Chapter 7 · Linear Integral Equations

7.1 Fredholm and Volterra Equations; Separable Kernels

Linear integral equations of Fredholm and Volterra type: classification, the passage from initial and boundary value problems to integral equations and back, separable kernels and the Fredholm alternative, successive approximations for Volterra equations, convolution kernels, the Laplace transform and Abel's equation.

An initial value problem such as $y''+q(x)\,y=0$, $y(0)=1$, $y'(0)=0$ asks for a function whose second derivative is prescribed in terms of the function itself. Integrating twice turns it into a single equation, $$y(x)=1-\int_0^x(x-t)\,q(t)\,y(t)\,dt,$$ in which the unknown function appears under an integral sign and the initial data are built into the right-hand side. A boundary value problem, solved with its Green's function from §2.4, turns into an equation of the same shape in which the integral runs over the whole interval. Equations of this kind — linear integral equations — are often easier to analyse than the differential equations they come from: they need no differentiability of the unknown, they contain the side conditions automatically, and they can be attacked by iteration, by linear algebra and by transforms. They also arise directly, without any differential equation behind them: in Abel's problem of recovering the shape of a slide from the descent times, in population models in which births depend on the whole age structure, in potential theory, in tomography and in the theory of materials with memory.

This section sets up the basic theory. We classify integral equations (first and second kind, Fredholm and Volterra), convert initial and boundary value problems into integral equations and back, and solve two large classes completely. For separable kernels $K(x,t)=\sum a_i(x)b_i(t)$ the equation reduces to a linear system $(I-\lambda A)c=F$, and the familiar dichotomy of linear algebra becomes the Fredholm alternative. For Volterra equations of the second kind, successive approximations converge for every value of the parameter, so the solution always exists and is unique; for convolution kernels the Laplace transform gives it in closed form. CSIR NET questions on this topic are short and frequent: an equivalent integral equation for an IVP, the characteristic numbers of a kernel like $x+t$ or $\cos(x+t)$, the solution of a convolution equation, and Part C statements about when solutions exist.

Classification of Linear Integral Equations

Throughout, $[a,b]$ is a bounded interval, $f$ is a given continuous function, $\lambda$ is a real or complex parameter, and the kernel $K(x,t)$ is a given function of two variables. We look for a continuous unknown function $y$.

📘 Definition 7.1 — Linear integral equations: kind, type, homogeneity

Let $K$ be continuous on the square $Q=[a,b]\times[a,b]$ (Fredholm case) or on the triangle $T=\{(x,t):a\le t\le x\le b\}$ (Volterra case), and let $f\in C[a,b]$.

Fredholm, first kind$\displaystyle\int_a^bK(x,t)\,y(t)\,dt=f(x)$
Fredholm, second kind$\displaystyle y(x)$ $\displaystyle$ $\displaystyle =f(x)+\lambda\int_a^bK(x,t)\,y(t)\,dt$
Volterra, first kind$\displaystyle\int_a^xK(x,t)\,y(t)\,dt=f(x)$
Volterra, second kind$\displaystyle y(x)$ $\displaystyle$ $\displaystyle =f(x)+\lambda\int_a^xK(x,t)\,y(t)\,dt$
  • In a Fredholm equation the limits of integration are fixed; in a Volterra equation the upper limit is the variable $x$.
  • In an equation of the first kind the unknown appears only under the integral; in one of the second kind it also appears outside. (An equation $h(x)y(x)=f(x)+\lambda\int K y$ with $h$ vanishing somewhere is sometimes said to be of the third kind; we shall not need it.)
  • A second-kind equation is homogeneous if $f\equiv0$, and non-homogeneous otherwise.
  • The equation is linear because the unknown enters linearly; $y(x)=1+\int_0^xy(t)^2\,dt$ is a nonlinear Volterra equation. It is singular if the interval is infinite or the kernel is unbounded, as in Abel's equation $\int_0^x\frac{y(t)}{\sqrt{x-t}}\,dt=f(x)$.

A Volterra equation is a Fredholm equation in disguise: extend the kernel by $K(x,t)=0$ for $t\gt x$, and $\int_a^x K(x,t)y(t)\,dt$ $=\int_a^bK(x,t)y(t)\,dt$. The extended kernel is in general discontinuous on the diagonal, but it is bounded, and this "triangular" structure is exactly what makes Volterra equations so much better behaved (Theorem 7.7). The two kinds behave very differently. A second-kind equation is a perturbation of the identity, $y-\lambda\mathcal{K}y=f$, and behaves like a well-conditioned linear system. A first-kind equation $\mathcal{K}y=f$ is typically ill-posed: integration smooths, so $f$ must already be smooth to be in the range, and tiny wiggles in $f$ correspond to huge wiggles in $y$. For Volterra equations there is a way out — differentiate.

📐 Theorem 7.1 — Volterra equations of the first kind become equations of the second kind
Suppose $K$ and $\partial K/\partial x$ are continuous on $T$, $K(x,x)\ne0$ for all $x\in[a,b]$, and $f\in C^1[a,b]$. Then a function $y\in C[a,b]$ satisfies $$\int_a^xK(x,t)\,y(t)\,dt=f(x)\qquad(a\le x\le b)$$ if and only if $f(a)=0$ and $$y(x)+\int_a^x\frac{K_x(x,t)}{K(x,x)}\,y(t)\,dt=\frac{f'(x)}{K(x,x)}.$$ Consequently (by Theorem 7.7 below) the first-kind equation has exactly one continuous solution if $f(a)=0$ and none if $f(a)\ne0$.

Proof. We use the Leibniz rule (Unit 1): if $K$ and $K_x$ are continuous on $T$ and $y$ is continuous, then $\Phi(x)=\int_a^xK(x,t)y(t)\,dt$ is continuously differentiable with $$\Phi'(x)=K(x,x)\,y(x)+\int_a^xK_x(x,t)\,y(t)\,dt.$$ If $\Phi=f$ on $[a,b]$, then $f(a)=\Phi(a)=0$, and differentiating gives $K(x,x)y(x)+\int_a^xK_xy\,dt=f'(x)$; dividing by $K(x,x)\ne0$ gives the second-kind equation. Conversely, if $f(a)=0$ and $y$ solves the second-kind equation, multiplying by $K(x,x)$ shows $\Phi'(x)=f'(x)$ on $[a,b]$; since also $\Phi(a)=0=f(a)$, the mean value theorem gives $\Phi\equiv f$. The last sentence follows because the second-kind equation has exactly one continuous solution. ∎

If $K(x,x)\equiv0$ but $K_x(x,x)\ne0$ (for instance $K=x-t$), differentiate twice; each differentiation adds a compatibility condition ($f(a)=0$, then $f'(a)=0$, …).

✏️ Example 1 — Classifying integral equations, and a first-kind equation solved
Classify each equation, and solve (b). (a) $y(x)=x+\int_0^1xt\,y(t)\,dt$. (b) $\int_0^xe^{x-t}y(t)\,dt=\sin x$. (c) $y(x)=\lambda\int_0^{\pi}\sin(x+t)\,y(t)\,dt$. (d) $\int_0^x\frac{y(t)}{\sqrt{x-t}}\,dt=x$. (e) $y(x)=1+\int_0^xy(t)^2\,dt$.
(a) Fredholm, second kind, non-homogeneous, $\lambda=1$, kernel $K=xt$ (continuous, symmetric and separable).
(b) Type Volterra, first kind, convolution kernel $e^{x-t}$. Here $K(x,x)=1\ne0$, $K_x=e^{x-t}$ and $f(0)=\sin0=0$, so Theorem 7.1 gives the equivalent second-kind equation $$y(x)+\int_0^xe^{x-t}y(t)\,dt=\cos x.$$
(b) Solution Write the original equation as $e^x\,v(x)=\sin x$ with $v(x)=\int_0^xe^{-t}y(t)\,dt$. Then $v(x)=e^{-x}\sin x$ and $e^{-x}y(x)$ $=v'(x)$ $=e^{-x}(\cos x-\sin x)$, so $$y(x)=\cos x-\sin x.$$ Check: $\frac{d}{dt}\left(e^{-t}\sin t\right)$ $=e^{-t}(\cos t-\sin t)$, hence $\int_0^xe^{x-t}(\cos t-\sin t)\,dt$ $=e^x\cdot e^{-x}\sin x$ $=\sin x$.
(c), (d) (c) Fredholm, second kind, homogeneous, with symmetric separable kernel $\sin(x+t)$ $=\sin x\cos t+\cos x\sin t$; it has nonzero solutions only for special $\lambda$ (Exercise III.6). (d) Abel's equation: Volterra, first kind, singular (the kernel $(x-t)^{-1/2}$ is unbounded); we solve it in Theorem 7.9.
(e) Volterra, second kind, but nonlinear. It is the integral form of $y'=y^2$, $y(0)=1$, whose solution $y=\frac1{1-x}$ blows up at $x=1$ (§1.1). Linear Volterra equations never blow up (Theorem 7.7).

From Differential Equations to Integral Equations and Back

The key tool for initial value problems is the formula that collapses repeated integrals into a single one.

📐 Theorem 7.2 — Cauchy's formula for repeated integration
Let $g\in C[a,b]$ and $n\in\mathbb{N}$. The $n$-fold integral $$I_ng(x)=\int_a^x\int_a^{x_{n-1}}\cdots\int_a^{x_1}g(t)\,dt\,dx_1\cdots dx_{n-1}$$ equals $$I_ng(x)=\frac{1}{(n-1)!}\int_a^x(x-t)^{n-1}g(t)\,dt.$$ Equivalently, $u(x)=\frac{1}{(n-1)!}\int_a^x(x-t)^{n-1}g(t)\,dt$ is the unique solution of $u^{(n)}=g$, $u(a)=u'(a)=\dots=u^{(n-1)}(a)=0$.

Proof. Let $u$ be the right-hand side. For $n=1$ it is $\int_a^xg$. For $n\ge2$ the Leibniz rule gives $u'(x)=\frac{1}{(n-2)!}\int_a^x(x-t)^{n-2}g(t)\,dt$ (the boundary term vanishes because $(x-t)^{n-1}=0$ at $t=x$). Repeating, $u^{(k)}(x)$ $=\frac{1}{(n-1-k)!}\int_a^x(x-t)^{n-1-k}g(t)\,dt$ for $k\le n-1$, so $u^{(n-1)}=\int_a^xg$, $u^{(n)}=g$, and $u^{(k)}(a)=0$ for $k\le n-1$. The $n$-fold integral $I_ng$ has the same properties, and a function with $n$-th derivative $g$ and vanishing derivatives of orders $0,\dots,n-1$ at $a$ is unique (the difference is a polynomial of degree $\lt n$ with all derivatives zero at $a$). ∎

📐 Theorem 7.3 — Initial value problems are equivalent to Volterra equations
Let $p,q,g\in C[a,b]$ and consider the initial value problem (IVP) $$y''+p(x)\,y'+q(x)\,y=g(x),$$ $$y(a)=\alpha,\qquad y'(a)=\beta.$$
  1. (Equation for $y$.) If $p\in C^1[a,b]$, a function $y\in C^2[a,b]$ solves (IVP) if and only if $y\in C[a,b]$ solves $$y(x)=F(x)+\int_a^xK(x,t)\,y(t)\,dt,$$ where $$K(x,t)=(x-t)\big[p'(t)-q(t)\big]-p(t),$$ $$F(x)=\alpha+\big[\beta+p(a)\alpha\big](x-a)+\int_a^x(x-t)\,g(t)\,dt.$$ In particular, for $y''+q(x)y=g$ the kernel is simply $-(x-t)\,q(t)$.
  2. (Equation for $u=y''$.) With no extra smoothness of $p$: $y$ solves (IVP) if and only if $u=y''$ is continuous and solves $$u(x)=g(x)-\beta p(x)-q(x)\big[\alpha+\beta(x-a)\big]-\int_a^x\big[p(x)+q(x)(x-t)\big]u(t)\,dt,$$ and then $y(x)=\alpha+\beta(x-a)+\int_a^x(x-t)\,u(t)\,dt$.
The same procedure turns an $n$-th order linear IVP into a Volterra equation; for $u=y^{(n)}$ the kernel is $-\sum_{k=0}^{n-1}p_k(x)\,\frac{(x-t)^{n-1-k}}{(n-1-k)!}$.

Proof. (1) Suppose $y$ solves (IVP). Integrate $y''=g-py'-qy$ from $a$ to $x$, integrating the term $py'$ by parts: $$y'(x)-\beta=\int_a^xg\,dt-\Big[p\,y\Big]_a^x+\int_a^xp'(t)y(t)\,dt-\int_a^xq(t)y(t)\,dt,$$ so $$y'(x)=\beta+p(a)\alpha-p(x)y(x)+\int_a^x\big[p'(t)-q(t)\big]y(t)\,dt+\int_a^xg\,dt.$$ Integrate once more from $a$ to $x$ and use Theorem 7.2 ($n=2$) on the two integrals: $$y(x)=\alpha+\big[\beta+p(a)\alpha\big](x-a)-\int_a^xp(t)y(t)\,dt+\int_a^x(x-t)\big[p'(t)-q(t)\big]y(t)\,dt+\int_a^x(x-t)g(t)\,dt,$$ which is the stated equation.

Conversely let $y\in C[a,b]$ solve the integral equation. The right-hand side is $C^1$ (Leibniz rule), so $y\in C^1$, and since $K(x,x)=-p(x)$ and $K_x(x,t)=p'(t)-q(t)$, $$y'(x)=\beta+p(a)\alpha+\int_a^xg\,dt-p(x)y(x)+\int_a^x\big[p'(t)-q(t)\big]y(t)\,dt.$$ The right-hand side is again $C^1$ (now that $y\in C^1$ and $p\in C^1$), so $y\in C^2$ and $y''=g-p'y-py'+(p'-q)y=g-py'-qy$. Finally $y(a)=F(a)=\alpha$ and $y'(a)=\beta+p(a)\alpha-p(a)\alpha$ $=\beta$.

(2) If $y$ solves (IVP) and $u=y''$, then by Theorem 7.2, $y'=\beta+\int_a^xu$ and $y=\alpha+\beta(x-a)+\int_a^x(x-t)u\,dt$; substituting into $u=g-py'-qy$ gives the equation for $u$. Conversely, if $u$ is a continuous solution, define $y$ by the last formula; then $y''=u$, $y(a)=\alpha$, $y'(a)=\beta$, and the equation for $u$ says exactly that $y''=g-py'-qy$. ∎

The integral equation is a complete substitute for the IVP: the initial data sit in $F$, and by Theorem 7.7 it has exactly one continuous solution on the whole of $[a,b]$ — a second proof of the global existence theorem for linear equations (§2.1).

✏️ Example 2 — One IVP, two Volterra equations
Convert $y''+x\,y'+y=0$, $y(0)=1$, $y'(0)=0$ into a Volterra equation (a) for $y$, (b) for $u=y''$. Then solve the problem and check both equations.
(a) $p(x)=x$, $q=1$, $g=0$, $a=0$, $\alpha=1$, $\beta=0$. Then $K(x,t)$ $=(x-t)(p'(t)-q(t))-p(t)$ $=(x-t)(1-1)-t=-t$ and $F(x)=1+(0+0\cdot1)x=1$. So $$y(x)=1-\int_0^xt\,y(t)\,dt.$$ The integration by parts has absorbed the whole $(x-t)$-term.
(b) $K(x,t)$ $=-[p(x)+q(x)(x-t)]$ $=-(x+x-t)=t-2x$ and $F(x)=g-\beta p-q(\alpha+\beta x)=-1$: $$u(x)=-1+\int_0^x(t-2x)\,u(t)\,dt.$$
Solve From (a), differentiating: $y'=-xy$, $y(0)=1$, so $y=e^{-x^2/2}$. (Directly: $y''+xy'+y=(y'+xy)'$, so $y'+xy$ is constant, equal to $0$ at $x=0$.)
Check (a) $\int_0^xte^{-t^2/2}dt$ $=1-e^{-x^2/2}$, so $1-(1-e^{-x^2/2})=y$. (b) $u=y''=(x^2-1)e^{-x^2/2}$; a computer-algebra (or by-hand) evaluation of $-1+\int_0^x(t-2x)(t^2-1)e^{-t^2/2}dt$ returns $(x^2-1)e^{-x^2/2}$. Equation (a) is clearly the better one to work with: always try the $y$-form first when $p$ is differentiable.

For boundary value problems the right tool is the Green's function of §2.4. There we worked with $L[u]=(p\,u')'+q\,u$ on $[a,b]$ ($p\in C^1$, $p\gt0$, $q$ continuous) and separated homogeneous conditions $B_a[u]$ $=\alpha_1u(a)+\alpha_2u'(a)=0$, $B_b[u]=\beta_1u(b)+\beta_2u'(b)=0$, and proved: if the homogeneous problem has only the trivial solution, there is a unique, symmetric Green's function $G(x,t)$, and for every $h\in C[a,b]$ the function $u=\int_a^bG(x,t)h(t)\,dt$ is the unique solution of $L[u]=h$ with the boundary conditions.

📐 Theorem 7.4 — Boundary value problems are equivalent to Fredholm equations
Let $L,B_a,B_b$ be as above, suppose the homogeneous problem $L[u]=0$, $B_a[u]=B_b[u]=0$ has only the trivial solution, and let $G$ be its Green's function. Let $r,h\in C[a,b]$. Then $y\in C^2[a,b]$ solves $$L[y]+\lambda\,r(x)\,y=h(x),\qquad B_a[y]=B_b[y]=0$$ if and only if $y\in C[a,b]$ solves the Fredholm equation of the second kind $$y(x)=\int_a^bG(x,t)\,h(t)\,dt-\lambda\int_a^bG(x,t)\,r(t)\,y(t)\,dt.$$ In particular the values of $\lambda$ for which the homogeneous boundary value problem $L[y]+\lambda ry=0$ has nontrivial solutions are exactly the values for which $y=-\lambda\int G\,r\,y$ has nontrivial solutions.

Proof. If $y\in C^2$ solves the boundary value problem, then $L[y]=h-\lambda ry$ with $h-\lambda ry$ continuous, and the solution formula of §2.4 gives $y=\int_a^bG\,(h-\lambda ry)\,dt$. Conversely, if $y$ is continuous and satisfies the integral equation, put $\tilde h=h-\lambda ry\in C[a,b]$. The equation says $y=\int_a^bG\tilde h\,dt$, and by §2.4 this function is $C^2$, satisfies both boundary conditions, and $L[y]=\tilde h=h-\lambda ry$. ∎

Two remarks. First, the kernel $G(x,t)r(t)$ is not symmetric unless $r$ is constant; if $r\gt0$, the substitution $z=\sqrt{r}\,y$ produces the symmetric kernel $\sqrt{r(x)}\,G(x,t)\sqrt{r(t)}$, which is how Sturm–Liouville problems with a weight become symmetric integral equations (§7.2). Second, the hypothesis on the homogeneous problem matters: for Neumann conditions $u'(0)=u'(1)=0$ the operator $u''$ has no Green's function (constants solve the homogeneous problem), and one must first move a multiple of $y$ to the other side (Exercise II.5).

✏️ Example 3 — A boundary value problem with a free end
Convert $y''+\lambda y=x$, $y(0)=0$, $y'(1)=0$, into a Fredholm equation of the second kind.
Green's function For $L[u]=u''$ ($p=1$, $q=0$) with $u(0)=0$, $u'(1)=0$: the homogeneous problem has only $u=0$. Take $u_1=x$ (satisfies $u_1(0)=0$) and $u_2=1$ (satisfies $u_2'(1)=0$); $C=p\,(u_1u_2'-u_1'u_2)=-1$. By the construction of §2.4, $$G(x,t)=\frac{u_1(\min(x,t))\,u_2(\max(x,t))}{C}=-\min(x,t).$$ Check with $h\equiv1$: $-\int_0^1\min(x,t)\,dt$ $=-\left(\tfrac{x^2}2+x(1-x)\right)$ $=\tfrac{x^2}{2}-x$, whose second derivative is $1$, which vanishes at $0$ and has derivative $0$ at $1$.
Apply Theorem 7.4 Here $r=1$, $h=x$: $y=\int G\,t\,dt-\lambda\int G\,y\,dt$, i.e. $$y(x)=-\int_0^1\min(x,t)\,t\,dt+\lambda\int_0^1\min(x,t)\,y(t)\,dt.$$
The free term $-\int_0^1\min(x,t)\,t\,dt$ $=-\left[\int_0^xt^2\,dt+x\int_x^1t\,dt\right]$ $=-\left[\tfrac{x^3}{3}+\tfrac{x(1-x^2)}{2}\right]$ $=\tfrac{x^3}{6}-\tfrac{x}{2}$. So $$y(x)=\frac{x^3}{6}-\frac x2+\lambda\int_0^1\min(x,t)\,y(t)\,dt.$$
Check $F=\tfrac{x^3}6-\tfrac x2$ satisfies $F''=x$, $F(0)=0$, $F'(1)=\tfrac12-\tfrac12=0$: it is the solution for $\lambda=0$, as it must be. The kernel $\min(x,t)$ is symmetric and reappears in §7.2, where we find its characteristic numbers $\left(\tfrac{(2n-1)\pi}{2}\right)^2$.
✏️ Example 4 — From integral equations back to differential equations
(a) Let $G_0(x,t)$ $=\min(x,t)\,\big(1-\max(x,t)\big)$ on $[0,1]^2$. Show that $y(x)=\lambda\int_0^1G_0(x,t)\,y(t)\,dt$ is equivalent to $y''+\lambda y=0$, $y(0)=y(1)=0$, and find all $\lambda$ for which it has a nonzero solution. (b) Find the initial value problem equivalent to $y(x)=x-\int_0^x(x-t)\,y(t)\,dt$ and solve it.
(a) Split $G_0=t(1-x)$ for $t\le x$ and $x(1-t)$ for $t\ge x$, so $$y(x)=\lambda\left[(1-x)\int_0^xt\,y(t)\,dt+x\int_x^1(1-t)\,y(t)\,dt\right].$$ A continuous $y$ makes the right side $C^1$, hence $y\in C^1$; then the right side is $C^2$, so $y\in C^2$.
Differentiate by the Leibniz rule the two boundary terms $\pm\lambda x(1-x)y(x)$ cancel, so $$y'=\lambda\left[-\int_0^xty\,dt+\int_x^1(1-t)y\,dt\right],$$ and $y''=\lambda\left[-xy(x)-(1-x)y(x)\right]$ $=-\lambda y$.
Hidden conditions Put $x=0$: both integrals are multiplied by $0$ or taken over $[0,0]$, so $y(0)=0$; similarly $y(1)=0$. Conversely Theorem 7.4 (with $L[u]=u''$, whose Dirichlet Green's function is $-G_0$, and $r=1$, $h=0$) shows the BVP implies the integral equation. Nonzero solutions: $y''+\lambda y=0$, $y(0)=y(1)=0$ forces $\lambda=n^2\pi^2$, $y=\sin n\pi x$ (§2.3).
(b) Differentiate twice (Leibniz): $y'=1-\int_0^xy\,dt$, $y''=-y$. At $x=0$: $y(0)=0$, $y'(0)=1$. So $y''+y=0$, $y(0)=0$, $y'(0)=1$, and $y=\sin x$. Check: $\int_0^x(x-t)\sin t\,dt=x-\sin x$, and $x-(x-\sin x)=\sin x$.
🧭 Problem-solving strategy — converting between differential and integral equations
  1. IVP → Volterra. Solve for the highest derivative, integrate $n$ times from the initial point, integrate terms like $p\,y'$ by parts, and collapse repeated integrals with $\int_a^x\!\int_a^s g$ $=\int_a^x(x-t)g(t)\,dt$. The initial values go into the free term; the kernel never contains them.
  2. BVP → Fredholm. Find the Green's function $G$ of the part of the operator you keep on the left (it must have no nontrivial homogeneous solution — shift a multiple of $y$ across if necessary), then write $y=\int G\cdot(\text{everything else})$. Inhomogeneous boundary values: subtract a function satisfying them first (Exercise II.3).
  3. Integral → differential. Split the kernel at $t=x$, differentiate with the Leibniz rule until the integrals disappear, and read off the initial or boundary conditions by putting $x=a$ (and $x=b$) in the equation and in its derivatives.
  4. Check by putting $\lambda=0$: the free term must solve the problem with the $\lambda$-term removed.

Separable Kernels and the Fredholm Alternative

When the kernel is a finite sum of products, the integral $\int_a^bK(x,t)y(t)\,dt$ is a combination of finitely many fixed functions of $x$, with coefficients that are numbers depending on $y$. The equation then reduces to finitely many unknowns.

📘 Definition 7.2 — Separable (degenerate) kernels
A kernel $K$ on $[a,b]^2$ is separable (or degenerate, or of finite rank) if $$K(x,t)=\sum_{i=1}^na_i(x)\,b_i(t)$$ with $a_i,b_i\in C[a,b]$. Grouping terms we may always assume that $a_1,\dots,a_n$ are linearly independent and $b_1,\dots,b_n$ are linearly independent; the smallest such $n$ is the rank of $K$. Associated with this representation are the $n\times n$ matrix $A$ and, for given $f$, the vector $F$: $$A_{ij}=\int_a^bb_i(t)\,a_j(t)\,dt,\qquad F_i=\int_a^bb_i(t)\,f(t)\,dt.$$

Examples: $xt$ (rank $1$), $x+t=x\cdot1+1\cdot t$ (rank $2$), $1+3xt$, $\cos(x-t)$ $=\cos x\cos t+\sin x\sin t$, $\sin(x+t)$, $e^{x+t}=e^xe^t$, and every polynomial in $x$ and $t$. The kernels $e^{xt}$, $\min(x,t)$ and $\frac1{1+(x-t)^2}$ are not separable.

📐 Theorem 7.5 — Separable kernels: reduction to a linear system
Let $K=\sum_{i=1}^na_i(x)b_i(t)$ be separable, $f\in C[a,b]$, and $A$, $F$ as in Definition 7.2.
  1. If $y\in C[a,b]$ solves $y(x)=f(x)+\lambda\int_a^bK(x,t)y(t)\,dt$, then the numbers $c_i=\int_a^bb_i(t)y(t)\,dt$ satisfy $$(I-\lambda A)\,c=F,\qquad\text{and}\qquad y(x)=f(x)+\lambda\sum_{i=1}^nc_i\,a_i(x).$$
  2. Conversely, if $c\in\mathbb{C}^n$ solves $(I-\lambda A)c=F$, then $y=f+\lambda\sum c_ia_i$ solves the integral equation, and $\int_a^b b_iy\,dt=c_i$.
The maps $y\mapsto c$ and $c\mapsto y$ are inverse bijections between the solutions of the integral equation and the solutions of the linear system.

Proof. (1) $\int_a^bK(x,t)y(t)\,dt$ $=\sum_ia_i(x)\int_a^bb_i(t)y(t)\,dt$ $=\sum_ic_ia_i(x)$, so $y=f+\lambda\sum_jc_ja_j$. Multiply this identity by $b_i$ and integrate: $c_i=F_i+\lambda\sum_jA_{ij}c_j$, i.e. $(I-\lambda A)c=F$. (2) If $(I-\lambda A)c=F$ and $y=f+\lambda\sum_jc_ja_j$, then $\int_a^bb_iy\,dt$ $=F_i+\lambda\sum_jA_{ij}c_j=c_i$; hence $\int_a^bKy\,dt=\sum_ia_ic_i$ and $f+\lambda\int Ky$ $=f+\lambda\sum c_ia_i=y$. Starting from $y$, (1) produces $c$ and the formula $y=f+\lambda\sum c_ia_i$ returns the same $y$; starting from $c$, (2) produces $y$ with $\int b_iy=c_i$. So the maps are mutually inverse. ∎

To state the alternative we need two definitions that §7.2 develops much further.

📘 Definition 7.3 — Characteristic numbers, eigenfunctions and the transposed equation
Let $K$ be continuous on $[a,b]^2$. A number $\lambda$ is a characteristic number of $K$ if the homogeneous equation $$y(x)=\lambda\int_a^bK(x,t)\,y(t)\,dt$$ has a solution $y\in C[a,b]$, $y\not\equiv0$; such a $y$ is an eigenfunction belonging to $\lambda$. ($\lambda=0$ is never a characteristic number, since it forces $y=0$.) The transposed kernel is $K^{T}(x,t)=K(t,x)$, and the transposed (adjoint) homogeneous equation is $$z(x)=\lambda\int_a^bK(t,x)\,z(t)\,dt.$$ Many books call a characteristic number an "eigenvalue"; the eigenvalue of the integral operator $y\mapsto\int Ky$ is then $\mu=1/\lambda$.
📐 Theorem 7.6 — The Fredholm alternative for separable kernels
Let $K=\sum_{i=1}^na_i(x)b_i(t)$ with $\{a_i\}$ and $\{b_i\}$ each linearly independent, and put $D(\lambda)=\det(I-\lambda A)$.
  1. $D$ is a polynomial of degree at most $n$ with $D(0)=1$. The characteristic numbers of $K$ are exactly the roots of $D$; there are at most $n$ of them, and there may be none, or only non-real ones.
  2. If $D(\lambda)\ne0$, then for every $f\in C[a,b]$ the equation $y=f+\lambda\int_a^bKy\,dt$ has exactly one solution, namely $$y(x)=f(x)+\lambda\int_a^bR(x,t;\lambda)\,f(t)\,dt,\qquad R(x,t;\lambda)=\sum_{i,j=1}^na_i(x)\big[(I-\lambda A)^{-1}\big]_{ij}\,b_j(t).$$
  3. If $D(\lambda)=0$, then the homogeneous equation has exactly $r=n-\operatorname{rank}(I-\lambda A)\ge1$ linearly independent solutions; the transposed homogeneous equation also has exactly $r$ linearly independent solutions $z_1,\dots,z_r$; and the non-homogeneous equation is solvable if and only if $$\int_a^bf(x)\,z_k(x)\,dx=0\qquad(k=1,\dots,r).$$ In that case its general solution is $y_p+\sum_{k}C_ky_k$, where $y_p$ is any one solution and $y_1,\dots,y_r$ are independent eigenfunctions.

Proof. (1) Every entry of $I-\lambda A$ is a polynomial of degree $\le1$ in $\lambda$, so $D$ has degree $\le n$, and $D(0)=\det I=1$. By Theorem 7.5 with $f=0$, the solutions of the homogeneous equation are the functions $y=\lambda\sum c_ia_i$ with $(I-\lambda A)c=0$. For $\lambda\ne0$ the map $c\mapsto\lambda\sum c_ia_i$ is injective because the $a_i$ are linearly independent; so a nonzero solution exists iff $(I-\lambda A)c=0$ has a nonzero solution, i.e. iff $D(\lambda)=0$.

(2) If $D(\lambda)\ne0$, the system has the unique solution $c=(I-\lambda A)^{-1}F$, so by Theorem 7.5 the integral equation has exactly one solution, $y=f+\lambda\sum_ia_i\sum_j[(I-\lambda A)^{-1}]_{ij}F_j$. Substituting $F_j=\int b_j(t)f(t)\,dt$ gives the formula with $R$.

(3) The injective map above identifies the solution space of the homogeneous equation with the null space of $I-\lambda A$, of dimension $n-\operatorname{rank}(I-\lambda A)=r\ge1$. The transposed kernel $K(t,x)=\sum_ib_i(x)a_i(t)$ is separable with the roles of $a_i$ and $b_i$ exchanged, so its matrix is $A'_{ij}=\int a_ib_j=A_{ji}$, i.e. $A'=A^{T}$. Its homogeneous solutions are $z=\lambda\sum d_ib_i$ with $(I-\lambda A^{T})d=0$, and since $\operatorname{rank}(I-\lambda A^{T})$ $=\operatorname{rank}(I-\lambda A)$, there are again exactly $r$ independent ones. By Theorem 7.5 the non-homogeneous equation is solvable iff $(I-\lambda A)c=F$ is solvable, and by linear algebra this holds iff $F\cdot d=0$ for every $d$ with $(I-\lambda A)^{T}d=0$ (the column space of a matrix is the orthogonal complement of the null space of its transpose). For such $d$ and the corresponding $z=\lambda\sum d_ib_i$, $$F\cdot d=\sum_id_i\int_a^bb_i f\,dx=\int_a^bf\sum_id_ib_i\,dx=\frac1\lambda\int_a^bf\,z\,dx,$$ so the condition is $\int fz=0$ for every solution $z$ of the transposed homogeneous equation. The description of the general solution is linearity. ∎

This is the integral-equation version of the Fredholm alternative for boundary value problems (§2.4), and of the fact that $Ax=b$ is solvable iff $b\perp\ker A^{T}$. Fredholm's great discovery (§7.2) is that the same three statements hold for every continuous kernel, with $D(\lambda)$ replaced by an entire function. Note that the solvability condition involves the eigenfunctions of the transposed kernel; they coincide with those of $K$ only when $K$ is symmetric.

🧭 Problem-solving strategy — equations with separable kernels
  1. Write $K=\sum a_i(x)b_i(t)$ with as few terms as possible (expand $\cos(x\pm t)$, $\sin(x\pm t)$, $(x+t)^2$, …).
  2. Name the unknown constants $c_i=\int b_i y$, so that $y=f+\lambda\sum c_ia_i$. Multiply by each $b_i$ and integrate: this gives $(I-\lambda A)c=F$ with $A_{ij}=\int b_ia_j$ — mind the order of the indices.
  3. Characteristic numbers: roots of $D(\lambda)=\det(I-\lambda A)$. Eigenfunctions: $\sum c_ia_i$ for $c$ in the null space of $I-\lambda A$.
  4. At a characteristic number, test solvability with the transposed equation (or simply look for inconsistency in the linear system), and add the eigenfunctions to a particular solution.
  5. Check by substituting the answer back: two or three integrals, and they catch nearly every sign error.
✏️ Example 5 — The kernel $x+t$
Find the characteristic numbers of $K(x,t)=x+t$ on $[0,1]$, and solve $y(x)=1+\lambda\int_0^1(x+t)\,y(t)\,dt$ for every $\lambda$ that is not a characteristic number. What is the solution for $\lambda=1$?
Set-up $a_1=x$, $b_1=1$, $a_2=1$, $b_2=t$. Then $A_{11}=\int_0^1t\,dt=\tfrac12$, $A_{12}=\int_0^11\,dt=1$, $A_{21}=\int_0^1t^2\,dt=\tfrac13$, $A_{22}=\int_0^1t\,dt=\tfrac12$, and $F=(\int_0^11,\int_0^1t)$ $=(1,\tfrac12)$.
Characteristic numbers $$D(\lambda)=\left(1-\tfrac\lambda2\right)^2-\tfrac{\lambda^2}{3}=1-\lambda-\tfrac{\lambda^2}{12}.$$ $D=0\iff\lambda^2+12\lambda-12=0$ $\iff\lambda$ $=-6\pm4\sqrt3$, i.e. $\lambda_1\approx0.928$ and $\lambda_2\approx-12.93$.
Solve $(I-\lambda A)c=F$ reads $(1-\tfrac\lambda2)c_1-\lambda c_2=1$, $-\tfrac\lambda3c_1+(1-\tfrac\lambda2)c_2$ $=\tfrac12$. Cramer's rule gives $c_1=\dfrac{12}{12-12\lambda-\lambda^2}$, $c_2=\dfrac{6+\lambda}{12-12\lambda-\lambda^2}$, hence $$y(x)=1+\lambda(c_1x+c_2)=\frac{6\,(2\lambda x-\lambda+2)}{12-12\lambda-\lambda^2}.$$
$\lambda=1$ $y=\dfrac{6(2x+1)}{-1}=-12x-6$. Check: $\int_0^1(x+t)(-12t-6)\,dt$ $=x(-6-6)+(-4-3)$ $=-12x-7$, and $1+(-12x-7)=-12x-6$. Notice that $\lambda=1$ lies just beyond $\lambda_1\approx0.928$: the solution is large because $D(1)=-\tfrac1{12}$ is small (see Interactive 1).
Interactive 1 — Separable Kernels: the Determinant $D(\lambda)$ and the Fredholm Alternative

Left: $D(\lambda)=\det(I-\lambda A)$ for the chosen kernel; its real zeros (dots) are the real characteristic numbers. Right: the solution $y=f+\lambda\sum c_ia_i$ of $y=f+\lambda\int K y$ and the right-hand side $f$ (dashed). Drag $\lambda$ towards a zero of $D$ and watch $y$ blow up — unless $f$ satisfies the solvability condition of Theorem 7.6. Press Jump to a characteristic number to land exactly on one and see which alternative holds. Play sweeps $\lambda$.

—
λ—
D(λ)—
c = (c₁, c₂)—
real characteristic numbers—
✏️ Example 6 — Prove or disprove: solvability at a characteristic number
Let $K(x,t)=\cos(x+t)$ on $[0,2\pi]$. (a) Find the characteristic numbers and eigenfunctions. For $\lambda=\tfrac1\pi$ decide whether the equation $y(x)=f(x)+\tfrac1\pi\int_0^{2\pi}\cos(x+t)\,y(t)\,dt$ is solvable, and if so solve it, for (b) $f=\cos x$, (c) $f=\sin x$, (d) $f=x$.
(a) $\cos(x+t)$ $=\cos x\cos t-\sin x\sin t$: $a_1=\cos x$, $b_1=\cos t$, $a_2=\sin x$, $b_2=-\sin t$. Then $A_{11}=\int_0^{2\pi}\cos^2=\pi$, $A_{22}=-\int_0^{2\pi}\sin^2=-\pi$, $A_{12}=A_{21}=0$ (since $\int_0^{2\pi}\sin t\cos t\,dt=0$). $D(\lambda)$ $=(1-\pi\lambda)(1+\pi\lambda)$: characteristic numbers $\lambda=\tfrac1\pi$ with eigenfunction $\cos x$, and $\lambda=-\tfrac1\pi$ with eigenfunction $\sin x$. The kernel is symmetric, so the transposed equation has the same eigenfunctions.
(b) No solution By Theorem 7.6(3) we need $\int_0^{2\pi}f\cos x\,dx=0$; for $f=\cos x$ this integral is $\pi\ne0$. Directly: the first equation of the system is $(1-\pi\lambda)c_1=F_1$, i.e. $0\cdot c_1=\pi$.
(c) Infinitely many $\int\sin x\cos x=0$, so solvable. $F=(0,-\pi)$: the equations are $0\cdot c_1=0$ ($c_1$ free) and $2c_2=-\pi$. Hence $y=\sin x+\tfrac1\pi(c_1\cos x-\tfrac\pi2\sin x)$ $=\tfrac12\sin x+C\cos x$, $C$ arbitrary.
(d) Surprise: solvable $\int_0^{2\pi}x\cos x\,dx$ $=\big[x\sin x+\cos x\big]_0^{2\pi}=0$, so the condition holds although $f=x$ looks nothing like an eigenfunction. $F_2=-\int_0^{2\pi}t\sin t\,dt=2\pi$, so $c_2=\pi$ and $$y=x+\sin x+C\cos x.$$ Check: $\int_0^{2\pi}\cos(x+t)(t+\sin t)\,dt$ $=\cos x\cdot0-\sin x\,(-2\pi+\pi)$ $=\pi\sin x$, and $x+\tfrac1\pi\cdot\pi\sin x$ $=y-C\cos x$; the $C\cos x$ term reproduces itself.

Volterra Equations: Successive Approximations and the Laplace Transform

For Fredholm equations, iteration converges only for small $|\lambda|$ (§7.2). Volterra equations are different: each integration over $[a,x]$ produces a factor $(x-a)^n/n!$, and factorials beat every geometric growth.

📐 Theorem 7.7 — Volterra equations of the second kind: existence, uniqueness, successive approximations
Let $K$ be continuous on $T=\{a\le t\le x\le b\}$ with $M=\max_T|K|$, let $f\in C[a,b]$ with $\|f\|=\max|f|$, and let $\lambda\in\mathbb{C}$ be arbitrary. Define $$\varphi_0=f,\qquad\varphi_n(x)=\int_a^xK(x,t)\,\varphi_{n-1}(t)\,dt\quad(n\ge1).$$ Then:
  1. $|\varphi_n(x)|$ $\le\|f\|\,\dfrac{M^n(x-a)^n}{n!}$ for all $n$ and $x$;
  2. the series $y=\sum_{n=0}^\infty\lambda^n\varphi_n$ converges uniformly on $[a,b]$, and $y$ is the unique continuous solution of $$y(x)=f(x)+\lambda\int_a^xK(x,t)\,y(t)\,dt;$$
  3. the successive approximations $y_0=f$, $y_{n+1}=f+\lambda\int_a^xKy_n\,dt$ satisfy $y_n=\sum_{k=0}^n\lambda^k\varphi_k$ and $$\|y-y_n\|\le\|f\|\sum_{k=n+1}^{\infty}\frac{\big(|\lambda|M(b-a)\big)^k}{k!}\le\|f\|\,\frac{\big(|\lambda|M(b-a)\big)^{n+1}}{(n+1)!}\,e^{|\lambda|M(b-a)}.$$
In particular a continuous Volterra kernel has no characteristic numbers: the homogeneous equation has only the solution $y\equiv0$, for every $\lambda$.

Proof. Each $\varphi_n$ is continuous (an integral of a continuous function over $[a,x]$ with a continuous kernel). (1) by induction: true for $n=0$; if it holds for $n-1$, then $$|\varphi_n(x)|\le\int_a^xM\,\|f\|\frac{M^{n-1}(t-a)^{n-1}}{(n-1)!}\,dt=\|f\|\frac{M^n(x-a)^n}{n!}.$$ (2) By (1), $\|\lambda^n\varphi_n\|$ $\le\|f\|\frac{(|\lambda|M(b-a))^n}{n!}$, the terms of a convergent exponential series, so the Weierstrass M-test gives uniform convergence of $\sum\lambda^n\varphi_n$ to a continuous $y$. Uniform convergence lets us integrate term by term: $$\lambda\int_a^xK(x,t)y(t)\,dt=\sum_{n\ge0}\lambda^{n+1}\int_a^xK\varphi_n\,dt,$$ $$\sum_{n\ge0}\lambda^{n+1}\int_a^xK\varphi_n\,dt=\sum_{n\ge0}\lambda^{n+1}\varphi_{n+1}=y-f.$$ So $y$ is a solution. For uniqueness let $y,\tilde y$ be two continuous solutions and $w=y-\tilde y$, $W=\max|w|$. Then $w=\lambda\int_a^xKw\,dt$, so $|w(x)|\le|\lambda|MW(x-a)$, and inductively $|w(x)|$ $\le W\frac{(|\lambda|M(x-a))^n}{n!}$ for every $n$. The right side tends to $0$, so $w\equiv0$. (3) $y_n=\sum_{k\le n}\lambda^k\varphi_k$ by induction ($y_{n+1}$ $=f+\lambda\int K\sum_{k\le n}\lambda^k\varphi_k$ $=f+\sum_{k\le n}\lambda^{k+1}\varphi_{k+1}$), and the tail estimate follows from (1); the last inequality uses $\sum_{k\ge n+1}\frac{u^k}{k!}$ $\le\frac{u^{n+1}}{(n+1)!}e^u$. ∎

Compare Picard's theorem (§1.1): there the existence interval had length $h=\min(a,b/M)$ because the nonlinear right-hand side was controlled only near the initial point. In the linear case the Lipschitz condition holds on the whole strip, and the solution exists on all of $[a,b]$.

✏️ Example 7 — Iteration and the Laplace transform for the same equation
Solve $y(x)=x+\int_0^x(x-t)\,y(t)\,dt$ (a) by successive approximations, (b) by the Laplace transform, (c) by converting to an initial value problem. How many iterations guarantee an error below $10^{-6}$ on $[0,1]$?
(a) Iterates $\varphi_0=x$, $\varphi_1$ $=\int_0^x(x-t)t\,dt$ $=\tfrac{x^3}{6}$, $\varphi_2$ $=\int_0^x(x-t)\tfrac{t^3}{6}\,dt$ $=\tfrac{x^5}{120}$, and by induction $\varphi_n$ $=\dfrac{x^{2n+1}}{(2n+1)!}$ (using $\int_0^x(x-t)t^{m}\,dt$ $=\frac{x^{m+2}}{(m+1)(m+2)}$). With $\lambda=1$, $$y=\sum_{n\ge0}\frac{x^{2n+1}}{(2n+1)!}=\sinh x.$$
(b) Laplace The integral is the convolution $x*y$. With $Y=\mathcal{L}[y]$: $Y=\frac1{s^2}+\frac{1}{s^2}Y$, so $Y=\dfrac{1}{s^2-1}$ and $y=\sinh x$.
(c) IVP Differentiating twice: $y'=1+\int_0^xy$, $y''=y$, with $y(0)=0$, $y'(0)=1$: again $y=\sinh x$.
Error Here $M=\max_{0\le t\le x\le1}(x-t)=1$, $\|f\|=1$, $b-a=1$, so Theorem 7.7(3) gives $\|y-y_n\|$ $\le\sum_{k\ge n+1}\frac1{k!}$, which is $\lt10^{-6}$ once $n\ge9$ ($\sum_{k\ge10}1/k!\approx3.0\times10^{-7}$, while $\sum_{k\ge9}1/k!\approx3.0\times10^{-6}$). The true error is much smaller: $y-y_n=\sum_{k\gt n}\frac{x^{2k+1}}{(2k+1)!}$, already $\approx2.5\times10^{-8}$ for $n=4$ — the general bound ignores the special structure of this kernel.
Interactive 2 — Successive Approximations for Volterra Equations: Factorials Win

The faint curves are the successive approximations $y_0,y_1,\dots$, the bold curve is $y_n=\sum_{k\le n}\lambda^k\varphi_k$, and the dashed curve is the exact solution on $[0,2]$. Press Play to add iterations. Try large $|\lambda|$ and negative $\lambda$: the iterates may overshoot wildly at first, but the factor $\tfrac{(x-a)^n}{n!}$ in Theorem 7.7 always wins in the end. The readouts compare the true error with the guaranteed bound $\|f\|\sum_{k\gt n}\frac{(|\lambda|M(b-a))^k}{k!}$.

1.002
iterations n—
true error max|y − yₙ|—
bound of Theorem 7.7—
exact solution—

When the kernel depends only on $x-t$ and $a=0$, the integral is a convolution, $(k*y)(x)=\int_0^xk(x-t)y(t)\,dt$, and the Laplace transform $\mathcal{L}[g](s)$ $=\int_0^\infty e^{-sx}g(x)\,dx$ turns it into a product.

📐 Theorem 7.8 — Convolution kernels and the Laplace transform
Let $k$ and $f$ be continuous on $[0,\infty)$ and of exponential order ($|k(x)|,|f(x)|\le Ce^{\sigma x}$). Then the unique continuous solution of $$y(x)=f(x)+\lambda\int_0^xk(x-t)\,y(t)\,dt\qquad(x\ge0)$$ is of exponential order, and for all sufficiently large $s$ $$Y(s)=\frac{F(s)}{1-\lambda\,\hat k(s)},$$ where $Y,F,\hat k$ are the Laplace transforms of $y,f,k$. For the first-kind equation $\int_0^xk(x-t)y(t)\,dt=f(x)$ a continuous solution, if it exists, satisfies $\hat k(s)Y(s)=F(s)$.

Proof (sketch). Theorem 7.7, applied on every $[0,b]$, gives a unique continuous solution on $[0,\infty)$. Put $\tilde y=e^{-\sigma x}y$; it satisfies a Volterra equation with right-hand side $e^{-\sigma x}f$ (bounded by $C$) and kernel $e^{-\sigma(x-t)}k(x-t)$ (bounded by $C$), so the estimate in Theorem 7.7(2) gives $|\tilde y(x)|\le Ce^{|\lambda|Cx}$, and $y$ is of exponential order. The convolution theorem $\mathcal{L}[k*y]=\hat k\,Y$ then turns the equation into $Y=F+\lambda\hat kY$; for large $s$, $\hat k(s)\to0$, so $1-\lambda\hat k(s)\ne0$. ∎

The most famous convolution equation has a singular kernel. In Abel's problem a bead slides from rest at height $h$ along a smooth wire to the bottom; if $\phi(y)=\frac{ds}{dy}$ describes the wire, the descent time is $T(h)=\frac{1}{\sqrt{2g}}\int_0^h\frac{\phi(y)}{\sqrt{h-y}}\,dy$. Prescribing $T$ and finding $\phi$ is the following equation.

📐 Theorem 7.9 — Abel's integral equation
Let $f\in C^1[0,b]$ with $f(0)=0$. Then the equation $$\int_0^x\frac{y(t)}{\sqrt{x-t}}\,dt=f(x)\qquad(0\le x\le b)$$ has exactly one continuous solution, $$y(x)=\frac1\pi\,\frac{d}{dx}\int_0^x\frac{f(t)}{\sqrt{x-t}}\,dt=\frac{1}{\pi}\int_0^x\frac{f'(t)}{\sqrt{x-t}}\,dt.$$

Proof. The key integral is, for $t\lt x$, $$\int_t^x\frac{ds}{\sqrt{(x-s)(s-t)}}=\pi$$ (substitute $s=t+(x-t)\sin^2\theta$, $0\le\theta\le\tfrac\pi2$: then $ds=2(x-t)\sin\theta\cos\theta\,d\theta$ and $\sqrt{(x-s)(s-t)}$ $=(x-t)\sin\theta\cos\theta$). Uniqueness and the formula. If $y$ is a continuous solution, divide the equation (with $x$ replaced by $s$) by $\sqrt{x-s}$ and integrate over $s\in[0,x]$; interchanging the order of integration (the integrand is absolutely integrable) gives $$\int_0^x\frac{f(s)}{\sqrt{x-s}}\,ds=\int_0^xy(t)\int_t^x\frac{ds}{\sqrt{(x-s)(s-t)}}\,dt,$$ which equals $\pi\int_0^xy(t)\,dt$, and differentiating gives the first formula, so $y$ is determined by $f$. Writing $\int_0^xf(t)(x-t)^{-1/2}dt$ $=\int_0^xf(x-u)u^{-1/2}du$ and differentiating under the integral (the boundary term $f(0)x^{-1/2}$ vanishes) gives the second form, which is continuous. Existence. For this $y$, the same interchange gives $\int_0^x\frac{y(t)}{\sqrt{x-t}}dt$ $=\frac1\pi\int_0^xf'(s)\int_s^x\frac{dt}{\sqrt{(x-t)(t-s)}}\,ds$ $=\int_0^xf'(s)\,ds$ $=f(x)-f(0)=f(x)$. ∎

For example $f(x)=x$ gives $y=\frac1\pi\int_0^x\frac{dt}{\sqrt{x-t}}$ $=\frac{2\sqrt x}{\pi}$; check: $\int_0^x\frac{2\sqrt t}{\pi\sqrt{x-t}}\,dt$ $=\frac{2x}{\pi}B\!\left(\tfrac32,\tfrac12\right)$ $=\frac{2x}{\pi}\cdot\frac\pi2=x$. For the tautochrone, $T$ is constant, so $f$ is a nonzero constant $C$ and $f(0)\ne0$; the first formula still applies and gives $\phi(y)=\frac{C}{\pi\sqrt y}$, an integrable but unbounded solution — and the curve with $\frac{ds}{dy}\propto y^{-1/2}$ is a cycloid.

⚠️ Counterexamples and pitfalls
  1. First-kind Volterra equations need compatible data. $\int_0^xy(t)\,dt=1+x$ has no continuous solution ($x=0$ gives $0=1$); $\int_0^x(x-t)y(t)\,dt=\sin x$ has none either, because differentiating once gives $\int_0^xy=\cos x$, false at $x=0$. Theorem 7.1 needs $f(a)=0$ and $K(x,x)\ne0$.
  2. First-kind Fredholm equations are worse. $\int_0^1y(t)\,dt=x$ has no solution (the left side is constant); $\int_0^1y(t)\,dt=0$ has infinitely many; and $\int_0^1(x+t)y(t)\,dt=f(x)$ is solvable only when $f$ is linear (Exercise III.8). Second-kind equations have no such range restriction.
  3. Volterra second kind is always uniquely solvable (continuous kernel) — but this fails for unbounded kernels: with $K(x,t)=\frac1x$ on $0\lt t\le x\le1$, every $\lambda\ge1$ has the eigenfunction $x^{\lambda-1}$ (Exercise IV.7).
  4. A nonzero kernel may have no characteristic numbers. $\sin x\cos t$ on $[0,\pi]$: $A=\int_0^\pi\cos t\sin t\,dt=0$, so $D(\lambda)\equiv1$. Every continuous Volterra kernel is another example.
  5. Characteristic numbers may all be non-real. $\sin(x-t)$ on $[0,2\pi]$ gives $D=1+\pi^2\lambda^2$; $x-t$ on $[0,1]$ gives $D=1+\frac{\lambda^2}{12}$, roots $\pm2\sqrt3\,i$. Only symmetric kernels are guaranteed real ones (§7.2).
  6. Rank $n$ does not mean $n$ characteristic numbers. $D$ has degree $\le n$; the degree drops when $\det A=0$ (e.g. $1+\sin\pi x\cos\pi t$ on $[0,1]$ has rank $2$ but $D=1-\lambda$), and a root may be repeated ($\cos(x-t)$ on $[0,2\pi]$ has the double characteristic number $\frac1\pi$ with the two eigenfunctions $\cos x$, $\sin x$).
  7. The solvability condition uses the transposed kernel. For $K=t$ on $[0,1]$ (characteristic number $2$, eigenfunction $1$), the transposed eigenfunction is $x$, and $y=f+2\int_0^1t\,y(t)\,dt$ is solvable iff $\int_0^1xf\,dx=0$. With $f=3x-2$ we have $\int_0^1f\cdot1\,dx=-\tfrac12\ne0$ but $\int_0^1xf\,dx=0$: solvable.
  8. Index order in $A$. $A_{ij}=\int b_i\,a_j$, not $\int a_ib_j$; using the wrong one gives $A^{T}$, which has the same determinant (so the same characteristic numbers) but wrong eigenfunctions and wrong solutions.
  9. Green's function must exist. $y''+\lambda y=0$, $y'(0)=y'(1)=0$ cannot be converted with the Green's function of $u''$ — none exists, since constants solve the homogeneous Neumann problem. Write $y''-y=-(\lambda+1)y$ instead (Exercise II.5).
  10. Do not drop the initial data. For $y''=g-qy$ with $y(a)=\alpha$, $y'(a)=\beta$ the free term is $\alpha+\beta(x-a)+\int(x-t)g$; forgetting $\beta(x-a)$, or writing $(t-x)$ for $(x-t)$, are the commonest errors in Part B options.
  11. Laplace transforms need convolution form. They apply to $\int_0^xk(x-t)y(t)\,dt$, not to Fredholm integrals $\int_0^1$, nor to kernels like $k(x+t)$ or $xt$.
  12. Nonlinear is different. $y=1+\int_0^xy^2$ blows up at $x=1$; Theorem 7.7 is a theorem about linear equations.

Exercises

I
Classification, first-kind equations and initial value problems.
  • 1 Classify: (a) $y(x)=e^x+\int_0^x\sin(x-t)\,y(t)\,dt$; (b) $\int_0^1e^{xt}y(t)\,dt=g(x)$; (c) $y(x)=\lambda\int_{-1}^1(1+xt)\,y(t)\,dt$; (d) $y(x)=x+\lambda\int_0^1\sqrt{y(t)}\,dt$.
  • 2 Show that $y=\cos x$ solves $y(x)=1-\int_0^x(x-t)\,y(t)\,dt$.
  • 3 Convert $\int_0^x(2+x-t)\,y(t)\,dt=x$ into a second-kind equation and solve it.
  • 4 Convert $y''+3y'+2y=0$, $y(0)=1$, $y'(0)=0$, into a Volterra equation for $y$, and verify it with the exact solution.
  • 5 Show that $y'''-2y=0$, $y(0)=1$, $y'(0)=y''(0)=0$, is equivalent to $y(x)=1+\int_0^x(x-t)^2y(t)\,dt$.
  • 6 Find the initial value problem equivalent to $y(x)=1+\int_0^x(x-2t)\,y(t)\,dt$, and solve it.
  • 7 Solve $\int_0^x\cos(x-t)\,y(t)\,dt$ $=\sin x$ (a) by Theorem 7.1, (b) by the Laplace transform.
  • 8 Show that neither $\int_0^xy(t)\,dt=1+x$ nor $\int_0^x(x-t)\,y(t)\,dt=\sin x$ has a continuous solution.
1 (a) Volterra, second kind, non-homogeneous, convolution kernel. (b) Fredholm, first kind (non-separable kernel $e^{xt}$). (c) Fredholm, second kind, homogeneous, symmetric separable kernel. (d) Not linear: the unknown enters through $\sqrt{y}$.
2 $\int_0^x(x-t)\cos t\,dt$ $=\big[(x-t)\sin t\big]_0^x+\int_0^x\sin t\,dt$ $=1-\cos x$, and $1-(1-\cos x)=\cos x$.
3 $K(x,x)=2$, $K_x=1$, $f(0)=0$, $f'=1$: $y(x)+\tfrac12\int_0^xy(t)\,dt$ $=\tfrac12$. Differentiating, $y'=-\tfrac12y$, and $y(0)=\tfrac12$, so $y=\tfrac12e^{-x/2}$. Check: $\int_0^x(2+x-t)\tfrac12e^{-t/2}dt=x$ (integrate by parts).
4 $p=3$, $q=2$: $K=(x-t)(0-2)-3=-3-2(x-t)$, $F=1+(0+3\cdot1)x=1+3x$. So $y(x)=1+3x-\int_0^x\big[3+2(x-t)\big]y(t)\,dt$. The exact solution $y=2e^{-x}-e^{-2x}$ satisfies it (direct integration).
5 Integrate $y'''=2y$ three times from $0$: by Theorem 7.2, $y=1+0\cdot x+0\cdot\tfrac{x^2}{2}+\tfrac{1}{2!}\int_0^x(x-t)^2\,2y(t)\,dt$. Conversely, differentiating the integral equation three times gives $y'''=2y$, and putting $x=0$ in it and in its first two derivatives gives $y(0)=1$, $y'(0)=y''(0)=0$.
6 $y'=\int_0^xy\,dt+(x-2x)y(x)$ $=\int_0^xy\,dt-xy$, $y''=y-y-xy'=-xy'$; $y(0)=1$, $y'(0)=0$. Then $(y')'=-xy'$ with $y'(0)=0$ forces $y'\equiv0$, so $y\equiv1$. Check: $\int_0^x(x-2t)\,dt=x^2-x^2=0$.
7 (a) $K(x,x)=1$, $K_x=-\sin(x-t)$: $y(x)-\int_0^x\sin(x-t)y(t)\,dt$ $=\cos x$; $y\equiv1$ works: $1-(1-\cos x)=\cos x$. (b) $\frac{s}{s^2+1}Y=\frac1{s^2+1}$, so $Y=\frac1s$ and $y=1$. Check: $\int_0^x\cos(x-t)\,dt=\sin x$.
8 At $x=0$ the first equation reads $0=1$. For the second, a continuous solution would make the left side $C^1$ with derivative $\int_0^xy\,dt$, so $\int_0^xy=\cos x$; at $x=0$ this reads $0=1$. In Theorem 7.1 terms: $K(x,x)=0$, and the extra compatibility condition $f'(0)=0$ fails.
II
Boundary value problems and Fredholm equations. Throughout, $G_0(x,t)=\min(x,t)\,(1-\max(x,t))$ on $[0,1]^2$ is the Green's function of $-u''=h$, $u(0)=u(1)=0$.
  • 1 Show that $y''+\lambda y=0$, $y(0)=y(1)=0$, is equivalent to $y=\lambda\int_0^1G_0(x,t)\,y(t)\,dt$.
  • 2 Convert $y''+\lambda y=1$, $y(0)=y(1)=0$, into a Fredholm equation.
  • 3 Convert $y''+\lambda y=0$, $y(0)=0$, $y(1)=1$ (inhomogeneous boundary condition) into a Fredholm equation.
  • 4 Show that $y''+\lambda y=0$, $y'(0)=0$, $y(1)=0$, is equivalent to $y=\lambda\int_0^1\big(1-\max(x,t)\big)\,y(t)\,dt$.
  • 5 Convert $y''+\lambda y=0$, $y'(0)=y'(1)=0$, into a Fredholm equation by writing it as $y''-y=-(\lambda+1)y$.
  • 6 Find the boundary value problem equivalent to $y(x)=x+\lambda\int_0^1\big(1-\max(x,t)\big)\,y(t)\,dt$.
  • 7 The IVP $y''+\lambda y=0$, $y(0)=0$, $y'(0)=1$ has a solution for every $\lambda$, but the BVP $y''+\lambda y=1$, $y(0)=y(1)=0$ does not. Explain this in terms of the two integral equations.
  • 8 Whirling string. A string of length $L$, tension $T$ and linear density $\rho(x)\gt0$ rotates with angular speed $\omega$ about its equilibrium axis; a small steady deflection $y$ satisfies $Ty''+\omega^2\rho(x)y=0$, $y(0)=y(L)=0$. Write this as a Fredholm equation, show how to make its kernel symmetric, and find the critical speeds when $\rho\equiv\rho_0$.
1 Apply Theorem 7.4 with $L[u]=u''$ ($p=1$, $q=0$), whose Dirichlet Green's function is $-G_0$, and $r=1$, $h=0$: $y=-\lambda\int(-G_0)\,y\,dt$ $=\lambda\int_0^1G_0\,y\,dt$. (This is Example 4(a), done the other way round.)
2 $-y''=\lambda y-1$, so $y=\int_0^1G_0(\lambda y-1)\,dt$ $=-\tfrac{x(1-x)}{2}+\lambda\int_0^1G_0\,y\,dt$, since $\int_0^1G_0(x,t)\,dt$ $=\tfrac{x(1-x)}2$.
3 Put $w=y-x$: $w''+\lambda w=-\lambda x$, $w(0)=w(1)=0$, so $w=\int G_0\,\lambda(t+w)\,dt$, i.e. $y-x=\lambda\int G_0\,y\,dt$: $$y(x)=x+\lambda\int_0^1G_0(x,t)\,y(t)\,dt.$$
4 Green's function of $u''$ with $u'(0)=0$, $u(1)=0$: $u_1=1$, $u_2=x-1$, $C=u_1u_2'-u_1'u_2=1$, $G=\max(x,t)-1$. Then $y''=-\lambda y$ gives $y=\int G(-\lambda y)$ $=\lambda\int(1-\max(x,t))y\,dt$. Converse by Theorem 7.4.
5 For $u''-u$ with $u'(0)=u'(1)=0$ the homogeneous problem has only $u=0$. Take $u_1=\cosh x$, $u_2=\cosh(1-x)$; $C=u_1u_2'-u_1'u_2=-\sinh1$. So $G=-\frac{\cosh(\min(x,t))\cosh(1-\max(x,t))}{\sinh1}$, and $y=\int G\cdot(-(\lambda+1)y)$: $$y(x)=(\lambda+1)\int_0^1\frac{\cosh(\min(x,t))\,\cosh(1-\max(x,t))}{\sinh1}\,y(t)\,dt.$$ Its characteristic values are $\lambda+1=1+n^2\pi^2$, $n=0,1,2,\dots$
6 Let $u=\int_0^1(1-\max(x,t))y\,dt$ $=(1-x)\int_0^xy+\int_x^1(1-t)y$. Then $u'=-\int_0^xy$, $u''=-y$, $u'(0)=0$, $u(1)=0$. From $y=x+\lambda u$: $y''=-\lambda y$, $y'(0)=1$, $y(1)=1$.
7 The IVP is equivalent to the Volterra equation $y=x-\lambda\int_0^x(x-t)y\,dt$, which has a unique solution for every $\lambda$ (Theorem 7.7). The BVP is equivalent to the Fredholm equation $y=-\frac{x(1-x)}2+\lambda\int G_0y$, and at the characteristic numbers $\lambda=n^2\pi^2$ the Fredholm alternative applies: solvable iff $\int_0^1\frac{x(1-x)}{2}\sin n\pi x\,dx=0$, which fails for odd $n$ (the integral is $\frac{2}{n^3\pi^3}$).
8 $-y''=\frac{\omega^2}{T}\rho y$ with $G_L(x,t)$ $=\frac{\min(x,t)\,(L-\max(x,t))}{L}$ (Green's function of $-u''$ on $[0,L]$): $y(x)=\frac{\omega^2}{T}\int_0^LG_L(x,t)\rho(t)y(t)\,dt$. The kernel $G_L(x,t)\rho(t)$ is not symmetric; with $z=\sqrt\rho\,y$: $z(x)=\frac{\omega^2}{T}\int_0^L\sqrt{\rho(x)}G_L(x,t)\sqrt{\rho(t)}\,z(t)\,dt$, symmetric. Critical speeds: $\frac{\omega^2}{T}$ must be a characteristic number; for $\rho\equiv\rho_0$, $\frac{\omega^2\rho_0}{T}$ $=\frac{n^2\pi^2}{L^2}$, i.e. $\omega_n$ $=\frac{n\pi}{L}\sqrt{T/\rho_0}$, $n\in\mathbb{N}$ — the frequencies of the vibrating string.
III
Separable kernels and the Fredholm alternative.
  • 1 Solve $y(x)=x+\lambda\int_0^1xt\,y(t)\,dt$. For which $\lambda$ is there no solution?
  • 2 Solve $y(x)=x^2+\int_0^1(x+t)\,y(t)\,dt$.
  • 3 Find the characteristic numbers and eigenfunctions of $\cos(x-t)$ on $[0,2\pi]$.
  • 4 Find the characteristic numbers and eigenfunctions of $1+xt$ on $[-1,1]$.
  • 5 For which $a$ is $y(x)=x+a+2\int_0^1t\,y(t)\,dt$ solvable? Find all solutions for that $a$.
  • 6 Find the characteristic numbers and eigenfunctions of $\sin(x+t)$ on $[0,\pi]$, and solve $y(x)=1+\lambda\int_0^\pi\sin(x+t)\,y(t)\,dt$.
  • 7 Prove that a separable kernel of rank $n$ has at most $n$ characteristic numbers. Give a kernel of rank $2$ with no real characteristic number.
  • 8 Show that $\int_0^1(x+t)\,y(t)\,dt=f(x)$ has a continuous solution if and only if $f$ is a polynomial of degree $\le1$, and find all solutions for $f(x)=x$.
1 $y=x+\lambda cx$ with $c=\int_0^1ty\,dt$ $=\tfrac13+\tfrac\lambda3c$, so $c=\frac1{3-\lambda}$ and $y=\dfrac{3x}{3-\lambda}$ for $\lambda\ne3$. At $\lambda=3$: $0\cdot c=\tfrac13$, no solution ($\int_0^1x\cdot x\,dx\ne0$; the kernel is symmetric).
2 As in Example 5 with $\lambda=1$ and $F=(\int t^2,\int t^3)$ $=(\tfrac13,\tfrac14)$: $\tfrac12c_1-c_2=\tfrac13$, $-\tfrac13c_1+\tfrac12c_2=\tfrac14$, so $c_1=-5$, $c_2=-\tfrac{17}6$, and $y=x^2-5x-\tfrac{17}{6}$. (Substituting back confirms it.)
3 $a=(\cos x,\sin x)$, $b=(\cos t,\sin t)$, $A=\pi I$, $D=(1-\pi\lambda)^2$. One characteristic number $\lambda=\tfrac1\pi$, of multiplicity $2$: every $C_1\cos x+C_2\sin x\ne0$ is an eigenfunction.
4 $a=(1,x)$, $b=(1,t)$: $A=\begin{pmatrix}2&0\\0&2/3\end{pmatrix}$. $\lambda=\tfrac12$ with eigenfunction $1$, and $\lambda=\tfrac32$ with eigenfunction $x$.
5 $c=\int_0^1ty\,dt$ satisfies $c=\int_0^1t(t+a)\,dt+2\cdot\tfrac12c$ $=\tfrac13+\tfrac a2+c$, so solvable iff $a=-\tfrac23$ (equivalently $\int_0^1x\,f\,dx=0$, $x$ being the transposed eigenfunction). Then $c$ is arbitrary and $y=x-\tfrac23+C$, $C\in\mathbb{R}$.
6 $a=(\sin x,\cos x)$, $b=(\cos t,\sin t)$: $A=\begin{pmatrix}0&\pi/2\\ \pi/2&0\end{pmatrix}$, $D=1-\frac{\pi^2\lambda^2}{4}$. $\lambda=\tfrac2\pi$: eigenfunction $\sin x+\cos x$; $\lambda=-\tfrac2\pi$: $\sin x-\cos x$. With $F=(0,2)$: $$y=1+\frac{4\pi\lambda^2\sin x+8\lambda\cos x}{4-\pi^2\lambda^2}\qquad(\lambda\ne\pm\tfrac2\pi).$$ At $\lambda=\pm\frac2\pi$ there is no solution, because $\int_0^\pi(\sin x\pm\cos x)\,dx$ $=2\ne0$.
7 By Theorem 7.6(1) the characteristic numbers are the roots of $D$, a polynomial of degree $\le n$ with $D(0)=1\ne0$, so $D\not\equiv0$ and has at most $n$ roots. Example: $\sin(x-t)$ $=\sin x\cos t-\cos x\sin t$ on $[0,2\pi]$, with $D=1+\pi^2\lambda^2$ (roots $\pm i/\pi$); or $x-t$ on $[0,1]$, $D=1+\frac{\lambda^2}{12}$.
8 The left side equals $c_1x+c_2$ with $c_1=\int y$, $c_2=\int ty$, so $f$ must be linear. Conversely for $f=\alpha x+\beta$ try $y=p+qt$: $p+\tfrac q2=\alpha$, $\tfrac p2+\tfrac q3=\beta$ (determinant $\tfrac1{12}\ne0$). For $f=x$: $p=4$, $q=-6$, $y=4-6t$; check $\int_0^1(x+t)(4-6t)\,dt$ $=x(4-3)+(2-2)=x$. All solutions: $4-6x+g(x)$ with $\int_0^1g=\int_0^1xg=0$, e.g. $g=C(6x^2-6x+1)$ — infinitely many.
IV
Volterra equations, transforms and proofs.
  • 1 Solve $y(x)=1+\int_0^x(x-t)\,y(t)\,dt$ by successive approximations.
  • 2 Solve $y(x)=1+2\int_0^xt\,y(t)\,dt$ by successive approximations, and check by converting to an IVP.
  • 3 Solve $y(x)=\sin x+2\int_0^x\cos(x-t)\,y(t)\,dt$ by the Laplace transform.
  • 4 Solve $y(x)=x+\int_0^x\sin(x-t)\,y(t)\,dt$.
  • 5 Solve Abel's equation $\int_0^x\frac{y(t)}{\sqrt{x-t}}\,dt=x^2$.
  • 6 Prove, using Gronwall's inequality (§1.1), that the homogeneous equation $y(x)=\lambda\int_a^xK(x,t)y(t)\,dt$ with continuous $K$ has only the trivial solution.
  • 7 Show that for the unbounded kernel $K(x,t)=\frac1x$ ($0\lt t\le x\le1$), every $\lambda\ge1$ admits a nonzero continuous solution of $y(x)=\lambda\int_0^x\frac{y(t)}{x}\,dt$. Which hypothesis of Theorem 7.7 fails?
  • 8 Renewal equation. In a simple population model the birth rate $B(t)$ satisfies $B(t)=e^{-t}+\int_0^t\beta e^{-(t-s)}B(s)\,ds$ ($t$ in generations, $\beta\gt0$ the reproduction rate). Solve it and decide when the population grows.
1 $\varphi_0=1$, $\varphi_1=\tfrac{x^2}{2}$, $\varphi_n=\tfrac{x^{2n}}{(2n)!}$ (same integral as in Example 7), so $y=\cosh x$. (IVP: $y''=y$, $y(0)=1$, $y'(0)=0$.)
2 $\varphi_0=1$, $\varphi_n$ $=\int_0^x2t\,\varphi_{n-1}\,dt$ gives $\varphi_n=\tfrac{x^{2n}}{n!}$, so $y=e^{x^2}$. IVP: $y'=2xy$, $y(0)=1$, whose solution is $e^{x^2}$.
3 $Y=\frac1{s^2+1}+\frac{2s}{s^2+1}Y$, so $Y(s^2-2s+1)=1$, $Y=\frac1{(s-1)^2}$, $y=xe^x$. (Substitution confirms it.)
4 $Y=\frac1{s^2}+\frac{Y}{s^2+1}$, so $Y\frac{s^2}{s^2+1}=\frac1{s^2}$, $Y=\frac{s^2+1}{s^4}$ $=\frac1{s^2}+\frac1{s^4}$, and $y=x+\tfrac{x^3}{6}$.
5 $f=x^2$, $f(0)=0$: $y=\frac1\pi\int_0^x\frac{2t}{\sqrt{x-t}}\,dt$ $=\frac{2}{\pi}\cdot\frac43x^{3/2}$ $=\frac{8}{3\pi}x^{3/2}$ (using $\int_0^xt(x-t)^{-1/2}dt$ $=x^{3/2}B(2,\tfrac12)$ $=\tfrac43x^{3/2}$). Check: $\int_0^x\frac{8t^{3/2}}{3\pi\sqrt{x-t}}dt$ $=\frac{8}{3\pi}x^2B(\tfrac52,\tfrac12)$ $=\frac{8}{3\pi}\cdot\frac{3\pi}{8}x^2=x^2$.
6 $u(x)=|y(x)|$ is continuous and $u(x)\le|\lambda|M\int_a^xu(t)\,dt$. Gronwall's inequality with constant term $0$ gives $u(x)\le0\cdot e^{|\lambda|M(x-a)}=0$, so $y\equiv0$.
7 For $y=x^k$ ($k\ge0$): $\lambda\int_0^x\frac{t^k}{x}\,dt$ $=\frac{\lambda}{k+1}x^k$, which equals $x^k$ iff $\lambda=k+1$. So every $\lambda\ge1$ is a characteristic number, with eigenfunction $x^{\lambda-1}$ (continuous on $[0,1]$). The kernel is unbounded near $(0,0)$, so $M=\max|K|$ does not exist and the estimate (1) of Theorem 7.7 breaks down.
8 Laplace: $\hat B$ $=\frac1{s+1}+\frac{\beta}{s+1}\hat B$, so $\hat B=\frac1{s+1-\beta}$ and $B(t)=e^{(\beta-1)t}$. The births grow iff $\beta\gt1$, decay iff $\beta\lt1$ — the threshold "basic reproduction number $=1$" of epidemiology and demography.
🎯 Exam corner CSIR Part BCSIR Part CGATE MA

E1. CSIR Part B GATE MA The initial value problem $y''-xy=0$, $y(0)=1$, $y'(0)=2$, is equivalent to: (A) $y(x)=1+2x+\int_0^x(x-t)\,t\,y(t)\,dt$; (B) $y(x)=1+2x-\int_0^x(x-t)\,t\,y(t)\,dt$; (C) $y(x)=2+x+\int_0^x(x-t)\,t\,y(t)\,dt$; (D) $y(x)=1+2x+\int_0^x(x-t)\,x\,y(t)\,dt$.

Idea $y''=xy$; integrate twice from $0$ and use $\int_0^x\!\int_0^sg$ $=\int_0^x(x-t)g(t)\,dt$ with $g(t)=t\,y(t)$.
Options (A) is exactly that: $y=y(0)+y'(0)x+\int_0^x(x-t)\,t\,y(t)\,dt$. (B) has the sign of $q$ wrong; (C) swaps the initial values; (D) puts the variable $x$ instead of $t$ in the integrand (differentiating twice would give $y''=xy+2\int_0^xy\,dt$, not $xy$).
Answer (A).

E2. CSIR Part B The characteristic numbers of the kernel $K(x,t)=1+3xt$ on $[0,1]$ are: (A) $4\pm2\sqrt3$; (B) $2\pm\sqrt3$; (C) $\tfrac{2\pm\sqrt3}{2}$; (D) $-4\pm2\sqrt3$.

Matrix $a=(1,x)$, $b=(1,3t)$: $A_{11}=\int1=1$, $A_{12}=\int x=\tfrac12$, $A_{21}=\int3t=\tfrac32$, $A_{22}=\int3t^2=1$.
Determinant $D=(1-\lambda)^2-\tfrac34\lambda^2$ $=1-2\lambda+\tfrac{\lambda^2}4=0$ $\iff\lambda^2-8\lambda+4=0$ $\iff\lambda$ $=4\pm2\sqrt3$. Option (C) lists the reciprocals $\frac{1}{4\pm2\sqrt3}$ $=\frac{2\mp\sqrt3}{2}$, the eigenvalues of the integral operator, not the characteristic numbers.
Answer (A).

E3. CSIR Part B The continuous solution of $y(x)=e^x-\int_0^xe^{x-t}\,y(t)\,dt$ is: (A) $1$; (B) $e^x$; (C) $e^{2x}$; (D) $\cosh x$.

Laplace $Y=\frac{1}{s-1}-\frac{1}{s-1}Y$, so $Y\cdot\frac{s}{s-1}=\frac1{s-1}$ and $Y=\frac1s$: $y\equiv1$.
Check $\int_0^xe^{x-t}\,dt=e^x-1$, and $e^x-(e^x-1)=1$. The solution is unique by Theorem 7.7, so the other options are wrong (e.g. (B) gives $e^x-xe^x\ne e^x$).
Answer (A).

E4. CSIR Part C Consider $y(x)=f(x)+\lambda\int_0^1x\,y(t)\,dt$ on $[0,1]$. Which of the following statements are true? (One or more options may be correct.) (A) $\lambda=2$ is the only characteristic number. (B) For $\lambda=2$ and $f(x)=2x-1$ the equation has infinitely many solutions. (C) For $\lambda=2$ and $f(x)=3x-2$ the equation has a solution. (D) For $\lambda=1$ and $f\equiv1$ the solution is $y=1+2x$.

(A) True Rank one: $a=x$, $b=1$, $A=\int_0^1x\,dx=\tfrac12$, $D=1-\tfrac\lambda2$. Eigenfunction $x$. The transposed kernel $K(t,x)=t$ has eigenfunction $z=1$.
(B) True Solvable iff $\int_0^1f\cdot1\,dx=0$; $\int_0^1(2x-1)\,dx=0$. Then $y=2x-1+Cx$ for every $C$.
(C) False $\int_0^1(3x-2)\,dx=-\tfrac12\ne0$. (The trap: $f$ is orthogonal to the eigenfunction $x$ of $K$, since $\int_0^1x(3x-2)\,dx=0$ — but the condition uses the transposed eigenfunction.)
(D) True $c=\int_0^1y=1+\tfrac{\lambda}{2}c$, so $c=2$ at $\lambda=1$ and $y=1+2x$. Check: $\int_0^1x(1+2t)\,dt=2x$.
Answer (A), (B), (D).

E5. CSIR Part C Let $K$ be continuous on $[0,1]^2$, not identically zero. Which of the following are necessarily true? (A) For every $\lambda$ and every $f\in C[0,1]$, $y(x)=f(x)+\lambda\int_0^xK(x,t)y(t)\,dt$ has exactly one continuous solution. (B) If $|\lambda|\max|K|\lt1$, then $y(x)=f(x)+\lambda\int_0^1K(x,t)y(t)\,dt$ has exactly one continuous solution for every $f$. (C) $\int_0^xK(x,t)y(t)\,dt=f(x)$ has a continuous solution for every $f\in C^1[0,1]$ with $f(0)=0$. (D) $K$ has at least one characteristic number.

(A) True Theorem 7.7: Volterra equations of the second kind with continuous kernel.
(B) True Here $b-a=1$; the Neumann series of §7.2 converges when $|\lambda|M(b-a)\lt1$ and gives the unique solution. (Directly: the map $y\mapsto f+\lambda\int Ky$ is a contraction in the max norm, with constant $|\lambda|M\lt1$.)
(C) False $K=x-t$ and $f=x$: differentiating gives $\int_0^xy=1$, impossible at $x=0$. Theorem 7.1 needs $K(x,x)\ne0$.
(D) False $K=\sin(\pi x)\cos(\pi t)$ on $[0,1]$: $A=\int_0^1\cos\pi t\sin\pi t\,dt=0$, $D\equiv1$, no characteristic numbers.
Answer (A), (B).

E6. CSIR Part C GATE MA Consider the boundary value problem $y''+\lambda y=0$, $y(0)=0$, $y'(1)=0$. Which of the following are true? (A) It is equivalent to $y(x)=\lambda\int_0^1\min(x,t)\,y(t)\,dt$. (B) Its characteristic numbers are $n^2\pi^2$, $n\in\mathbb{N}$. (C) $y''+\lambda y=x$, $y(0)=0$, $y'(1)=0$ is equivalent to $y(x)=\frac{x^3}{6}-\frac{x}{2}+\lambda\int_0^1\min(x,t)\,y(t)\,dt$. (D) For $\lambda=\frac{\pi^2}{4}$ the equation $y(x)=x+\lambda\int_0^1\min(x,t)\,y(t)\,dt$ has a solution.

(A) True The Green's function of $u''$ with these conditions is $-\min(x,t)$ (Example 3), so $y''=-\lambda y$ becomes $y=\lambda\int\min(x,t)\,y\,dt$.
(B) False $y=\sin(\sqrt\lambda x)$ with $\cos\sqrt\lambda=0$: $\lambda$ $=\left(\frac{(2n-1)\pi}{2}\right)^2$. ($\lambda\le0$ gives only $y=0$.)
(C) True Example 3.
(D) False $\lambda=\frac{\pi^2}{4}$ is a characteristic number with eigenfunction $\sin\frac{\pi x}{2}$ (the kernel is symmetric). Solvability needs $\int_0^1x\sin\frac{\pi x}{2}\,dx=0$, but this integral is $\frac{4}{\pi^2}\ne0$.
Answer (A), (C).
🧑‍🏫 For the teacher

Suggested plan — 3 lectures × 60 min.

  1. Classification and conversion. Open with Abel's bead problem and the population-with-memory story; Definition 7.1 and Example 1 (ask the class to classify before you do; part (e) catches those who forget linearity). Theorem 7.2 on the board, then Theorem 7.3 with Example 2 — let students derive both forms and compare. Finish with Theorem 7.1 and Exercise I.8.
  2. Boundary value problems and separable kernels. Recall the Green's function from §2.4; Theorem 7.4, Examples 3–4 and the strategy box. Then Definition 7.2, Theorem 7.5 (the proof is five lines — do it fully) and Example 5. Project Interactive 1 with the kernel $x+t$ and let the class predict where $y$ blows up.
  3. The alternative and Volterra theory. Theorem 7.6 with its linear-algebra proof; Example 6 and Interactive 1 with $\cos(x+t)$ (solvable at $\frac1\pi$, not at $-\frac1\pi$) and with $1-2x+3t$ (non-symmetric). Theorem 7.7 with full proof, Example 7, Interactive 2; Laplace transforms and Abel's equation as a short coda. Do E4 and E5 live.

Discussion questions. (1) Why does an IVP lead to a Volterra equation and a BVP to a Fredholm equation — what is it about "information at one end" versus "information at both ends"? (2) The Fredholm alternative says "uniqueness implies existence". Where else have you met this principle (square matrices, Green's functions)? (3) Why are first-kind equations ill-posed? What happens to $\int_0^1\sin(n\pi t)\,y(t)\,dt$-type data as $n\to\infty$?

Misconceptions to watch for. Forgetting the $\beta(x-a)$ term or writing $(t-x)$ for $(x-t)$; testing solvability against the eigenfunctions of $K$ rather than of $K^{T}$; believing every kernel has characteristic numbers, or that they must be real; using $A_{ij}=\int a_ib_j$; assuming a first-kind equation is solvable whenever the second-kind one is; applying the Laplace transform to a Fredholm integral.

Suggested homework. I(3, 4, 6, 8), II(2, 3, 4, 8), III(1, 2, 5, 6, 8), IV(1, 3, 5, 7, 8); Exam corner E2, E4, E6.

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

Research question. A flame, a plasma column or a gas cloud around a star is often radially symmetric, but a camera only sees its brightness integrated along lines of sight. Can the radial emission profile be recovered from the side-on image, and how badly do measurement errors spoil the answer?

Why it matters. This is Abel inversion, used daily in combustion research, laser-plasma diagnostics in fusion laboratories, and astrophysics. Its mathematics is Abel's equation — an integral equation of the first kind — so it inherits the ill-posedness of first-kind problems: the inversion formula contains a derivative of measured data.

What to do.

  1. For a radial profile $\varepsilon(r)$ on $0\le r\le1$ the side-on projection is $P(y)=2\int_y^1\frac{\varepsilon(r)\,r}{\sqrt{r^2-y^2}}\,dr$. Show that the substitutions $x=1-y^2$, $t=1-r^2$ turn this into an equation of Abel type, and verify that $\varepsilon(r)=1-r^2$ gives $P(y)=\tfrac43(1-y^2)^{3/2}$.
  2. Implement the inversion formula $\varepsilon(r)$ $=-\frac1\pi\int_r^1\frac{P'(y)}{\sqrt{y^2-r^2}}\,dy$ numerically (finite differences for $P'$, a quadrature that copes with the square-root singularity), and recover $\varepsilon$ from exact data on a grid of $100$ points.
  3. Add random noise of relative size $1\%$ to $P$ and invert again. Measure the error of $\varepsilon$ near $r=0$ and near $r=1$.
  4. Smooth the noisy data first (for instance by a least-squares fit of a low-degree polynomial in $y^2$) and repeat. Explain the trade-off you see between fitting the data and stabilising the inversion.

Tools: Python with NumPy (or any language with arrays); a plotting library.

Hand in: Plots of the true and recovered profiles with and without noise and smoothing, and a paragraph explaining why the noise is amplified.

Going further. Read about Tikhonov regularisation for first-kind equations, and about the Herglotz–Wiechert inversion that recovers the seismic velocity inside the Earth from travel times — another Abel equation.

Instructor note: the exact inversion of $P=\tfrac43(1-y^2)^{3/2}$ gives back $\varepsilon(0)$ $=\frac1\pi\int_0^14\sqrt{1-y^2}\,dy=1$. With $1\%$ noise and naive differencing on $100$ points, errors of tens of per cent near the axis are typical, because differentiation multiplies the noise by roughly $1/h$; a quadratic fit in $y^2$ reduces the error to well below $1\%$ for this smooth profile. The point to draw out is that the second-kind theory of this section (unique, stable solutions) does not cover first-kind problems.

🧪 Self-check quiz

0/5 correct

Q1. The equation $\int_0^1K(x,t)\,y(t)\,dt=f(x)$ is

Q2. How many characteristic numbers does a continuous Volterra kernel $K(x,t)$ have?

Q3. The IVP $y''+y=0$, $y(0)=0$, $y'(0)=1$ is equivalent to

Q4. For a separable kernel $K=\sum_{i=1}^na_i(x)b_i(t)$, solving $y=f+\lambda\int Ky$ reduces to

Q5. If $\lambda$ is a characteristic number of a separable kernel $K$, the equation $y=f+\lambda\int_a^bKy$ is solvable if and only if