Math 5110/6830 Homework 4.1 Solutions 1. (a) A fixed point satisfies the equation x∗ = r2 x∗ [1 − (r + 1)x∗ + 2rx∗ 2 − rx∗ 3 ]. x∗ = 0 is immediately a solution. Now, notice that r−1 r is a fixed point of the original logistic equation, f (x). So, it satisfies x∗ = f (x∗ ). Since f 2 (x) = f (f (x)), and a fixed point of f 2 (x) satisfies x∗ = f (f (x∗ )), it follows that if x∗ is a fixed point of the original equation, f (f (x∗ )) = f (x∗ ) = x∗ , and so x∗ is a fixed point of the second-iterate equation. Thus, 2 2 2 3 x∗ = r−1 r is a fixed point of f (x) = r x[1 − (r + 1)x + 2rx − rx ]. As a fixed point, r−1 r must be root of the following equation: −r2 [1 − (r + 1)x + 2rx2 − rx3 ] + 1 = 0, which implies that x − r−1 r must be a factor of the left-hand side (LHS). We can therefore divide the LHS by x − r−1 r to determine the other two roots of the equation. r3 x2 − r2 (r + 1)x + r(r + 1) x− r−1 r r 3 x3 − 2r3 x2 + r2 (r + 1)x + (1 − r2 ) r3 x3 − r2 (r − 1)x2 − r2 (r + 1)x2 + r2 (r + 1)x − r2 (r + 1)x2 + r(r2 − 1)x r(r + 1)x + (1 − r2 ) r(r + 1)x − (r2 − 1) From this, it follows that r3 x2 − r2 (r + 1)x + r(r + 1) = r2 x2 − r(r + 1)x + (r + 1) = 0, which has the roots p r(r + 1) ± r2 (r + 1)2 − 4r2 (r + 1) x1,2 = 2 √ 2r r(r + 1) ± r r2 + 2r + 1 − 4r − 4 = 2r2 √ r + 1 ± r2 − 2r − 3 = p2r r + 1 ± (r − 3)(r + 1) = , 2r √ r+1± (r−3)(r+1) and so the second-iterate map has the 4 fixed points x∗ = 0, r−1 , . r 2r √ √ r+1+ (r−3)(r+1) r+1− (r−3)(r+1) (b) The fixed points u = and v = form the 2-cycle of the 2r 2r original system. The 2-cycle exists when these two fixed points are distinct, real, defined, and positive. They are distinct and real whenever (r − 3)(r + 1) > 0, which occurs when r < −1 and r > 3. Since r > 0, the fixed points are defined,√and r > 3 must be true. Further, r+1− (r−3)(r+1) > 0, which occurs when u is always positive, but v is only positive when 2r p r + 1 − (r − 3)(r + 1) > 0. Multiplying both sides by the conjugate (which is positive) 1 gives the inequality (r + 1 − p p (r − 3)(r + 1))(r + 1 + (r − 3)(r + 1)) > 0 ⇒ (r + 1)2 − (r + 1)(r − 3) > 0 ⇒ 4(r + 1) > 0. Since we already know that r > 3 must be true, this inequality will always hold. Thus, we have a 2-cycle for r > 3. The other points, as mentioned earlier, correspond to the fixed points of the original logistic equation. (c) d 2 d 2 r x(1 − (r + 1)x + 2rx2 − rx3 ) f (x) = dx dx = r2 (1 − (r + 1)x + 2rx2 − rx3 ) + r2 x(−(r + 1) + 4rx − 3rx2 ) = r2 − 2r2 (r + 1)x + 6r3 x2 − 4r3 x3 (d) To determine when the non-trivial 2-cycle is stable, we need to find out when the fixed points of the second-iterate map that corresponding to the 2-cycle are stable. Stability d 2 of the 2-cycle occurs when dx f (x) x=u,v < 1. After some algebra, it turns out that d 2 d 2 f (x) = f (x) = −r2 + 2r + 4. So, stability exists when | − r2 + 2r + 4| < 1. dx dx x=u x=v We need to solve two inequalities: i. −r2 + 2r + 4 < 1 ⇒ −r2 + 2r + 3 < 0 ⇒ r > 3 √ ii. −r2 + 2r + 4 > −1 ⇒ −r2 + 2r + 5 > 0 ⇒ 0 < r < 1 + 6 √ Therefore, the 2-cycle is stable for 3 < r < 1 + 6. 2. (a) For r = 3.2 and x1 = 0.513: x1 0.5130 x2 0.7995 x3 0.5130 x4 0.7995 x5 0.5130 x6 0.7995 x7 0.5130 x8 0.7995 x9 0.5130 x10 0.7995 1 xn+1=f(xn) 0.9 0.7 xn+1=xn cobweb 0.8 0.6 0.7 xn 0.5 0.6 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 0 0 1 2 3 4 5 6 7 8 9 10 0 0.2 0.4 0.6 0.8 1 n The 2-cycle bounces between 0.5130 and 0.7995. (b) For r = 3.55 and x1 = 0.8874: x1 0.8874 x2 0.3547 x3 0.8126 x4 0.5407 x5 0.8816 2 x6 0.3705 x7 0.8279 x8 0.5057 x9 0.8874 x10 0.3548 1 xn+1=f(xn) 0.8 0.9 0.7 0.8 xn+1=xn cobweb 0.7 0.6 0.6 xn 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 0 0 1 2 3 4 5 6 7 8 9 10 0 0.2 0.4 0.6 0.8 1 n There may be a pattern that restarts at x9 , depending on the decimal place to which the solution is rounded. (c) For r = 3.8 and x1 = 0.5: x1 0.5 x2 0.95 x3 0.1805 x4 0.5621 x5 0.9353 x6 0.2298 x7 0.6726 x8 0.8369 x9 0.5188 x10 0.9487 1 xn+1=f(xn) 0.9 0.9 xn+1=xn 0.8 cobweb 0.8 0.7 0.7 0.6 xn 0.6 0.5 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 0 1 2 3 4 5 6 7 8 9 10 0 0 0.2 0.4 0.6 n There doesn’t seem to be any observable pattern in the solution. Matlab code r = 3.2; % define parameter (change manually for parts b and c) x = []; % defines x as an empty vector x(1) = 0.513; % sets the first entry in x to be the given x_1 (change % manually for parts b and c) f = @(x) r*x.*(1-x); % define map as a function of x N=10; % maximum n % this for-loop evaluates the solution through n=10 and puts % the solution at each successive time point in the vector x for n=1:N-1 x(n+1) = f(x(n)); end 3 0.8 1 % print the solution as a column vector in the command window % (no semi-colon) x' % plot and label the solution in the specified range plot(x,'-o') xlabel('n') ylabel('x_n') axis([1 N 0 max(x)]) Note: Any single quotations in the code will not copy correctly into Matlab (a LA TE X issue), so you will need to retype them after pasting the code into the Editor. 3. Feigenbaum diagram of the logistic map 1 0.9 0.8 0.7 Stability of x*=(r−1)/r for 1<r<3 x50 0.6 0.5 Stable 2−cycle chaos 0.4 0.3 Stable 4−cycle 0.2 Stability of x*=0 for 0<r<1 0.1 0 0 0.5 1 1.5 2 r 4 2.5 3 3.5 4 Homework 4.2 Solutions 1. (a) Fixed points: p∗ = p∗ 2 − q ∗ q ∗ = p∗ q ∗ − 100p∗ has 3 solutions in the form P = (p∗ , q ∗ ): P1 = (0, 0), P2 = (11, 110) and P3 = (−9, 90) . Stability: The Jacobian for this system is 2p∗ −1 ∗ ∗ J(p , q ) = ∗ q − 100 p∗ 0 −1 For P1 = (0, 0): J(0, 0) = , which has eigenvalues λ1 = −10 and λ2 = 10. −100 0 Since |λ1,2 | > 1, P1 is unstable. 22 −1 , which has eigenvalues λ1 = 12 and λ2 = 21. For P2 = (11, 110): J(11, 110) = 10 11 Since |λ1,2 | > 1, P2 is unstable. −18 −1 , which has eigenvalues λ1 = −19 and For P3 = (−9, 90): J(−9, 90) = −10 −9 λ2 = −8. Since |λ1,2 | > 1, P3 is unstable. (b) Fixed points: x∗ = rx∗ − x∗ 2 + x∗ has 2 solutions, x∗ = 0, r . Stability: Since this is a one-dimensional system, we can look at the derivative of the righthand side evaluated at the fixed points to determine stability. If f (xn ) = rxn − x2n + xn , then f 0 (xn ) = r − 2xn + 1. For x∗ = 0: f 0 (0) = r + 1, which implies stability provided |r + 1| < 1. But, since r > 0, this can never happen. Thus, the fixed point x∗ = 0 is always unstable. For x∗ = r: f 0 (r) = 1 − r, which implies stability provided |1 − r| < 1. So the fixed point x∗ = r is stable when 0 < r < 2 and unstable for r > 2. Bifurcation diagram (solid line: stable, dashed line: unstable) 5 5 4.5 4 3.5 x * 3 2.5 2 1.5 1 0.5 0 0 1 2 3 4 5 r 2. (a) When aR + pJ = 1 and aJ + pR = 1, Jn + Rn = Jo + Ro . We can see this by adding Rn+1 and Jn+1 . Rn+1 + Jn+1 = (aR + pJ )Rn + (aJ + pR )Jn Because the total amount of feeling (Jn + Rn ) is preserved, this mode is called feeling preserving. (b) Fixed points: R ∗ = aR R ∗ + p R J ∗ J ∗ = p J R ∗ + aJ J ∗ 0 = (aR − 1)R∗ + pR J ∗ J ∗ = pJ R∗ + (aJ − 1)J ∗ ∗ 0 aR − 1 pR R = 0 J∗ pJ aJ − 1 The determinant of this matrix is (aR − 1)(aJ − 1) − (pR pJ ). Substituting 1 − aR = pJ and 1 − aJ = pR , we see that the determinant is 0. Because the determinant is 0, we know that pJ the two equations are redundant and we can write J ∗ in terms of R∗ . So J ∗ = 1−a R∗ . J (c) The Jacobian matrix is: a p A= R R pJ aJ To find the eigenvalues we must find the determinant of A − λI. aR − λ pR A − λI = pJ aJ − λ Then, the det A − λI = (aR − λ)(aJ − λ) − pR pJ or det A − λI = λ2 − λ(aR + aJ ) + (−1 + aJ + aR ). To find the values of λ set detA − λI = 0. 6 Therefore, p 1 λ1,2 = (aR + aJ ± (aR + aJ )2 − 4(−1 + aJ + aR )) 2 p 1 λ1,2 = (aR + aJ ± (2 − aR − aJ )2 ) 2 λ1 = 1 λ2 = aR + aJ − 1 7