When the variables do not separate
cannot be written as f(x) times g(y), so the variables do not separate as it stands. The question gives a substitution, u = x + y, that turns it into an equation that does.
The first step is to differentiate the substitution with respect to x: . This links the new derivative to the old one.
Replace every y
From , . Put that and u = x + y into the equation: , so .
Only u and x are left, and the right side depends on u alone, so the variables separate.
Solve, then return to y
Separate and integrate: , so and u = tan(x + C).
The answer must be in x and y, so put x + y back for u: x + y = tan(x + C), and y = tan(x + C) − x.
Check by differentiating: , and tan(x + C) is x + y, so .
The substitution y = vx
is solved with y = vx, where v is a new variable that depends on x. Since y is a product of v and x, the product rule gives .
Substitute: , so . Separate: , so v = ln|x| + C.
Multiply back by x: y = x ln|x| + Cx. Check: , and .
Two more examples
with u = 2x + y. Now , so . Integrating gives , and then , where .
with y = vx. The right side is , so and . Then gives , and with , .
The usual slips
Treating as . For u = 2x + y, : the 2x term changes with x too.
Writing for y = vx. v changes with x, so the product rule gives .
Stopping at u or v. The question asks for y, so put the substitution back and make y the subject.