15.08.2015 Views

Introduction to the Modeling and Analysis of Complex Systems

introduction-to-the-modeling-and-analysis-of-complex-systems-sayama-pdf

introduction-to-the-modeling-and-analysis-of-complex-systems-sayama-pdf

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

5.7. LINEAR STABILITY ANALYSIS OF DISCRETE-TIME NONLINEAR... 91around its equilibrium point x eq . By definition, x eq satisfiesx eq = F (x eq ). (5.57)To analyze <strong>the</strong> stability <strong>of</strong> <strong>the</strong> system around this equilibrium point, we switch our perspectivefrom a global coordinate system <strong>to</strong> a local one, by zooming in <strong>and</strong> capturing asmall perturbation added <strong>to</strong> <strong>the</strong> equilibrium point, ∆x t = x t − x eq . Specifically, we apply<strong>the</strong> following replacementx t ⇒ x eq + ∆x t (5.58)<strong>to</strong> Eq. (5.56), <strong>to</strong> obtainx eq + ∆x t = F (x eq + ∆x t−1 ). (5.59)The right h<strong>and</strong> side <strong>of</strong> <strong>the</strong> equation above is still a nonlinear function. If x t is scalar<strong>and</strong> thus F (x) is a scalar function, <strong>the</strong> right h<strong>and</strong> side can be easily approximated using<strong>the</strong> Taylor expansion as follows:F (x eq + ∆x t−1 ) = F (x eq ) + F ′ (x eq )∆x t−1 + F ′′ (x eq )2!∆x 2 t−1 + F ′′′ (x eq )∆x 3 t−1 + . . .3!(5.60)≈ F (x eq ) + F ′ (x eq )∆x t−1 (5.61)This means that, for a scalar function F , F (x eq + ∆x) can be linearly approximated by <strong>the</strong>value <strong>of</strong> <strong>the</strong> function at x eq plus a derivative <strong>of</strong> <strong>the</strong> function times <strong>the</strong> displacement fromx eq . Toge<strong>the</strong>r with this result <strong>and</strong> Eq. (5.57), Eq. (5.59) becomes <strong>the</strong> following very simplelinear difference equation:∆x t ≈ F ′ (x eq )∆x t−1 (5.62)This means that, if |F ′ (x eq )| > 1, ∆x grows exponentially, <strong>and</strong> thus <strong>the</strong> equilibrium pointx eq is unstable. Or if |F ′ (x eq )| < 1, ∆x shrinks exponentially, <strong>and</strong> thus x eq is stable.Interestingly, this conclusion has some connection <strong>to</strong> <strong>the</strong> cobweb plot we discussed before.|F ′ (x eq )| is <strong>the</strong> slope <strong>of</strong> function F at an equilibrium point (where <strong>the</strong> function curvecrosses <strong>the</strong> diagonal straight line in <strong>the</strong> cobweb plot). If <strong>the</strong> slope is <strong>to</strong>o steep, ei<strong>the</strong>rpositively or negatively, trajec<strong>to</strong>ries will diverge away from <strong>the</strong> equilibrium point. If <strong>the</strong>slope is less steep than 1, trajec<strong>to</strong>ries will converge <strong>to</strong> <strong>the</strong> point. You may have noticedsuch characteristics when you drew <strong>the</strong> cobweb plots. Linear stability analysis <strong>of</strong>fers ama<strong>the</strong>matical explanation <strong>of</strong> that.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!