Exact Differential
Exact Differential
- Implying, that if we have a differential equation $$M(x,y)\dd x + N(x,y)\dd y = 0$$and $$\frac{\partial }{\partial y}M(x,y) = \frac{\partial}{\partial x}N(x,y)$$then the differential equation is exact.
- Solution of this type is obviously
.
How to solve exact differential types
Integrating factors
- When the ODE is not exact, but can be made exact by multiplying with an arbitrary function
, then is called the integrating factor of the differential equation - We claim that the solution set of both these equations are "essentially same", but it might result in
- Loss of one or more solutions of the original
- Gain of one or more new solutions which are not solutions of the original
- Or, both of these can happen simultaneously.
Comment
If an equation has an integrating factor, then it is infinitely many integrating factors.
Finding integrating factors
- If $$\frac{1}{N}\left( \frac{\partial M}{\partial y} - \frac{\partial N}{\partial y} \right) = p(x)$$with
being a function of only, then the integrating factor is
$$e^{\int p(x)dx}$$
- A first order linear equation of the fashion $$y' +p(x)y = r(x)$$has an integrating factor of
, thus the solution being $$e^{\int p(x)dx}\times y = \int^xr(s)e^{\int p(s)ds}ds + C$$ - In a similar fashion if $$-\frac{1}{M}\left( \frac{\partial M}{\partial y} - \frac{\partial N}{\partial x} \right) = g(y)$$then integrating factor is
$$e^{\int g(y)dy}$$
- Once we find the integrating factor, we multiply and try to reduce to exact differential, or apply the method to solve exact differentials.
Condition for arbitrary integrating factor
If we want an integrating factor as a function of another function, ie $$\mu = \exp\left( \int f(z)dz \right)$$where