Consider the system
x' = F(x, y, α), y' = G(x, y, α),
where
α is a parameter. The equations
F(x, y, α) = 0, G(x, y, α) = 0
determine the
x- and
y-nullclines, respectively; any point where an
x-nullcline and a
y-nullcline intersect is a critical point. As
α varies and the configuration of the nullclines changes, it may well happen that, at a certain value of
α, two critical points coalesce into one. For further variation in
α, the critical point may once again separate into two critical points, or it may disappear altogether. Or the process may occur in reverse: For a certain value of
α, two formerly nonintersecting nullclines may come together, creating a critical point, which, for further changes in
α, may split into two. A value of
α at which such phenomena occur is a bifurcation point. It is also common for a critical point to experience a change in its type and stability properties at a bifurcation point. Thus both the number and the kind of critical points may change abruptly as
α passes through a bifurcation point. Since a phase portrait of a system is very dependent on the location and nature of the critical points, an understanding of bifurcations is essential to an understanding of the global behavior of the system's solutions.
Consider the following system. (A computer algebra system is recommended.)
x' =
−16x +
y +
x2,
y' =
α − y
(a) Sketch the nullclines. (Let
α1 <
α2 <
α3.)
Describe how the critical points move as
α increases.
(b) Find the critical points.
c1 | | (x, y) | = |
| (smaller x-value) |
c2 | | (x, y) | = |
| (larger x-value) |
(c) Let
α = 14.
Classify each critical point by investigating the corresponding approximate linear system.
c1 | |
|
c2 | |
|
(d) Find the bifurcation point
α0
at which the critical points coincide.
Locate this critical point.
Find the eigenvalues of the approximate linear system. (Enter your answers as a comma-separated list.)