First-order first-degree autonomous differential equation: Difference between revisions

From Calculus
 
(One intermediate revision by the same user not shown)
Line 51: Line 51:
| <math>f(x) := x^\gamma, 0 < \gamma < 1</math> || <math>x = O(t^{1/(1 - \gamma)})</math> (i.e., it grows roughly like a power function of <math>t</math>)
| <math>f(x) := x^\gamma, 0 < \gamma < 1</math> || <math>x = O(t^{1/(1 - \gamma)})</math> (i.e., it grows roughly like a power function of <math>t</math>)
|-
|-
| linear function <math>f(x) := mx, m > 0</math> || exponential function <math>x := \exp(mt)</math>  
| linear function <math>f(x) := mx, m > 0</math> || exponential function <math>x := \exp(mt)</math>
|-
|-
| linear times logarithmic, something like <math>x (\ln x + 1)</math> || <math>x</math> grows something like a doubly exponential function of <math>t</math>
| linear times logarithmic, something like <math>x (\ln x + 1)</math> || <math>x</math> grows something like a doubly exponential function of <math>t</math> (note: if we wanted growth between exponential and double exponential, we would need something like <math>x(\ln x + 1)^\gamma, 0 < \gamma < 1</math>, and if we wanted triple exponential growth, we would multiply by a double logarithmic term)
|-
|-
| <math>f(x) := x^\gamma, \gamma > 1</math> || <math>x</math> grows so fast in terms of <math>t</math> that it reaches <math>\infty</math> in finite time.
| <math>f(x) := x^\gamma, \gamma > 1</math> || <math>x</math> grows so fast in terms of <math>t</math> that it reaches <math>\infty</math> in finite time.
|}
|}

Latest revision as of 21:51, 5 July 2012

Definition

Following the convention for autonomous differential equation, we denote the dependent variable by x and independent variable by t.

Form of the differential equation

The differential equation is of the form:

dxdt=f(x)

Solution method and formula

We convert the differential equation to an integration problem:

dxf(x)=dt

and carry out the integrations on both sides. If p is an antiderivative of 1/f, the solution will be:

p(x)=t+C

with C the freely varying parameter over R: every particular value of C gives a solution. To express x as a function of t, we need to invert p. If we can do so, we'd have:

x=p1(t+C)

as the general solution.

In addition, there may be stationary solutions. These are solutions that correspond to constant functions x=k that satisfy f(k)=0.

Related notions

Analysis

Starting at time zero with value one

Suppose we want to solve the initial value problem for the differential equation:

dxdt=f(x)

subject to the initial condition that at t=0, x=1.

We consider various possibilities for f a function that sends 1 and higher numbers to positive numbers, and make cases based on the growth rate of f:

Nature of f Nature of x in terms of t?
constant function f(x):=c linear function x:=1+ct
f(x):=xγ,0<γ<1 x=O(t1/(1γ)) (i.e., it grows roughly like a power function of t)
linear function f(x):=mx,m>0 exponential function x:=exp(mt)
linear times logarithmic, something like x(lnx+1) x grows something like a doubly exponential function of t (note: if we wanted growth between exponential and double exponential, we would need something like x(lnx+1)γ,0<γ<1, and if we wanted triple exponential growth, we would multiply by a double logarithmic term)
f(x):=xγ,γ>1 x grows so fast in terms of t that it reaches in finite time.