Which function has derivative ? Every function of the form does, one for each value of the constant, so the equation with describes a whole family of curves rather than a single one. Picking one member takes extra data, given either at one point or at the two ends of an interval, and the two kinds of data behave very differently. This lesson sorts out the kinds of solutions, shows how to verify a solution completely, and explains why every solution comes with an interval.
Key terms
Explicit and implicit solutions
Definition Let . An explicit solution of an equation of order on an open interval is a function that is times differentiable on and satisfies the equation at every in . A relation is an implicit solution on if it defines at least one explicit solution on , that is, a function with for every in that solves the equation there. A relation is checked by implicit differentiation, treating as a function of , so the chain rule brings in a factor .
In plain words An explicit solution hands you as a formula in . An implicit one hands you a curve, and the solutions are the pieces of that curve that are graphs of differentiable functions.
Example The circle is an implicit solution of on . Differentiating the relation gives , so wherever . The full circle is not the graph of a function, because each in is the first coordinate of two points of the circle. It splits into two explicit solutions on , the upper and the lower half,
For the upper half the derivative is , which equals , and the lower half checks the same way.
General and particular solutions
Definition A general solution of an equation on an interval is a formula with arbitrary constants that produces every solution on as the constants vary. For a linear equation of order whose coefficients and forcing term are continuous on , with a leading coefficient that never vanishes on , it contains exactly constants. A particular solution is one member of the family, with every constant fixed.
In a nonhomogeneous linear equation the phrase has a second use. There a particular solution is any one solution of the forced equation, chosen without regard to initial data, and the general solution is plus the general solution of the homogeneous equation with the same left side.
In plain words The general solution is the whole collection of answers, with its constants still free. A particular solution is one curve from that collection. The surrounding problem tells you whether a particular solution means the member selected by the data or just some solution of a forced equation.
Example Every solution of on has the form , because two functions with the same derivative on an interval differ by a constant. So is the general solution, and is a particular solution. For the general solution is , with two constants for a second-order equation. For the forced equation , the function is a particular solution, since , and the general solution is .
A family found by a calculation can miss solutions. The family solves , and so does the zero function, which no choice of produces, so that family alone is not the general solution.
Solution verification
Definition To verify that a function solves a problem on an interval , check three things separately. First, is defined on and has there every derivative the equation needs. Second, substituting and its derivatives into the equation gives an identity, true for every in . Third, meets each initial or boundary condition. Passing one check says nothing about the others.
In plain words You never have to solve an equation to test an answer. Substitute and simplify, then check the data and the interval.
Example Test against with . Its derivative is for every real , and , so it passes every check on . The function has the right derivative but gives . The function passes through , yet its derivative equals only at , so it does not satisfy the equation on any interval. Each wrong candidate fails exactly one of the checks.
Domain of a solution
Definition The domain of a solution is the interval on which it is claimed to solve the equation. It must be an interval. It must leave out every point where the function or one of its needed derivatives fails to exist and every point where the equation itself is undefined. For an initial-value problem it must contain the initial point. The set where a formula makes sense can be larger than this interval, or not an interval at all.
In plain words A solution is a formula together with an interval. The same formula on two separate intervals counts as two different solutions.
Example The function solves , since . Its formula makes sense for every , which is not an interval, so it gives one solution on and another on . With the condition , the solution is on only. In the same way, the upper half of the circle is continuous on , but at it has vertical tangents and the equation would divide by zero. Its domain as a solution is the open interval .
Initial-value problem
Definition An initial-value problem for an equation of order consists of the equation together with the values of at one point . For a first-order equation it reads
A solution is a function on an open interval containing that solves the equation there and meets every condition. For a linear equation of order whose coefficients and forcing term are continuous on an open interval , with a leading coefficient that never vanishes on , every initial-value problem with in has exactly one solution, and that solution exists on all of .
In plain words Initial data describe the state at one moment, a height for a first-order equation, and a height and a slope for a second-order one. For linear equations that meet these conditions, that is exactly enough information to pick out one solution.
Example Solve with and . Start from , so . The derivative is , so . The solution is
on all of . As a check, , the value at is , and the slope at is .
Boundary-value problem
Definition A boundary-value problem imposes conditions at two or more different points, usually the endpoints of an interval, for example and for a second-order equation. Other boundary conditions prescribe , or a combination of and , at an endpoint. A solution satisfies the equation on the interval from to and every boundary condition. Even for a linear equation with continuous coefficients and forcing term and a leading coefficient that never vanishes, where every initial-value problem has exactly one solution, a boundary-value problem can have exactly one solution, infinitely many or none, and which case occurs is decided by solving the boundary equations for the constants.
In plain words The data are split between two ends. The equation fixes how the curve bends, and the question is whether that bending can be made to hit both targets.
Example Take with general solution . The condition forces and leaves . Each of the three cases then occurs.
- With on , the second condition gives , so the only solution is .
- With on , every works, because , so there are infinitely many solutions.
- With on , the second condition reads for every , so there is no solution.
Check yourself
- Find the solution of with , and verify it.
Show the answer
Every solution on has the form . The condition gives , so and . To verify it, note that the derivative of is for every real and that . The solution is valid on all of .
- A classmate solves with and , finds , and calls it the solution. What is wrong with that conclusion? What changes if the second condition is instead?
Show the answer
Start from . The first condition gives . The second gives , which then holds automatically and puts no restriction on . So every function solves the problem, and is only the case . The boundary-value problem has infinitely many solutions.
With the second condition demands , which contradicts , so there is no solution at all.
- Show that the relation is an implicit solution of . Find the explicit solution through the point and give its interval.
Show the answer
Differentiating with respect to gives , so wherever . The point lies on the lower branch, so the explicit solution is . Its derivative is
since . The interval is . At the branch reaches , where the equation would divide by zero and is not differentiable, and for no real satisfies .