If $|\lambda|B\lt1$ (in particular if $|\lambda|M(b-a)\lt1$), the series $R(x,t;\lambda)$ $=\sum\lambda^{n-1}K_n$ converges absolutely and uniformly on $[a,b]^2$ to a continuous kernel satisfying the resolvent equations $$R(x,t;\lambda)=K(x,t)+\lambda\int_a^bK(x,s)R(s,t;\lambda)\,ds,$$ $$R(x,t;\lambda)=K(x,t)+\lambda\int_a^bR(x,s;\lambda)K(s,t)\,ds.$$
  3. For such $\lambda$ and every $f\in C[a,b]$, the equation $y=f+\lambda\mathcal{K}y$ has exactly one continuous solution, $$y(x)=f(x)+\lambda\int_a^bR(x,t;\lambda)\,f(t)\,dt,$$ and the successive approximations converge to it uniformly.

Proof. (1) By induction, $|K_{n+1}(x,t)|$ $\le\int_a^bM\cdot M^n(b-a)^{n-1}ds$ $=M^{n+1}(b-a)^n$. Next, Cauchy–Schwarz in $s$ gives $|K_{n+1}(x,t)|^2$ $\le\int|K(x,s)|^2ds\int|K_n(s,t)|^2ds$; integrating over $x$ and $t$, $\|K_{n+1}\|_2^2\le B^2\|K_n\|_2^2$. Finally $K_{n+2}(x,t)$ $=\iint K(x,s)K_n(s,r)K(r,t)\,ds\,dr$, and Cauchy–Schwarz in $(s,r)$ gives $|K_{n+2}(x,t)|$ $\le k(x)\tilde k(t)\|K_n\|_2$ $\le k(x)\tilde k(t)B^n$.

(2) $k$ and $\tilde k$ are bounded by $M\sqrt{b-a}$, so the terms of $\sum_{n\ge3}\lambda^{n-1}K_n$ are bounded by $M^2(b-a)\,|\lambda|\,(|\lambda|B)^{n-2}$: a convergent geometric series when $|\lambda|B\lt1$. The Weierstrass M-test gives absolute and uniform convergence, so $R$ is continuous. Multiplying the series by $K(x,s)$ and integrating term by term, $$\lambda\int_a^bK(x,s)R(s,t;\lambda)\,ds=\sum_{n\ge1}\lambda^nK_{n+1}(x,t)=R-K;$$ the second resolvent equation is proved the same way using $K_{n+1}=\int K_nK$.

(3) Let $y=f+\lambda\int Rf$. Then, by Fubini and the first resolvent equation, $$\lambda\mathcal{K}y=\lambda\int_a^bK f\,dt+\lambda^2\!\int_a^b\!\left[\int_a^bK(x,s)R(s,t)\,ds\right]f(t)\,dt=\lambda\int_a^bR(x,t)f(t)\,dt=y-f.$$ So $y$ is a solution. It is the only one: the difference of two solutions solves the homogeneous equation, which has only the trivial solution because $|\lambda|B\lt1$ (Theorem 7.10(1)). The successive approximations are the partial sums $f+\sum_{n\le N}\lambda^n\mathcal{K}^nf$, which converge uniformly by (2). ∎

The condition $|\lambda|B\lt1$ is sufficient, not necessary. The truth is described by Fredholm's theory.

📐 Theorem 7.13 — The resolvent beyond the Neumann disc (statement)
Let $K$ be continuous on $[a,b]^2$. The resolvent $R(x,t;\lambda)$, defined by the Neumann series near $\lambda=0$, extends to a meromorphic function of $\lambda$ (for each $x,t$), $R=D(x,t;\lambda)/D(\lambda)$, whose poles are exactly the characteristic numbers of $K$. For every $\lambda$ that is not a characteristic number it satisfies the resolvent equations and $y=f+\lambda\int Rf$ is the unique solution. Consequently the radius of convergence of the Neumann series (as a power series in $\lambda$) equals $|\lambda_1|$, the smallest modulus of a characteristic number, and is $\infty$ if $K$ has none.

For a separable kernel all of this is explicit: by the Fredholm alternative for separable kernels (§7.1), $R=\sum_{i,j}a_i(x)[(I-\lambda A)^{-1}]_{ij}b_j(t)$, a rational function of $\lambda$ whose denominator is $D(\lambda)=\det(I-\lambda A)$. When the Neumann series diverges ($|\lambda|\gt|\lambda_1|$) the resolvent still exists at every non-characteristic $\lambda$ — the series is just the wrong way to compute it.

✏️ Example 3 — A resolvent, three radii, and a solution
Let $K(x,t)=xt+x^2t^2$ on $[-1,1]$. (a) Compute the iterated kernels and the resolvent. (b) Compare the true radius of convergence of the Neumann series with the guarantees $|\lambda|M(b-a)\lt1$ and $|\lambda|B\lt1$. (c) Solve $y(x)=1+\int_{-1}^1(xt+x^2t^2)\,y(t)\,dt$.
(a) Iterated kernels Since $\int_{-1}^1s^3\,ds=0$, $\int_{-1}^1s^2\,ds=\tfrac23$, $\int_{-1}^1s^4\,ds=\tfrac25$: $$K_2=\int_{-1}^1(xs+x^2s^2)(st+s^2t^2)\,ds=\tfrac23xt+\tfrac25x^2t^2,$$ and inductively $K_n=\left(\tfrac23\right)^{n-1}xt+\left(\tfrac25\right)^{n-1}x^2t^2$. Summing two geometric series: $$R(x,t;\lambda)=\frac{3xt}{3-2\lambda}+\frac{5x^2t^2}{5-2\lambda}.$$
(b) Radii The poles $\tfrac32$ and $\tfrac52$ are the characteristic numbers (eigenfunctions $x$ and $x^2$), so the Neumann series converges exactly for $|\lambda|\lt\tfrac32$. The crude bound: $M=2$, $b-a=2$, guaranteeing only $|\lambda|\lt\tfrac14$. The Hilbert–Schmidt bound: $B^2=\iint(xt+x^2t^2)^2$ $=\tfrac49+0+\tfrac4{25}$ $=\tfrac{136}{225}$, guaranteeing $|\lambda|$ $\lt\tfrac{15}{\sqrt{136}}\approx1.286$.
(c) Solution $\lambda=1\lt\tfrac32$, but we can use the closed form anyway: $R(x,t;1)=3xt+\tfrac53x^2t^2$ and $$y=1+\int_{-1}^1\left(3xt+\tfrac53x^2t^2\right)dt=1+\tfrac{10}{9}x^2.$$
Check $\int_{-1}^1(xt+x^2t^2)\left(1+\tfrac{10}9t^2\right)dt$ $=x\cdot0+x^2\left(\tfrac23+\tfrac{10}9\cdot\tfrac25\right)$ $=\tfrac{10}9x^2$. Because $f=1$ is orthogonal to the eigenfunction $x$, only the characteristic number $\tfrac52$ influences this solution.
Interactive 1 — The Neumann Series: Where It Converges and Where It Does Not

The faint curves are the partial sums $y_N=f+\lambda\mathcal{K}f+\dots+\lambda^N\mathcal{K}^Nf$ of the Neumann series, the bold curve is the current one, and the dashed curve is the exact solution of $y=f+\lambda\int K y$. Press Play to add terms. Below the guaranteed radius $1/B$ convergence is certain; between $1/B$ and the first characteristic number $|\lambda_1|$ it still converges (more slowly); beyond $|\lambda_1|$ the partial sums explode even though the equation still has a perfectly good solution.

—3
terms N—
max |y − y_N|—
1/(M(b − a))—
1/B—
|λ₁| (true radius)—

For Volterra kernels the factorials of §7.1 make the resolvent exist for every $\lambda$.

📐 Theorem 7.14 — The resolvent of a Volterra kernel is entire
Let $K$ be continuous on $T=\{a\le t\le x\le b\}$, $M=\max_T|K|$, and extend $K$ by $0$ for $t\gt x$. Then the iterated kernels vanish for $t\gt x$ and satisfy $$|K_n(x,t)|\le\frac{M^n(x-t)^{n-1}}{(n-1)!}\qquad(a\le t\le x\le b).$$ Hence $R(x,t;\lambda)$ $=\sum_{n\ge1}\lambda^{n-1}K_n(x,t)$ converges for every $\lambda\in\mathbb{C}$, uniformly on $T$ and on compact sets of $\lambda$; $R$ is an entire function of $\lambda$ with $|R|\le Me^{|\lambda|M(x-t)}$, and for every $\lambda$ and $f$ the unique solution of $y=f+\lambda\int_a^xKy\,dt$ is $$y(x)=f(x)+\lambda\int_a^xR(x,t;\lambda)\,f(t)\,dt.$$

Proof. If $K_n(s,t)=0$ for $s\lt t$, then $K_{n+1}(x,t)$ $=\int_a^bK(x,s)K_n(s,t)\,ds$ has a nonzero integrand only when $t\le s\le x$; so $K_{n+1}(x,t)=0$ for $t\gt x$ and $K_{n+1}(x,t)$ $=\int_t^xK(x,s)K_n(s,t)\,ds$ otherwise. The estimate follows by induction: $|K_{n+1}(x,t)|$ $\le\int_t^xM\frac{M^n(s-t)^{n-1}}{(n-1)!}\,ds$ $=\frac{M^{n+1}(x-t)^n}{n!}$. Thus $\sum|\lambda|^{n-1}|K_n|$ $\le M\sum\frac{(|\lambda|M(x-t))^{n-1}}{(n-1)!}$ $=Me^{|\lambda|M(x-t)}$, and the M-test gives uniform convergence for $|\lambda|\le r$, any $r$; a uniformly convergent series of polynomials in $\lambda$ is entire. The resolvent equations and the solution formula follow exactly as in Theorem 7.12, and uniqueness is the existence theorem for Volterra equations (§7.1). ∎

✏️ Example 4 — The Volterra resolvent of $e^{x-t}$
Find the resolvent of the Volterra kernel $K(x,t)=e^{x-t}$ ($0\le t\le x$) and solve $y(x)=1+\lambda\int_0^xe^{x-t}y(t)\,dt$. Check the case $\lambda=1$, and confirm the resolvent with the Laplace transform.
Iterated kernels $K_2(x,t)$ $=\int_t^xe^{x-s}e^{s-t}\,ds$ $=(x-t)e^{x-t}$, and by induction $K_n(x,t)$ $=\frac{(x-t)^{n-1}}{(n-1)!}e^{x-t}$ (the factor $e^{x-t}$ survives each integration unchanged).
Resolvent $$R(x,t;\lambda)=e^{x-t}\sum_{n\ge1}\frac{\lambda^{n-1}(x-t)^{n-1}}{(n-1)!}=e^{(1+\lambda)(x-t)},$$ entire in $\lambda$, as Theorem 7.14 promises.
Solution $y=1+\lambda\int_0^xe^{(1+\lambda)(x-t)}\,dt$ $=1+\frac{\lambda}{1+\lambda}\left(e^{(1+\lambda)x}-1\right)$ for $\lambda\ne-1$, and $y=1-x$ for $\lambda=-1$ (then $R\equiv1$). For $\lambda=1$: $y=\tfrac12\left(1+e^{2x}\right)$. Check: $\int_0^xe^{x-t}\cdot\tfrac12(1+e^{2t})\,dt$ $=\tfrac{e^x}2\left[(1-e^{-x})+(e^x-1)\right]$ $=\tfrac{e^{2x}-1}{2}$, and $1+\tfrac{e^{2x}-1}2=y$.
Laplace For a convolution kernel $k(x-t)$ the resolvent is also of convolution type, and the resolvent equation $R=k+\lambda\,k*R$ gives $\hat R$ $=\frac{\hat k}{1-\lambda\hat k}$. Here $\hat k=\frac1{s-1}$, so $\hat R=\frac{1}{s-1-\lambda}$ and $R=e^{(1+\lambda)(x-t)}$.
🧭 Problem-solving strategy — finding a resolvent kernel
  1. Separable kernel: compute $A$, invert $I-\lambda A$, and form $R=\sum a_i[(I-\lambda A)^{-1}]_{ij}b_j$. For rank one, $K=a(x)b(t)$, $K_n=\alpha^{n-1}K$ with $\alpha=\int ab$, and $R=\frac{K}{1-\lambda\alpha}$.
  2. Look for a pattern in $K_2$, $K_3$: often $K_n=c^{n-1}K$ (e.g. $e^{x-t}$ on $[0,1]$, Fredholm) or $K_n=\frac{(x-t)^{n-1}}{(n-1)!}\times(\dots)$ (Volterra); then sum the series.
  3. Convolution Volterra kernel: $\hat R=\hat k/(1-\lambda\hat k)$, then invert the Laplace transform.
  4. Symmetric kernel with known eigenfunctions: $\int R f$ $=\sum\frac{(f,\varphi_n)}{\lambda_n-\lambda}\varphi_n$ (Theorem 7.17).
  5. Check: $R(x,t;0)=K(x,t)$, the poles of $R$ are the characteristic numbers, and $R$ satisfies $R=K+\lambda\int KR$.

Symmetric Kernels

Green's functions of self-adjoint boundary value problems, covariance functions in probability, and kernels like $\min(x,t)$, $\cos(x-t)$, $e^{-|x-t|}$ and $x+t$ are symmetric. For them the theory becomes as clean as that of real symmetric matrices.

📘 Definition 7.6 — Symmetric, Hermitian and positive kernels
A real continuous kernel is symmetric if $K(x,t)=K(t,x)$ for all $x,t$; a complex one is Hermitian if $K(x,t)=\overline{K(t,x)}$. Then $(\mathcal{K}f,g)=(f,\mathcal{K}g)$ for all $f,g\in C[a,b]$. A symmetric kernel is positive (positive semidefinite) if $(\mathcal{K}f,f)$ $=\iint K(x,t)f(t)f(x)\,dt\,dx\ge0$ for every real $f\in C[a,b]$.
📐 Theorem 7.15 — Spectral properties of symmetric kernels
Let $K\not\equiv0$ be real, continuous and symmetric on $[a,b]^2$.
  1. Every characteristic number of $K$ is real, and the eigenfunctions can be chosen real.
  2. Eigenfunctions belonging to different characteristic numbers are orthogonal: $(\varphi,\psi)=0$.
  3. $K$ has at least one characteristic number. The one of smallest modulus satisfies $$\frac{1}{|\lambda_1|}=\max_{\|\varphi\|_2=1}\big|(\mathcal{K}\varphi,\varphi)\big|,$$ and $\|\mathcal{K}\varphi\|_2$ $\le\|\varphi\|_2/|\lambda_1|$ for every $\varphi$.
  4. The characteristic numbers, repeated according to multiplicity, can be listed as $|\lambda_1|$ $\le|\lambda_2|$ $\le\cdots$ with real orthonormal eigenfunctions $\varphi_1,\varphi_2,\dots$; the list is finite or $|\lambda_n|\to\infty$, and $\displaystyle\sum_n\frac{1}{\lambda_n^2}$ $\displaystyle \le\iint K^2$.
  5. If $K$ is positive, all its characteristic numbers are positive.

Proof. (1) Let $\varphi=\lambda\mathcal{K}\varphi$, $\varphi\not\equiv0$. Then $(\mathcal{K}\varphi,\varphi)$ $=\frac1\lambda\|\varphi\|_2^2$. By symmetry $(\mathcal{K}\varphi,\varphi)$ $=(\varphi,\mathcal{K}\varphi)$ $=\overline{(\mathcal{K}\varphi,\varphi)}$, so $(\mathcal{K}\varphi,\varphi)$ is real, and so is $\lambda$ (as $\|\varphi\|_2\gt0$; it is nonzero because $\lambda\ne0$). Since $K$ and $\lambda$ are real, the real and imaginary parts of $\varphi$ are again solutions, and at least one is nonzero.

(2) If $\varphi=\lambda\mathcal{K}\varphi$ and $\psi=\mu\mathcal{K}\psi$ with real $\lambda\ne\mu$, then $$\frac1\lambda(\varphi,\psi)=(\mathcal{K}\varphi,\psi)=(\varphi,\mathcal{K}\psi)=\frac1\mu(\varphi,\psi),$$ so $(\varphi,\psi)=0$.

(3) (Sketch.) Let $\mu_1=\sup_{\|\varphi\|=1}|(\mathcal{K}\varphi,\varphi)|$; one shows $\mu_1\gt0$ (otherwise polarisation gives $(\mathcal{K}f,g)=0$ for all $f,g$, hence $K\equiv0$). Take $\varphi_k$ with $\|\varphi_k\|=1$ and $(\mathcal{K}\varphi_k,\varphi_k)\to\pm\mu_1$. The functions $\mathcal{K}\varphi_k$ are uniformly bounded and equicontinuous (by uniform continuity of $K$ and Cauchy–Schwarz), so by the Arzelà–Ascoli theorem a subsequence converges uniformly; the identity $\|\mathcal{K}\varphi_k\mp\mu_1\varphi_k\|^2$ $=\|\mathcal{K}\varphi_k\|^2\mp2\mu_1(\mathcal{K}\varphi_k,\varphi_k)+\mu_1^2$ $\le2\mu_1^2\mp2\mu_1(\mathcal{K}\varphi_k,\varphi_k)\to0$ then shows that the limit $\psi$ satisfies $\mathcal{K}\psi=\pm\mu_1\psi$, $\psi\ne0$. So $\pm1/\mu_1$ is a characteristic number; and no characteristic number has smaller modulus, because $\varphi=\lambda\mathcal{K}\varphi$ with $\|\varphi\|=1$ gives $\frac1{|\lambda|}$ $=|(\mathcal{K}\varphi,\varphi)|$ $\le\mu_1$. (Here we used the standard fact that for symmetric kernels $\|\mathcal{K}\varphi\|$ $\le\mu_1\|\varphi\|$, which follows from the polarisation identity.)

(4) Choose an orthonormal basis of each (finite-dimensional, Theorem 7.10) eigenspace; by (2) the union is orthonormal. For any finitely many of them the Bessel argument of Theorem 7.10(2) gives $\sum_n\frac{\varphi_n(x)^2}{\lambda_n^2}$ $\le\int_a^bK(x,t)^2\,dt$, and integrating, $\sum_n\frac1{\lambda_n^2}$ $\le\iint K^2$. Hence only finitely many $\lambda_n$ lie in any disc, so they can be ordered by modulus and $|\lambda_n|\to\infty$ if there are infinitely many. (5) If $\varphi=\lambda\mathcal{K}\varphi$ is real and nonzero, $\frac1\lambda\|\varphi\|^2$ $=(\mathcal{K}\varphi,\varphi)\ge0$, so $\lambda\gt0$. ∎

📐 Theorem 7.16 — The Hilbert–Schmidt theorem; Mercer's theorem
Let $K$ be real, continuous and symmetric, with characteristic numbers $\lambda_n$ (repeated by multiplicity) and real orthonormal eigenfunctions $\varphi_n$ as in Theorem 7.15(4).
  1. (Hilbert–Schmidt.) Every function of the form $g=\mathcal{K}h$, $h\in C[a,b]$ ("sourcewise representable"), has the expansion $$g(x)=\sum_n(g,\varphi_n)\,\varphi_n(x)=\sum_n\frac{(h,\varphi_n)}{\lambda_n}\,\varphi_n(x),$$ converging absolutely and uniformly on $[a,b]$.
  2. (Iterated kernels and traces.) For $m\ge2$, $\displaystyle K_m(x,t)$ $\displaystyle =\sum_n\frac{\varphi_n(x)\varphi_n(t)}{\lambda_n^m}$ uniformly, and $\displaystyle\int_a^bK_m(x,x)\,dx$ $\displaystyle =\sum_n\frac{1}{\lambda_n^m}$. In particular $$\sum_n\frac1{\lambda_n^2}=\int_a^b\!\!\int_a^bK(x,t)^2\,dx\,dt,$$ and the bilinear series $\sum_n\frac{\varphi_n(x)\varphi_n(t)}{\lambda_n}$ converges to $K$ in the mean square on $[a,b]^2$.
  3. $K$ has only finitely many characteristic numbers if and only if it is separable, and then $K(x,t)$ $=\sum_n\frac{\varphi_n(x)\varphi_n(t)}{\lambda_n}$ exactly.
  4. (Mercer.) If $K$ is positive (all $\lambda_n\gt0$), the bilinear series converges to $K$ absolutely and uniformly on $[a,b]^2$, and $$\sum_n\frac1{\lambda_n}=\int_a^bK(x,x)\,dx.$$

Proof (of (1)–(3), with (1) sketched; (4) stated). (1) Uniform convergence. By Cauchy–Schwarz and the inequality $\sum_n\varphi_n(x)^2/\lambda_n^2$ $\le\int K(x,t)^2dt$ $\le M^2(b-a)$ from the proof of Theorem 7.15(4), $$\sum_{n=p}^q\left|\frac{(h,\varphi_n)\varphi_n(x)}{\lambda_n}\right|\le\Big(\sum_{n=p}^q(h,\varphi_n)^2\Big)^{1/2}M\sqrt{b-a},$$ and the right side tends to $0$ as $p,q\to\infty$ by Bessel's inequality for $h$, uniformly in $x$. The limit is $g$. The coefficients are right: $(g,\varphi_n)$ $=(\mathcal{K}h,\varphi_n)$ $=(h,\mathcal{K}\varphi_n)$ $=(h,\varphi_n)/\lambda_n$. The kernel $K^{(N)}$ $=K-\sum_{n\le N}\varphi_n(x)\varphi_n(t)/\lambda_n$ is symmetric, kills $\varphi_1,\dots,\varphi_N$ and has characteristic numbers $\lambda_{N+1},\lambda_{N+2},\dots$ only; by the extremal property in Theorem 7.15(3), $\|K^{(N)}h\|_2$ $\le\|h\|_2/|\lambda_{N+1}|\to0$ (or $K^{(N)}\equiv0$ if the list stops at $N$). So $g-\sum_{n\le N}(g,\varphi_n)\varphi_n$ $=\mathcal{K}^{(N)}h\to0$ in mean square, and the uniform limit must be $g$.

(2) $K_2(\cdot,t)$ $=\mathcal{K}[K(\cdot,t)]$ is sourcewise representable with $(K(\cdot,t),\varphi_n)$ $=\varphi_n(t)/\lambda_n$, so (1) gives $K_2(x,t)$ $=\sum\varphi_n(x)\varphi_n(t)/\lambda_n^2$, uniformly in $x$ (and in $t$, by the estimate above); similarly for $m\gt2$. Put $x=t$ and integrate term by term: $\int K_m(x,x)\,dx$ $=\sum1/\lambda_n^m$. For $m=2$, $\int K_2(x,x)\,dx$ $=\iint K(x,s)K(s,x)\,ds\,dx$ $=\iint K^2$ by symmetry. Finally, expanding the square and using orthonormality, $$\iint\Big(K-\sum_{n\le N}\frac{\varphi_n(x)\varphi_n(t)}{\lambda_n}\Big)^2=\iint K^2-\sum_{n\le N}\frac1{\lambda_n^2}\to0.$$ (3) If there are finitely many $\lambda_n$, the remainder kernel $K^{(N)}$ is symmetric with no characteristic numbers, hence zero by Theorem 7.15(3); the converse follows from the Fredholm alternative for separable kernels in §7.1. (4) is Mercer's theorem (1909); its proof uses positivity to show that the partial sums of $\sum\varphi_n(x)^2/\lambda_n$ increase to $K(x,x)$ and then Dini's theorem. ∎

📐 Theorem 7.17 — Schmidt's solution formula for symmetric kernels
Let $K$ be real, continuous and symmetric with $\lambda_n,\varphi_n$ as above, and $f\in C[a,b]$.
  1. If $\lambda$ is not a characteristic number, the unique solution of $y=f+\lambda\mathcal{K}y$ is $$y(x)=f(x)+\lambda\sum_n\frac{(f,\varphi_n)}{\lambda_n-\lambda}\,\varphi_n(x),$$ the series converging absolutely and uniformly. Formally, $R(x,t;\lambda)$ $=\sum_n\frac{\varphi_n(x)\varphi_n(t)}{\lambda_n-\lambda}$.
  2. If $\lambda=\lambda_k$, the equation is solvable iff $(f,\varphi_j)=0$ for every eigenfunction $\varphi_j$ belonging to $\lambda_k$; then the solutions are the same series, summed over $n$ with $\lambda_n\ne\lambda_k$, plus an arbitrary combination of those $\varphi_j$.

Proof. (1) Necessity of the formula. If $y$ solves the equation, $y-f=\mathcal{K}(\lambda y)$ is sourcewise representable, so by Theorem 7.16(1) $y-f=\sum c_n\varphi_n$ with $c_n=(y-f,\varphi_n)$ $=\lambda(\mathcal{K}y,\varphi_n)$ $=\frac{\lambda}{\lambda_n}(y,\varphi_n)$ $=\frac{\lambda}{\lambda_n}\big((f,\varphi_n)+c_n\big)$. Solving, $c_n(\lambda_n-\lambda)$ $=\lambda(f,\varphi_n)$, which gives the formula when no $\lambda_n$ equals $\lambda$, and also uniqueness. Sufficiency. Write $\frac{(f,\varphi_n)}{\lambda_n-\lambda}$ $=\frac{\lambda_n}{\lambda_n-\lambda}\cdot\frac{(f,\varphi_n)}{\lambda_n}$; the factors $\frac{\lambda_n}{\lambda_n-\lambda}$ are bounded (they tend to $1$), so the series converges absolutely and uniformly by the estimate in the proof of Theorem 7.16(1). Apply $\mathcal{K}$ term by term ($\mathcal{K}\varphi_n$ $=\varphi_n/\lambda_n$) and use the expansion $\mathcal{K}f$ $=\sum_n\frac{(f,\varphi_n)}{\lambda_n}\varphi_n$ of Theorem 7.16(1): $$\lambda\mathcal{K}y=\lambda\sum_n(f,\varphi_n)\varphi_n\left[\frac1{\lambda_n}+\frac{\lambda}{\lambda_n(\lambda_n-\lambda)}\right],$$ and the bracket equals $\frac{1}{\lambda_n-\lambda}$, so $\lambda\mathcal{K}y=y-f$. (2) If $\lambda=\lambda_k$, the relation $c_n(\lambda_n-\lambda)$ $=\lambda(f,\varphi_n)$ for the $\varphi_n$ belonging to $\lambda_k$ forces $(f,\varphi_n)=0$ and leaves those $c_n$ free; this is the Fredholm alternative with $K^{T}=K$. ∎

✏️ Example 5 — Two sums from the kernel $\min(x,t)$
Use Example 2 and Theorem 7.16 to evaluate $\displaystyle\sum_{n\ge1}\frac{1}{(2n-1)^2}$ and $\displaystyle\sum_{n\ge1}\frac{1}{(2n-1)^4}$.
Positivity All $\lambda_n=\frac{(2n-1)^2\pi^2}4$ are positive (indeed $\min(x,t)$ $=\int_0^1\mathbf 1_{s\lt x}\mathbf 1_{s\lt t}\,ds$, so $(\mathcal{K}h,h)$ $=\int_0^1\big(\int_s^1h\big)^2ds\ge0$: Exercise III.6). Mercer's theorem applies.
Trace $\int_0^1\min(x,x)\,dx$ $=\int_0^1x\,dx$ $=\tfrac12$ $=\sum_n\frac{4}{(2n-1)^2\pi^2}$, so $$\sum_{n\ge1}\frac1{(2n-1)^2}=\frac{\pi^2}{8}.$$
Hilbert–Schmidt $\iint\min(x,t)^2=\tfrac16$ (Example 2) $=\sum_n\frac{16}{(2n-1)^4\pi^4}$, so $$\sum_{n\ge1}\frac1{(2n-1)^4}=\frac{\pi^4}{96}.$$
Consistency Since $\sum_{n}\frac1{n^2}$ $=\sum_{\text{odd}}+\frac14\sum_n\frac1{n^2}$, the first sum gives $\sum\frac1{n^2}$ $=\frac43\cdot\frac{\pi^2}8$ $=\frac{\pi^2}6$; similarly $\sum\frac1{n^4}$ $=\frac{16}{15}\cdot\frac{\pi^4}{96}$ $=\frac{\pi^4}{90}$ — the values obtained from the string Green's function in §2.4.
Interactive 2 — Building a Symmetric Kernel from Its Eigenfunctions (Hilbert–Schmidt and Mercer)

The dashed curve is one slice $x\mapsto K(x,t_0)$ of a positive symmetric kernel; the bold curve is the partial bilinear sum $\sum_{n\le N}\frac{\varphi_n(x)\varphi_n(t_0)}{\lambda_n}$, and the faint one the previous partial sum. Press Play to add eigenfunctions and drag $t_0$ to move the slice. Because these kernels are positive, Mercer's theorem says the convergence is uniform; the kink at $x=t_0$ is where it is slowest. The readouts check the trace formula $\sum1/\lambda_n=\int K(x,x)\,dx$ and the Hilbert–Schmidt identity $\sum1/\lambda_n^2=\iint K^2$.

0.402
N—
max error on the slice—
Σ 1/λₙ vs ∫K(x,x)dx—
Σ 1/λₙ² vs ∬K²—
✏️ Example 6 — Prove or disprove
Decide, with proof or counterexample: (a) a real continuous symmetric kernel can have a non-real characteristic number; (b) the Neumann series of a continuous kernel converges only for $|\lambda|\lt1/(M(b-a))$; (c) a nonzero real continuous symmetric kernel can have only finitely many characteristic numbers; (d) if $\lambda$ is not a characteristic number, the resolvent exists even where the Neumann series diverges.
(a) False Theorem 7.15(1). (For non-symmetric kernels it can happen: $x-t$ on $[0,1]$ has $D=1+\frac{\lambda^2}{12}$, roots $\pm2\sqrt3\,i$.)
(b) False $K=xt$ on $[0,1]$: $M(b-a)=1$, but $K_n=\frac{xt}{3^{n-1}}$, so $R=\sum\lambda^{n-1}\frac{xt}{3^{n-1}}$ converges for $|\lambda|\lt3$. The radius is $|\lambda_1|=3$ (Theorem 7.13); here it happens to equal $1/B$.
(c) True $xt$ (one characteristic number) or $\cos(x-t)$ on $[0,2\pi]$ (one, of multiplicity two). By Theorem 7.16(3) this happens exactly for separable symmetric kernels.
(d) True For $xt$ on $[0,1]$ and $\lambda=5$ the series diverges, but $R=\frac{3xt}{3-\lambda}$ $=-\frac32xt$ satisfies the resolvent equation: $xt+5\int_0^1xs\left(-\tfrac32st\right)ds$ $=xt-\tfrac52xt$ $=-\tfrac32xt$. So $y=f+5\int_0^1(-\tfrac32xt)f(t)\,dt$ solves $y=f+5\int_0^1xty\,dt$ for every $f$.
✏️ Example 7 — Schmidt's formula against the exact solution
Solve $y(x)=1+\lambda\int_0^1\min(x,t)\,y(t)\,dt$ (a) exactly, by converting to a boundary value problem, and (b) by Schmidt's formula. Compare the two at $\lambda=1$, $x=1$. For which $\lambda$ is there no solution?
(a) Exact As in Example 2, $y''=-\lambda y$ with $y(0)=1$ and $y'(1)=0$. For $\lambda\gt0$: $y=\cos\sqrt\lambda x+B\sin\sqrt\lambda x$, and $y'(1)=0$ gives $B=\tan\sqrt\lambda$, so $$y(x)=\frac{\cos\big(\sqrt\lambda\,(1-x)\big)}{\cos\sqrt\lambda}.$$ At $\lambda=1$, $x=1$: $y(1)=\frac1{\cos1}\approx1.850816$.
(b) Schmidt $(1,\varphi_n)$ $=\sqrt2\int_0^1\sin\frac{(2n-1)\pi x}{2}dx$ $=\frac{2\sqrt2}{(2n-1)\pi}$, so $$y(x)=1+\lambda\sum_{n\ge1}\frac{4}{(2n-1)\pi}\cdot\frac{\sin\frac{(2n-1)\pi x}{2}}{\lambda_n-\lambda}.$$
Compare At $\lambda=1$, $x=1$ the terms are $\frac{4(-1)^{n+1}}{(2n-1)\pi(\lambda_n-1)}$: the first is $\frac{4}{\pi(\pi^2/4-1)}\approx0.8677$, and summing $20\,000$ terms gives $1.850816$, agreeing with (a) to six decimals. The terms decay like $n^{-3}$, so convergence is fast and uniform, as Theorem 7.17 says.
No solution At $\lambda=\lambda_k$ we need $(1,\varphi_k)=0$, but $(1,\varphi_k)$ $=\frac{2\sqrt2}{(2k-1)\pi}\ne0$. So there is no solution exactly at the characteristic numbers $\lambda_k=\frac{(2k-1)^2\pi^2}4$ — matching the zeros of $\cos\sqrt\lambda$ in (a).
⚠️ Counterexamples and pitfalls
  1. "Eigenvalue" means two things. In $y=\lambda\int Ky$, $\lambda$ is a characteristic number; the eigenvalue of the operator is $1/\lambda$. Exam options often list both — for $xt$ on $[0,1]$ the characteristic number is $3$ and the operator eigenvalue $\frac13$.
  2. No characteristic numbers at all. $\sin x\cos t$ on $[0,\pi]$ ($K_2\equiv0$, so $R=K$ for every $\lambda$), and every continuous Volterra kernel. An entire resolvent does not mean the kernel is Volterra.
  3. Only non-real ones. $x-t$ on $[0,1]$: $\pm2\sqrt3\,i$. Reality is guaranteed only for symmetric (Hermitian) kernels.
  4. The Neumann condition is sufficient, not necessary. $|\lambda|M(b-a)\lt1$ and $|\lambda|B\lt1$ are guarantees; the true radius is $|\lambda_1|$ (Theorem 7.13): $3$ for $xt$ (bound $1$), $\tfrac32$ for $xt+x^2t^2$ on $[-1,1]$ (bounds $\tfrac14$ and $1.286$).
  5. Divergent Neumann series, existing solution. Beyond $|\lambda_1|$ the series diverges, but for non-characteristic $\lambda$ the resolvent and the solution still exist (Example 6(d)).
  6. Non-symmetric kernels: no orthogonality. $K=1-2x+3t$ on $[0,1]$ has characteristic numbers $1$ and $2$ with eigenfunctions $1-x$ and $4x-3$, and $\int_0^1(1-x)(4x-3)\,dx$ $=-\tfrac56\ne0$.
  7. Hilbert–Schmidt expands only the range of $\mathcal{K}$. For $K=xt$ on $[0,1]$ ($\varphi_1=\sqrt3x$), the "expansion" of $f=1$ is $(1,\varphi_1)\varphi_1$ $=\tfrac32x\ne1$: $1$ is not of the form $\int_0^1xt\,h(t)\,dt$.
  8. Uniform convergence of the bilinear series needs Mercer. For general symmetric kernels it converges only in mean square, and the trace formula $\sum1/\lambda_n=\int K(x,x)\,dx$ is guaranteed only for positive (or separable) kernels; for $m\ge2$ the trace formulas always hold.
  9. Count with multiplicity. For $\cos(x-t)$ on $[0,2\pi]$, $\sum1/\lambda_n^2$ must count $\frac1\pi$ twice: $2\pi^2=\iint\cos^2(x-t)$. Counting it once gives $\pi^2$, which is wrong.
  10. $K_2$ is not $K$ squared pointwise. $K_2(x,t)=\int K(x,s)K(s,t)\,ds$; its characteristic numbers are $\lambda_n^2$, all positive for symmetric $K$.
  11. The null space is not an eigenspace. $\int_0^1xt\,y(t)\,dt=0$ for every $y\perp t$ — an infinite-dimensional null space — but $0$ is never a characteristic number.

Exercises

I
Characteristic numbers and eigenfunctions.
  • 1 Find the characteristic numbers and eigenfunctions of $e^{x+t}$ on $[0,1]$.
  • 2 Find the characteristic numbers and eigenfunctions of $x+t$ on $[-1,1]$, and check that the eigenfunctions are orthogonal.
  • 3 Find the characteristic numbers, with multiplicities, of $\cos^2(x-t)$ on $[0,\pi]$.
  • 4 Find the characteristic numbers and eigenfunctions of $1-\max(x,t)$ on $[0,1]$.
  • 5 For $K=xt$ on $[0,1]$ compute $B$, and show that the bound $|\lambda|\ge1/B$ is attained.
  • 6 Show that the set of solutions of $\int_0^1xt\,y(t)\,dt=0$ is infinite-dimensional, and explain why this does not make $0$ a characteristic number.
  • 7 Find the characteristic numbers and eigenfunctions of $K=1-2x+3t$ on $[0,1]$ and of its transpose, and show that $\int_0^1fz\,dx=0$ fails for $f=1$ at both characteristic numbers.
  • 8 Show directly (without the general Volterra theorem of §7.1) that the Volterra kernel $e^{x-t}$ on $0\le t\le x\le1$ has no characteristic numbers.
1 Rank one, $a=e^x$, $b=e^t$: $A=\int_0^1e^{2t}dt$ $=\frac{e^2-1}{2}$. One characteristic number $\lambda$ $=\frac{2}{e^2-1}\approx0.313$, eigenfunction $e^x$.
2 $a=(x,1)$, $b=(1,t)$: $A=\begin{pmatrix}0&2\\2/3&0\end{pmatrix}$, $D=1-\frac43\lambda^2$, $\lambda=\pm\frac{\sqrt3}2$. From $c_1=2\lambda c_2$: $\lambda=\frac{\sqrt3}2$ gives $\sqrt3x+1$, $\lambda=-\frac{\sqrt3}2$ gives $\sqrt3x-1$. $\int_{-1}^1(3x^2-1)\,dx=2-2=0$.
3 $\cos^2(x-t)$ $=\tfrac12+\tfrac12\cos2x\cos2t+\tfrac12\sin2x\sin2t$, three mutually orthogonal terms on $[0,\pi]$. $\tfrac12\int_0^\pi1=\tfrac\pi2$ gives $\lambda=\frac2\pi$ (eigenfunction $1$); $\tfrac12\int_0^\pi\cos^22t$ $=\tfrac\pi4$ gives $\lambda=\frac4\pi$ with eigenfunctions $\cos2x$ and $\sin2x$ — multiplicity $2$.
4 By Exercise II.4 of §7.1 the equation is equivalent to $y''+\lambda y=0$, $y'(0)=0$, $y(1)=0$. So $\lambda_n=\frac{(2n-1)^2\pi^2}{4}$, $\varphi_n$ $=\sqrt2\cos\frac{(2n-1)\pi x}{2}$.
5 $B^2=\int_0^1\!\int_0^1x^2t^2$ $=\frac19$, $B=\frac13$; the characteristic number is $\lambda=1/\int_0^1t^2dt=3=1/B$. (Equality in Cauchy–Schwarz: $K(x,\cdot)$ is proportional to the eigenfunction.)
6 The condition is $\int_0^1t\,y(t)\,dt=0$: one linear condition, satisfied by $1-\tfrac32t$, $t^2-\tfrac34t$, $\cos2\pi t$, … — infinitely many independent functions. They are eigenfunctions of $\mathcal{K}$ for the eigenvalue $\mu=0$, which corresponds to no $\lambda=1/\mu$; the equation $y=\lambda\mathcal{K}y$ for these $y$ would give $y=0$.
7 $a=(1,x)$, $b=(1+3t,-2)$: $D=\frac{(\lambda-1)(\lambda-2)}{2}$, with eigenfunctions $1-x$ ($\lambda=1$) and $4x-3$ ($\lambda=2$). The transposed kernel is $K(t,x)=1-2t+3x$, so a transposed eigenfunction has the form $z=\lambda(\alpha+3\beta x)$ with $\alpha=\int_0^1(1-2t)z$, $\beta=\int_0^1z$; substituting $z=p+qx$ gives $z\propto6x-1$ for $\lambda=1$ and $z\propto3x-1$ for $\lambda=2$. Since $\int_0^1(6x-1)\,dx=2\ne0$ and $\int_0^1(3x-1)\,dx=\frac12\ne0$, the equation with $f=1$ has no solution at either characteristic number (the Fredholm alternative for separable kernels in §7.1).
8 $y=\lambda\int_0^xe^{x-t}y\,dt$ means $e^{-x}y$ $=\lambda\int_0^xe^{-t}y\,dt$ $=:\lambda w$, with $w'=e^{-x}y=\lambda w$, $w(0)=0$. Hence $w\equiv0$ and $y\equiv0$, for every $\lambda$.
II
Iterated kernels, Neumann series and resolvents.
  • 1 Find the resolvent of $xt$ on $[0,1]$ and solve $y(x)=x^2+\lambda\int_0^1xt\,y(t)\,dt$.
  • 2 For $K=x+t$ on $[0,1]$ compute $K_2$ and $K_3$.
  • 3 For the Fredholm kernel $e^{x-t}$ on $[0,1]$ show $K_n=K$, find $R$, and solve $y(x)=1+\lambda\int_0^1e^{x-t}y(t)\,dt$.
  • 4 For $K=\sin x\cos t$ on $[0,\pi]$ show $K_2\equiv0$. What are $R$ and the solution of $y=f+\lambda\mathcal{K}y$?
  • 5 For $K=x+t$ on $[0,1]$ compare $\frac1{M(b-a)}$, $\frac1B$ and the radius of convergence of the Neumann series.
  • 6 Find the resolvent of the Volterra kernel $x-t$ and solve $y(x)=1+\lambda\int_0^x(x-t)\,y(t)\,dt$ for $\lambda\gt0$.
  • 7 Find the resolvent of the Volterra kernel $K\equiv1$ ($0\le t\le x$).
  • 8 Use the Laplace transform to find the resolvent of the Volterra kernel $\sin(x-t)$ for $\lambda=1$, and solve $y(x)=x+\int_0^x\sin(x-t)\,y(t)\,dt$ with it.
  • 9 Derive the resolvent equation $R(x,t;\lambda)-K(x,t)$ $=\lambda\int_a^bK(x,s)R(s,t;\lambda)\,ds$ from the series, and show that, conversely, a continuous $R$ satisfying both resolvent equations gives the unique solution $y=f+\lambda\int Rf$.
1 $K_n=\frac{xt}{3^{n-1}}$, $R=\frac{3xt}{3-\lambda}$. Then $y=x^2+\lambda\int_0^1\frac{3xt}{3-\lambda}t^2\,dt$ $=x^2+\frac{3\lambda x}{4(3-\lambda)}$ for $\lambda\ne3$.
2 $K_2=\int_0^1(x+s)(s+t)\,ds$ $=xt+\frac{x+t}{2}+\frac13$; $K_3=\int_0^1(x+s)K_2(s,t)\,ds$ $=xt+\frac{7}{12}(x+t)+\frac13$.
3 $K_2=\int_0^1e^{x-s}e^{s-t}ds$ $=e^{x-t}$, so $K_n=K$ and $R=\frac{e^{x-t}}{1-\lambda}$ ($\lambda\ne1$; characteristic number $1$). $y=1+\frac{\lambda}{1-\lambda}e^x\int_0^1e^{-t}dt$ $=1+\frac{\lambda(1-e^{-1})}{1-\lambda}e^x$.
4 $K_2=\sin x\cos t\int_0^\pi\cos s\sin s\,ds=0$, so all $K_n=0$ for $n\ge2$ and $R=K$ for every $\lambda$ (entire; no characteristic numbers). $y=f+\lambda\sin x\int_0^\pi\cos t\,f(t)\,dt$.
5 $M=2$: $\frac1{M(b-a)}=\frac12$. $B^2=\frac76$: $\frac1B\approx0.9258$. Radius $|\lambda_1|$ $=4\sqrt3-6\approx0.9282$ (Example 1). For this kernel the Hilbert–Schmidt bound is almost exact.
6 $K_n=\frac{(x-t)^{2n-1}}{(2n-1)!}$ (induction), so $R=\sum\lambda^{n-1}\frac{(x-t)^{2n-1}}{(2n-1)!}$ $=\frac{\sinh(\sqrt\lambda(x-t))}{\sqrt\lambda}$. Then $y=1+\lambda\int_0^x\frac{\sinh\sqrt\lambda(x-t)}{\sqrt\lambda}dt$ $=\cosh(\sqrt\lambda x)$.
7 $K_n=\frac{(x-t)^{n-1}}{(n-1)!}$, $R=e^{\lambda(x-t)}$.
8 $\hat k=\frac1{s^2+1}$, $\hat R$ $=\frac{\hat k}{1-\hat k}$ $=\frac1{s^2}$, so $R(x,t;1)=x-t$. Then $y=x+\int_0^x(x-t)\,t\,dt$ $=x+\frac{x^3}6$, in agreement with Exercise IV.4 of §7.1.
9 Multiply $R=\sum\lambda^{n-1}K_n$ by $\lambda K(x,s)$, integrate in $s$ term by term (uniform convergence) and use $\int K(x,s)K_n(s,t)\,ds$ $=K_{n+1}(x,t)$. Conversely, the computation in the proof of Theorem 7.12(3) shows $y=f+\lambda\int Rf$ is a solution (first equation). If $y$ is any solution, multiply $y=f+\lambda\mathcal{K}y$ by $\lambda R(x,s)$ and integrate; the second equation turns $\lambda^2\int R\,\mathcal{K}y$ into $\lambda\int(R-K)y$, which gives $\lambda\int Rf$ $=\lambda\mathcal{K}y=y-f$. So every solution equals $f+\lambda\int Rf$.
III
Symmetric kernels, Hilbert–Schmidt and Mercer.
  • 1 For $K=x+t$ on $[-1,1]$ (Exercise I.2) verify $\sum1/\lambda_n^2=\iint K^2$ and $\sum1/\lambda_n=\int K(x,x)\,dx$.
  • 2 Solve $y(x)=x+\lambda\int_0^1G_0(x,t)\,y(t)\,dt$, $G_0=\min(x,t)(1-\max(x,t))$, by Schmidt's formula, and identify the sum in closed form. For which $\lambda$ is there no solution?
  • 3 Solve $y(x)=x+\lambda\int_{-1}^1(1+xt)\,y(t)\,dt$ by Schmidt's formula; what happens at $\lambda=\tfrac12$ and at $\lambda=\tfrac32$?
  • 4 Prove that a real continuous symmetric kernel with $K_2\equiv0$ is identically zero. Is this true without symmetry?
  • 5 Prove that a real symmetric separable kernel of rank $n$ has exactly $n$ characteristic numbers, counted with multiplicity.
  • 6 Prove that $\min(x,t)$ is a positive kernel on $[0,1]$ by showing $\iint\min(x,t)h(x)h(t)\,dx\,dt$ $=\int_0^1\left(\int_s^1h\right)^2ds$.
  • 7 Check the trace formula for $e^{x+t}$ on $[0,1]$ (Exercise I.1).
  • 8 For $\cos(x-t)$ on $[0,2\pi]$ compute $K_2$, the resolvent, and the bilinear expansion, and check $\sum1/\lambda_n^2=\iint K^2$.
1 $\lambda=\pm\frac{\sqrt3}2$: $\sum\frac1{\lambda^2}$ $=\frac43+\frac43$ $=\frac83$ and $\iint_{[-1,1]^2}(x+t)^2$ $=\frac43+0+\frac43$ $=\frac83$. $\sum\frac1\lambda$ $=\frac2{\sqrt3}-\frac2{\sqrt3}=0$ $=\int_{-1}^12x\,dx$ (exact, since $K$ is separable — although $K$ is not positive).
2 $\lambda_n=n^2\pi^2$, $\varphi_n=\sqrt2\sin n\pi x$, $(x,\varphi_n)$ $=\frac{\sqrt2(-1)^{n+1}}{n\pi}$: $$y=x+\lambda\sum_{n\ge1}\frac{2(-1)^{n+1}\sin n\pi x}{n\pi\,(n^2\pi^2-\lambda)}.$$ The equation is equivalent to $-y''=\lambda y$, $y(0)=0$, $y(1)=1$ (Exercise II.3 of §7.1), so the sum equals $\frac{\sin\sqrt\lambda x}{\sin\sqrt\lambda}$ for $0\lt\lambda$ not a characteristic number (numerically confirmed, e.g. at $\lambda=3$). No solution at $\lambda=n^2\pi^2$, since $(x,\varphi_n)\ne0$ for every $n$.
3 $\lambda_1=\frac12$, $\varphi_1=\frac1{\sqrt2}$; $\lambda_2=\frac32$, $\varphi_2=\sqrt{\tfrac32}\,x$. $(x,\varphi_1)=0$, $(x,\varphi_2)$ $=\sqrt{\tfrac32}\cdot\tfrac23$. So $y=x+\lambda\frac{x}{3/2-\lambda}$ $=\frac{3x}{3-2\lambda}$. At $\lambda=\frac12$: solvable, since $(x,\varphi_1)=0$; $y=\frac32x+C$. At $\lambda=\frac32$: no solution.
4 $0=K_2(x,x)$ $=\int_a^bK(x,s)K(s,x)\,ds$ $=\int_a^bK(x,s)^2\,ds$ for every $x$, so $K\equiv0$. Without symmetry it fails: $\sin x\cos t$ on $[0,\pi]$ has $K_2\equiv0$.
5 By symmetry, $\operatorname{span}\{a_i\}$ $=\operatorname{span}\{b_i\}=:V$, an $n$-dimensional space, and $\mathcal{K}$ maps $C[a,b]$ onto $V$ and vanishes on $V^{\perp}$. So $\mathcal{K}|_V:V\to V$ is a bijective symmetric linear map; by the spectral theorem it has an orthonormal eigenbasis $\varphi_1,\dots,\varphi_n$ with nonzero real eigenvalues $\mu_i$, i.e. characteristic numbers $1/\mu_i$, counted with multiplicity. Any eigenfunction lies in $V$ (it is $\lambda\mathcal{K}y$), so there are no others.
6 $\min(x,t)$ $=\int_0^1\mathbf 1_{\{s\lt x\}}\mathbf 1_{\{s\lt t\}}\,ds$. Fubini: $\iint\min(x,t)h(x)h(t)$ $=\int_0^1\left(\int_s^1h(x)\,dx\right)\left(\int_s^1h(t)\,dt\right)ds\ge0$.
7 $\int_0^1e^{2x}dx$ $=\frac{e^2-1}2$ $=\frac1\lambda$. (One characteristic number, so the trace formula is exact.)
8 $K_2=\int_0^{2\pi}\cos(x-s)\cos(s-t)\,ds$ $=\pi\cos(x-t)$, so $K_n=\pi^{n-1}\cos(x-t)$ and $R=\frac{\cos(x-t)}{1-\pi\lambda}$. Eigenfunctions $\frac{\cos x}{\sqrt\pi}$, $\frac{\sin x}{\sqrt\pi}$, both for $\lambda=\frac1\pi$: $\sum\frac{\varphi_n(x)\varphi_n(t)}{\lambda_n}$ $=\cos x\cos t+\sin x\sin t$ $=\cos(x-t)$ exactly. $\sum\frac1{\lambda_n^2}$ $=2\pi^2$ $=\iint\cos^2(x-t)$.
IV
Prove or disprove.
  • 1 Every continuous kernel on $[a,b]^2$ has at least one characteristic number.
  • 2 If $\lambda$ is a characteristic number of a separable kernel $K$, it is a characteristic number of $K^{T}$.
  • 3 For a real continuous symmetric kernel, the Neumann series converges for $|\lambda|\lt|\lambda_1|$ and diverges (for suitable $f$) for $|\lambda|\gt|\lambda_1|$.
  • 4 If $K$ is real, continuous and symmetric with characteristic numbers $\lambda_n$, then the characteristic numbers of $K_2$ are exactly the $\lambda_n^2$.
  • 5 There is a continuous kernel on $[0,1]^2$ whose only characteristic numbers are $\pm2\sqrt3\,i$.
1 False: $\sin x\cos t$ on $[0,\pi]$, or any Volterra kernel (the existence theorem for Volterra equations in §7.1). It is true for nonzero symmetric kernels (Theorem 7.15(3)).
2 True: the matrix of $K^{T}$ is $A^{T}$, and $\det(I-\lambda A^{T})$ $=\det(I-\lambda A)$ (the Fredholm alternative for separable kernels in §7.1); for general continuous kernels this is Theorem 7.11(2).
3 True. By Theorem 7.17, $\int Rf$ $=\sum\frac{(f,\varphi_n)}{\lambda_n-\lambda}\varphi_n$ $=\sum_k\lambda^k\sum_n\frac{(f,\varphi_n)}{\lambda_n^{k+1}}\varphi_n$, and for $f=\varphi_1$ the Neumann series is $\sum_k\left(\frac\lambda{\lambda_1}\right)^k\frac{\varphi_1}{\lambda_1}$, a geometric series with ratio $\lambda/\lambda_1$.
4 True. $\mathcal{K}^2\varphi_n$ $=\varphi_n/\lambda_n^2$; conversely, if $\mathcal{K}^2y=\mu y$ with $\mu\ne0$, expanding $y=\mu^{-1}\mathcal{K}(\mathcal{K}y)$ by Theorem 7.16(1) gives $(y,\varphi_n)(\mu-\lambda_n^{-2})=0$ for all $n$, so $\mu=\lambda_n^{-2}$ for some $n$. (All are positive.)
5 True: $K=x-t$, with $D(\lambda)=1+\frac{\lambda^2}{12}$ (§7.1, ⚠️ box).
🎯 Exam corner CSIR Part BCSIR Part CGATE MA

E1. CSIR Part B GATE MA The resolvent kernel of $K(x,t)=xt$ on $[0,1]$ is: (A) $\frac{3xt}{3-\lambda}$; (B) $\frac{xt}{1-\lambda}$; (C) $xt\,e^{\lambda/3}$; (D) $\frac{3xt}{3+\lambda}$.

Idea Rank one with $\alpha$ $=\int_0^1t\cdot t\,dt$ $=\frac13$: $K_n=\alpha^{n-1}K$.
Sum $R=xt\sum_{n\ge1}\left(\frac\lambda3\right)^{n-1}$ $=\frac{xt}{1-\lambda/3}$ $=\frac{3xt}{3-\lambda}$. (B) uses $\alpha=1$; (C) confuses the Fredholm geometric series with a Volterra exponential; (D) has the wrong sign.
Answer (A).

E2. CSIR Part B The resolvent kernel of the Volterra equation $y(x)=f(x)+\lambda\int_0^xy(t)\,dt$ is: (A) $e^{\lambda(x-t)}$; (B) $e^{\lambda(x+t)}$; (C) $\frac1{1-\lambda}$; (D) $\cosh\lambda(x-t)$.

Iterate $K_2=\int_t^x1\,ds=x-t$, and $K_n=\frac{(x-t)^{n-1}}{(n-1)!}$.
Sum $R=\sum\frac{\lambda^{n-1}(x-t)^{n-1}}{(n-1)!}$ $=e^{\lambda(x-t)}$, entire in $\lambda$. (C) is the Fredholm resolvent of $K\equiv1$ on $[0,1]$.
Answer (A).

E3. CSIR Part B Let $\lambda_1,\lambda_2,\dots$ be the characteristic numbers of $K(x,t)=e^{-|x-t|}$ on $[0,1]$, repeated according to multiplicity. Then $\sum_n\frac1{\lambda_n^2}$ equals: (A) $\frac{1+e^{-2}}{2}$; (B) $1$; (C) $\frac{1-e^{-2}}{2}$; (D) $1-e^{-1}$.

Idea $K$ is continuous and symmetric, so $\sum1/\lambda_n^2=\iint K^2$ (Theorem 7.16(2)) — no eigenfunctions needed.
Compute $\iint e^{-2|x-t|}$ $=2\int_0^1\!\int_0^xe^{-2(x-t)}\,dt\,dx$ $=\int_0^1(1-e^{-2x})\,dx$ $=1-\frac{1-e^{-2}}{2}$ $=\frac{1+e^{-2}}{2}$. (Option (B) is $\int K(x,x)\,dx=\sum1/\lambda_n$, by Mercer, since this kernel is positive.)
Answer (A).

E4. CSIR Part C Let $K(x,t)=x-t$ on $[0,1]$. Which of the following are true? (One or more options may be correct.) (A) $K$ has no real characteristic number. (B) The Neumann series for $y=f+\lambda\int_0^1(x-t)y(t)\,dt$ converges for every $|\lambda|\lt2\sqrt3$. (C) For every real $\lambda$ and every continuous $f$ the equation has exactly one solution. (D) $K_2\equiv0$.

(A) True $a=(x,1)$, $b=(1,-t)$, $A=\begin{pmatrix}1/2&1\\-1/3&-1/2\end{pmatrix}$, $D=1+\frac{\lambda^2}{12}$, roots $\pm2\sqrt3\,i$.
(B) True The resolvent $R=\frac{-2\left[6\lambda xt-3\lambda(x+t)+2\lambda+6t-6x\right]}{\lambda^2+12}$ is rational with poles at $\pm2\sqrt3\,i$, so its Taylor series in $\lambda$ converges for $|\lambda|\lt2\sqrt3$ (Theorem 7.13). Note the guarantee $|\lambda|B\lt1$ gives only $|\lambda|\lt\sqrt6$, since $B^2=\tfrac16$.
(C) True $D(\lambda)\ne0$ for real $\lambda$ (the Fredholm alternative for separable kernels in §7.1).
(D) False $K_2=\int_0^1(x-s)(s-t)\,ds$ $=\frac{x+t}2-xt-\frac13\not\equiv0$ (e.g. $K_2(0,0)=-\frac13$).
Answer (A), (B), (C).

E5. CSIR Part C Let $K\not\equiv0$ be a real continuous symmetric kernel on $[a,b]^2$, with orthonormal eigenfunctions $\varphi_n$. Which of the following are necessarily true? (A) $K$ has at least one characteristic number. (B) All characteristic numbers of $K$ are real. (C) Every $f\in C[a,b]$ satisfies $f=\sum(f,\varphi_n)\varphi_n$ uniformly. (D) If $\lambda$ is a characteristic number of $K$, then $\lambda^2$ is a characteristic number of $K_2$.

(A) True Theorem 7.15(3).
(B) True Theorem 7.15(1).
(C) False Only sourcewise representable functions $f=\mathcal{K}h$ are guaranteed. For $K=xt$ on $[0,1]$ the eigenfunctions are just $\sqrt3x$, and $f=1$ is not $\frac32x$.
(D) True If $\varphi=\lambda\mathcal{K}\varphi$ then $\mathcal{K}^2\varphi$ $=\lambda^{-2}\varphi$, i.e. $\varphi=\lambda^2\int K_2\varphi$.
Answer (A), (B), (D).

E6. CSIR Part C Consider $y(x)=f(x)+\lambda\int_0^xe^{x-t}y(t)\,dt$ on $[0,1]$. Which of the following are true? (A) The resolvent kernel is $e^{(1+\lambda)(x-t)}$. (B) For $\lambda=-1$ the solution is $y=f$. (C) For $\lambda=2$ and $f\equiv1$ the solution is $y=\frac{1+2e^{3x}}{3}$. (D) For some $\lambda$ the homogeneous equation has a nonzero continuous solution.

(A) True Example 4.
(B) False For $\lambda=-1$, $R\equiv1$ and $y=f-\int_0^xf(t)\,dt$, which differs from $f$ unless $f\equiv0$ (e.g. $f=1$ gives $y=1-x$).
(C) True $y=1+2\int_0^xe^{3(x-t)}\,dt$ $=1+\frac23(e^{3x}-1)$ $=\frac{1+2e^{3x}}{3}$.
(D) False Volterra kernels have no characteristic numbers (the existence theorem for Volterra equations in §7.1; Exercise I.8).
Answer (A), (C).
🧑‍🏫 For the teacher

Suggested plan — 3 lectures × 60 min.

  1. Characteristic numbers. Definition 7.4 and the two meanings of "eigenvalue"; Theorem 7.10 with full proof (Cauchy–Schwarz, then Bessel); Examples 1 and 2 on the board. State Fredholm's theorems (7.11) and explain the proof idea "polynomial plus small" in five minutes.
  2. Neumann series and resolvents. Definition 7.5, Theorem 7.12 with proof; Example 3 (three radii); project Interactive 1 and move $\lambda$ across $1/B$ and then $|\lambda_1|$ — let the class watch the series diverge while the dashed solution stays perfectly finite. Theorem 7.14 and Example 4; the strategy box.
  3. Symmetric kernels. Theorem 7.15(1), (2), (5) with proofs, (3) as a sketch; Hilbert–Schmidt and Mercer (7.16) with the proof of the trace identity; Interactive 2 (start with $t_0$ near the end point, then at $t_0=0.5$); Examples 5–7; finish with E4 and E5 live.

Discussion questions. (1) For matrices, a symmetric matrix always has real eigenvalues and an orthonormal eigenbasis. Which part of this survives for symmetric kernels, and which part fails (think of $xt$)? (2) Why is the radius of convergence of the Neumann series the distance to the nearest characteristic number, and not $1/B$? (3) The kernel $\min(x,t)$ is the covariance of Brownian motion. What does Mercer's theorem say about random functions?

Misconceptions to watch for. Confusing $\lambda$ with $1/\lambda$; believing every kernel has characteristic numbers, or that they are always real; treating the sufficient condition $|\lambda|M(b-a)\lt1$ as the exact range; thinking that divergence of the Neumann series means there is no solution; expanding an arbitrary $f$ in eigenfunctions of a kernel of finite rank; counting a multiple characteristic number once in $\sum1/\lambda_n^2$; computing $K_2$ as $K^2$.

Suggested homework. I(1, 2, 3, 7), II(1, 3, 5, 6, 8), III(1, 3, 6, 8), IV(1, 3, 5); Exam corner E1, E3, E4, E6.

🔬 Research–Teaching NexusResearch-tutored · optional, about 60 minutes

Research question. Brownian motion $W(x)$, $0\le x\le1$ — the model behind stock prices in option pricing, diffusing molecules and noise in electronic circuits — has covariance $\mathbb{E}[W(x)W(t)]=\min(x,t)$. How many eigenfunctions of this kernel are needed to reproduce Brownian paths, and how well does a computer recover its characteristic numbers?

Why it matters. Mercer's theorem for a covariance kernel becomes the Karhunen–Loève expansion $W(x)=\sum_nZ_n\frac{\varphi_n(x)}{\sqrt{\lambda_n}}$ with independent standard normal $Z_n$ — the continuous version of principal component analysis, used in signal compression, uncertainty quantification in engineering, and in generating random fields for climate and groundwater models.

What to do.

  1. Using Example 2, write the Karhunen–Loève expansion of $W$, and check that $\mathbb{E}[W(x)W(t)]$ $=\sum_n\frac{\varphi_n(x)\varphi_n(t)}{\lambda_n}$ $=\min(x,t)$.
  2. The fraction of the total variance $\int_0^1\mathbb{E}[W(x)^2]\,dx$ $=\tfrac12$ captured by the first $N$ terms is $\frac{2}{1}\sum_{n\le N}\frac1{\lambda_n}$. Compute it for $N=1,2,3,5,10$.
  3. Simulate paths with $N=5$ and $N=100$ terms (seeded random numbers), and compare them with a scaled random walk with $1000$ steps.
  4. Nyström's method. Replace $\int_0^1$ by the trapezoidal rule on $m$ equally spaced points, so that the integral equation becomes a matrix eigenvalue problem, and compute the smallest characteristic number for $m=11,21,41,81$. How fast does it approach $\pi^2/4$?

Tools: Python with NumPy (random numbers and numpy.linalg.eigvalsh), or any language with an eigenvalue routine.

Hand in: The variance table, two plots of simulated paths, and a table of Nyström approximations with their errors and your estimate of the order of convergence.

Going further. Find the eigenfunctions of the Brownian bridge covariance $\min(x,t)-xt$ (it is $G_0$!), and of the exponential covariance $e^{-|x-t|}$ used for random fields, whose characteristic numbers solve a transcendental equation.

Instructor note: the variance fractions are $8/\pi^2\approx81.1\%$ ($N=1$), $90.1\%$, $93.3\%$, $96.0\%$ ($N=5$) and $98.0\%$ ($N=10$); the tail decays like $1/N$, which is why Brownian paths look rough. The trapezoidal Nyström values of $\lambda_1$ are about $2.4623$, $2.4661$, $2.4671$, $2.4673$ for $m=11,21,41,81$, with errors $\approx-5.1\times10^{-3}$, $-1.3\times10^{-3}$, $-3.2\times10^{-4}$, $-7.9\times10^{-5}$: each halving of the step divides the error by $4$, i.e. order $h^2$, the order of the trapezoidal rule (§5.4).

🧪 Self-check quiz

0/5 correct

Q1. The characteristic number of $K(x,t)=xt$ on $[0,1]$ is

Q2. Which condition guarantees that the Neumann series of a continuous kernel converges?

Q3. The resolvent kernel of a continuous Volterra kernel, as a function of $\lambda$, is

Q4. For a real continuous symmetric kernel, the characteristic numbers are always

Q5. For a real continuous symmetric kernel with characteristic numbers $\lambda_n$ (repeated by multiplicity), $\sum_n1/\lambda_n^2$ equals