[[
wikihub
]]
Search
⌘K
Explore
People
For Agents
Sign in
Explore
People
For Agents
Sign in
@jemoka / Jemoka Knowledge Base / wiki/concepts/non_linear_ode.md
Suggest edit
Cancel
Submit suggestion
Title
Name
Note
--- title: "Non-Linear ODE" type: concept related: [Ode Linearilzation, Mesoscopic Region] source: https://www.jemoka.com/posts/kbhnon_linear_ode/ confidence: high status: active --- Suppose we analyze first order non-linear system: \begin{equation} x’ = F(t,x) \end{equation} We can actually turn this into an autonomous system: \begin{equation} x_0 = t \end{equation} \begin{equation} x_0’ = 1 \end{equation} meaning suddenly we have an autonomous system: \begin{equation} \begin{cases} x_0’ = 1 \\ x_1’ = F(x_0, x_1) \end{cases} \end{equation} General strategy: Find zeros of the right side (which are the stationary solutions) Analyze near-stationary solutions through eigenvalues of the linearized Jacobian matrix: if both eigenvalues are zero Away from stationary solutions: basically guessing Three Examples that are Hopeless to Solve Lotha-Volterra Prey-Predictor Equation \begin{equation} \begin{cases} x_1’ = 2x_1-x_1x_2 \\ x_2’ = x_1x_2 - 3x_2 \end{cases} \end{equation} By default, if either \(x_1\) or \(x_2\) goes down, the system dies quickly. Example \begin{equation} \begin{cases} x_1’ = r_1x_1 \qty(1- \frac{x_1 + h_{12} x_2}{k_1})\\ x_2’ = r_2x_2 \qty(1- \frac{x_2 + h_{21} x_1}{k_2}) \end{cases} \end{equation} Example \begin{equation} \begin{cases} x_1’ = x_2 \\ x_2’ = -\sin x_1 - \gamma x_2 \end{cases} \end{equation} Strategy to Analyze when its Hopeless find a stationary solutions: \(x(t) = a\): where \(x’ = F(a) = 0\) and draw them as points on the \(x_1\) and \(x_2\) plane near each equilibrium point, approximate through Linearilzation study the mesoscopic region So, see ODE linearilzation. Phase Portrait Phase Portrait is a figure in the \(x_1, x_2\) plane where each solution exists as a curve on the figure. monotone function for linearilzation systems that are marginal (zero, negative real parts, one or more fully imaginary), we can’t use linearilzation itself to analyze the system. Therefore, we have to use a function for which \(\dv t V(y(t)) \geq 0\) or \(\dv V(y(t)) \leq 0\) for all \(t\) called a monotone function, which could give us hints about the function’s behavior. Meaning: \begin{align} \dv t V(y(T) &= \nabla V(y(t)) \cdot y’(t) \\ &= \nabla V(y(t)) \cdot F(y(t)) \end{align} The gradient of \(V\) is always perpendicular to the level curve of \(V\), and—when dotted with \(F\) the vector field of $y$—we obtain a value that’s either positive or negative. When positive, the angle between the vector field \(F\) and \(V\) would be less than \(\frac{\pi}{2}\), meaning the vector field point “outwards” from the level sets. Otherwise, it would be more than \(\frac{\pi}{2}\), meaning the vector field point “inwards”. conserved function its like a monotone function, but \(\dv{V}{t} = 0\). any solution curve would lie inside a level curve of \(V\) (parts of the level curve). Its basically the intuition of a monotone function, but the solution curves instead of pointing inwards and outwards, it just get stuck.