1. Romantic Relationships Background –Life would be very dull without the excitement (and sometimes pain) of romance! –Love affairs can be modelled by differential equations – Stven Strogatz suggested a model, 1988 An example: –Suppose William is in love with Zelda, but Zelda is a fickle lover. –The more William loves her, the more she dislikes him – but when he loses interest in her, her feelings fro him warm up. –William reacts to her: when she loves him, his love for her grows and when she loses interest, he also loses interest.

2. Romantic Relationships Variables: –Independent variable: time: t –State variables: w(t): William’s feelings for Zelda at time t z(t): Zelda’s feelings for William at time t Positive and negative values of w and z denote love and dislike, resp. The exact units for the variables will be left to the imagination of the reader (e.g. the love between Romeo and Juliet can be cannonized in the name of ``romjul’’ for the units!) Assumptions: The change in one person’s feelings in a fixed time interval is directly proportional to the other person’s feelings over the same time interval.

3. Romantic Relationships The Model: –Consider the time interval –Change in the time interval Where a and b are two positive constants The minus sign to show that z decreases when w is positive, and conversely, z increases when w is negative !!

4. Romantic Relationships –Rate of change Linear first order ODE system with constant coefficients Solution –Try ansatz –Plug into the equations, we get

5. Romantic Relationships –Solution form –Initial condition –Solution

6. Romantic Relationships Graph of w(t) and z(t) for the case when a>b

7. Romantic Relationships Interpretation: Periodic solutions –At time t 0, William’s love for Zelda is a maximum and Zelda is neutral in her attitude to William –As time moves on, Zelda starts to dislike William (z(t)<0) and William’s love also diminishes until it is zero at t 1 and Zelda’s dislike is at its worst! –But for t>t 1, William starts to dislike Zelda who is now looking with new eyes at William, but unfortunately William’s dislike for Zelda is at it worst! –For t>t 2, Zelda is falling in love with William, consequently William is also losing his dislike for her until at t 3 William is over his dislike and Zelda is madly in love. –But now, as William is falling in love, Zelda’s love is cooling until at t 4 we have again the situation as at t=t 0. –Now the whole process repeats itself. –William and Zelda are in a perpetual cycle of love and hate, in fact, they love each other simultaneously for only one quarter of the time!!!!

8. Romantic Relationships Phase plane analysis –Solution –Equation –Remark: you can also get the equation from the solution

9. Romantic Relationships Graph in phase plane: an ellipse in (z,w)-plane with t as a parameter

10. Romantic Relationships Interpretation: Periodic solutions –At time t 0, William’s love for Zelda is a maximum and Zelda is neutral in her attitude to William –As time moves on, Zelda starts to dislike William (z(t)<0) and William’s love also diminishes until it is zero at t 1 and Zelda’s dislike is at its worst! –But for t>t 1, William starts to dislike Zelda who is now looking with new eyes at William, but unfortunately William’s dislike for Zelda is at it worst! –For t>t 2, Zelda is falling in love with William, consequently William is also losing his dislike for her until at t 3 William is over his dislike and Zelda is madly in love. –But now, as William is falling in love, Zelda’s love is cooling until at t 4 we have again the situation as at t=t 0. –Now the whole process repeats itself. –William and Zelda are in a perpetual cycle of love and hate, in fact, they love each other simultaneously for only one quarter of the time!!!!

11. Romantic Relationships Demo: http://www.aw-bc.com/ide/idefiles/media/JavaTools/romejuli.html –Case 1: a=0 & b=0 –Case 2: a=1 & b=0 –Case 3: a=0 & b=1 –Case 4: a=1 & b=1 –Case 5: a=1 & b=-1 –Case 5: a=1 & b=0.1 –Case 6: a=0.1 & b=1

12. Romantic Relationships Two ways to look at the solutions w(t) & z(t) –First way: look at the graphs of w(t) & z(t) as functions of t –Second way: eliminate the independent variable (if possible) to obtain a relation between w and z. Phase Plane: The plane whose points can be identified with coordinates on the (z,w)-axes. For autonomous system, we can always draw the phase plane.

13. Marriage Background: population model re-visited –An example: Malthus Model b & d denote the birth rate and the mortality rate, respectively. In general, these birth and mortality rates will differ in the population; for example, the birth rate will be close to zero in a home for senior citizens (let us not underestimate them!) Here, we will look at subsets of the population and their interaction in the process of population growth

14. Marriage We subdivide the population into three classes: –(i) unmarried males –(ii) unmarried females –(iii) married persons Assumptions –The marriage is monogamous (a wife & a husband!!!) –A person who are married, but is now single due to divorce or the death of the husband/wife, is unmarried –Only married couples produce children

15. Marriage Interaction between different classes –Married couples produce children who are unmarried –Unmarried persons decide to get married –The growth of each of the three classes is interrelated!! Variables –Independent variable: time: t –State variables: m(t): number of unmarried males in the population at time t f(t): number of unmarried females in the population at time t w(t): number of married couples in the population at time t

16. Marriage The model –The rate of change of m(t) depends on The mortality rate a of the unmarried male population The rate b at which unmarried females get married The birth rate c of (unmarried !) males produced by the married couples The mortality rate d of married females

17. Marriage –The rate change of f(t) depends on The rate b at which unmarried females get married The mortality rate e of unmarried female population The birth rate g of females produced by married couples The mortality rate h of married males

18. Marriage –The rate change of w(t) depends on The rate b at which unmarried females get married The mortality rate h of married males The mortality rate d of married females The equations

19. Marriage In vector (or matrix) form: –Denote –ODE system

20. Marriage Analytical solution –If B is diagonalizable: Discussions –Nondimensionalization –Equilibrium and stability

21. Analytic & Numerical Solutions of 1st Order ODEs General form of first order ODE system: –t: independent variable (time, in dimensionless form) – : state variable – a vector function of n+1 variables Solution: Y=Y(t) satisfies the equation –Existence & uniqueness ??? –Same theorem as for single equation –Infinitely many different solutions!!!

22. Classification –Linear ODE system, i.e. the function F is linear in Y –Otherwise, it is nonlinear –Autonomous ODE system, i.e. F is independent of t Some cases which can be solved analytically

23. Solutions of ODE Solution structure of simple equation –If Y 1 (t) is a solution, then Y 1 (t)+c is also a solution for any vector c ! –If Y 1 (t) and Y 2 (t) are two solutions, then there exists a constant vector c such that Y 2 (t)=Y 1 (t)+c. –If Y 1 (t) is a specific solution, then the general solution is (or any solution can be expressed as) Y 1 (t)+c

24. Solutions of ODE Infinitely many different solutions!!! Solution structure of linear homogeneous problem –If Y 1 (t) is a solution, then c Y 1 (t) is also a solution for any constant c ! –If Y 1 (t) and Y 2 (t) are two nonzero solutions, then c 1 Y 1 (t) + c 2 Y 2 (t) is also a solution (superposition).

25. Solutions of ODE Solution structure of linear problem –If Y 1 (t) is a solution of the homogeneous equation, Y 2 (t) is also a specific solution. Then Y 2 (t)+c Y 1 (t) is also a solution for any constant c ! –If Y 1 (t) and Y 2 (t) are two solutions, Y 1 (t)-Y 2 (t) is a solution of the homogeneous equation.

26. Solutions of IVP Initial value problem (IVP): Solution: –Existence –Uniqueness

27. Solutions of IVP Theorem: Suppose that the real-valued function F(t,Y) is continuous on some rectangle in the (t,Y)-plane containing the point (t 0,Y 0 ) in its interior. Then the above initial value problem has at least one solution defined on some open interval J containing the point t 0. If, in addition, the partial derivative is continuous on that rectangle, then the solution is unique on some (perhaps smaller) open interval J 0 containing the point t=t 0. Proof: Omitted

28. Solution: Case I –F is independent of y, i.e. Analytical solution Example 1: –Solution

29. Solution: case II Linear homogeneous system Analytical solution –If B is diagonalizable: –The solution –An example

30. Equilibrium & stability Consider autonomous ODE Equilibrium: Asymptotically stable: Conditions: –Stable: real parts of all eigenvalues of the Jacobian matrix are negative –Unstable: at least the real part of one eigenvalue of the Jacobian matrix is positive

31. Equilibrium & satiability Example: –Romantic relationships model –Equilibrium: w=z=0 –Jacobian matrix –Two eigenvalues –Unstable!!!

32. Phase plane portrait For autonomous ODE system with two state variables Two ways to look at the solutions x(t) & y(t) –First way: look at the graphs of x(t) & y(t) as functions of t –Second way: eliminate the independent variable (if possible) to obtain a relation between y and x. Phase Plane: The plane whose points can be identified with coordinates on the (x,y)-axes. For autonomous system, we can always draw the phase plane. Examples –Lotka-Voleterra system –Romantic relationships

33. Numerical methods Consider initial-value problem (IVP): Choose a time step and partition [0,T] into n pieces Denote

34. Numerical methods –4 th order Runge-Kutta (RK) method