A first-order differential equation is separable if it can be rearranged so that every y (and dy) is on one side and every x (and dx) is on the other — that is, written as h(y)y′=g(x), or equivalently as f1(x)g1(y)dx+f2(x)g2(y)dy=0.
Method.
- Rearrange the given equation into the separated form
f2(x)f1(x)dx=−g1(y)g2(y)dy.
- Integrate both sides independently:
∫f2(x)f1(x)dx=−∫g1(y)g2(y)dy+C.
Only one arbitrary constant C is needed — the two constants that would arise from integrating each side separately combine into a single overall constant.
3. If an initial condition is given (e.g. y=y0 at x=x0), substitute it into the integrated equation to evaluate C and obtain the particular solution.
"Solving" a differential equation is therefore also called "integrating" it, since the whole process reduces to two ordinary integrations once the variables are separated.
Recognising a separable equation in disguise. Many equations that do not look separable at first become separable after a short algebraic step:
- Product-to-sum trig identities — e.g. cos(x+y)+cos(x−y)=2cosxcosy — turn a mixed trigonometric equation into one where x and y split apart.
- Recognising an exact differential — noticing that y2ydx−xdy=d(yx) lets an equation be separated in the single combined variable yx directly, without a full substitution.
- A constant a on one side (e.g. sindxdy=a) simply gives dxdy=sin−1a, a constant slope, integrating to a straight line.
Substitution method. When the right side is a function of a single linear combination ax+by+c rather than of x and y separately — i.e. dxdy=f(ax+by+c) — put z=ax+by+c. Then dxdz=a+bdxdy, so dxdy=b1(dxdz−a) (for b=0), and substituting turns the equation into one relating z and x only — which is separable. Solve for z by the separable method, then replace z by ax+by+c to get the answer back in x,y. (If a=0 or b=0, the original equation is already separable and no substitution is needed.)