15-382 C OLLECTIVE I NTELLIGENCE β S18 L ECTURE 10: D ISCRETE -T IME D YNAMICAL S YSTEMS 1 I NSTRUCTOR : G IANNI A. D I C ARO
(M ORE ) G ENERAL DEFINITION OF DYNAMICAL SYSTEMS A dynamical system is a 3-tuple π,π, Ξ¦ : π is a set of all possible states of the dynamical system (the state space) Β§ π is the set of values the time (evolution) parameter can take Β§ Ξ¦ is the evolution function of the dynamical system, that associates to Β§ each π β π a unique image in π depending on the time parameter π’, (not all pairs (π’, π) are feasible, that requires the subset πΈ) Ξ¦:πΈ β πΓπ β π Γ Ξ¦ 0, π = π (the initial condition) Γ Ξ¦ π’ 2 ,Ξ¦ π’ 3 ,π = Ξ¦(π’ 2 + π’ 3 ,π) , (property of states) for π’ 3 , π’ 3 + π’ 2 β π½(π¦) , π’ 2 β π½(Ξ¦(π’ 3 π¦)) , π½ π¦ = {π’ β π βΆ (π’, π) β πΈ} Γ The evolution function Ξ¦ provides the system state (the value ) at time π’ for any initial state π Γ πΏ ; = {Ξ¦ π’,π βΆ π’ β π½ π } orbit of the system through π , starting in π , the set of visited states as a function of time 2
T YPES OF D YNAMICAL SYSTEMS Β§ Informally: A dynamical system defines a deterministic rule that allows to know the current state as a function of past states Given an initial condition π < = π(0) β π , a deterministic trajectory Β§ π π’ , π’ β π½(π < ) is produced by π, π,Ξ¦ Β§ States can be βanythingβ mathematically well-behaved that represent situations of interest Continuous time dynamical systems (Flows): π open interval of β , Β§ Ξ¦ continous and differentiable function Γ Differential equations π¦Μ = π(π¦ π’ β π ) Delay models Β§ Ordinary Differential Equations E π¦Μ = π π¦ π’ + β« π π¦ π ππ Β§ Integro-Differential E F Equations , accounting for history I H I H 3 IE H π π¦, π’ = I; H π π¦,π’ Β§ Partial Differential Equations , G H accounting for space and time Discrete-time dynamical systems (Maps): π interval of β€ , Ξ¦ a function Β§ π¦ K = π(π¦ KL3 ,π¦ KL2 ,β¦, π¦ KLN ) , iterated updating 3
F ROM LOCAL RULES TO GLOBAL BEHAVIORS ? ππ ππ’ = π(π,π’) Flows Maps π¦ K = π(π¦ KL3 ,π¦ KL2 ,β¦, π¦ KLN ) βπ’ = 1 , when βπ’ β 0 Γ ππ’ Γ Differential eq. Β§ For an infinitesimal time, only the instantaneous variation, the velocity , makes sense Γ The next state is expressed implicitly, and all the instantaneous variations, local in time, must be integrated in order to obtain the global behavior π(π’) Β§ Also in maps, the time-local iteration rule is a local description that can give rise to extremely complex global behaviors Β§ Γ How do we integrate the local description into global behaviors? Β§ Γ How do we predict global behaviors from the local descriptions? 4
M APS Β§ Where maps can arise from? Β§ Inherently discrete-time processes: looking at populations in terms of generations, epidemics in terms of weeks, economy in terms of quarters or years, traffic models per hour, growth per days, β¦ Β§ Discretization of differential equations : Euler method: πΜ = π(π) Γ π KR3 = π K + βπ(π K ) , β¦Runge-Kutta,β¦ Β§ Β§ Discretization of algebraic equations: Newtonβs method for solving π π¦ = 0 Γ Expand in Taylor series Β§ near π¦ K :π π¦ = π π¦ K + π¦ β π¦ K π T π¦ K + β― taking the usual linear approximation: π π¦ β π π¦ K + π¦ β π¦ K π T π¦ K , equating to 0: π¦ KR3 = π¦ K β π π¦ K /π T π¦ K Β§ Letβs focus on one-dimensional maps Β§ Even in one dimension, iterated maps can produce incredibly complex behaviors, including deterministic chaos! Β§ Later on, we will consider multi-dimensional maps defined over a lattice (spatial grid) Γ Cellular Automata Γ even more complex behaviors 5
M APS : SAME TERMINOLOGY AS IN FLOWS π π¦ = 2π¦ , π is a map Β§ The orbit of π¦ under the map π is the set of points: Β§ ,β¦ } = {π¦, π π¦ , π 2 π¦ ,π Y π¦ ,β¦ } {π¦,π π¦ ,π π π¦ , π π π π¦ corresponding to the iterated application of the map Β§ The initial point provides the initial conditions A point π¦ β , such that π π¦ β = π¦ β is a fixed point, the orbits remain Β§ in π¦ β for all future iterations Cobweb plots for individual orbits 6
F IXED POINTS Β§ Starts at 1.6, converges to 1 Β§ Starts at 1.8, converges to -1 Stability of a fixed point π¦ β ? Fixed points correspond to intersection between graph π(π¦) and π¦ Β§ General map: π¦ KR3 = π(π¦ K ) , π(π¦ β ) = π¦ β Β§ Letβs consider a near orbit, π¦ K = π¦ β + π K : is the orbit attracted or Β§ repelled from π¦ β ? If itβs attracted we can say that π¦ β is stable Does the perturbation π K grow or decay with π ? Β§ By the definition, π¦ β + π KR3 = π¦ KR3 = π(π¦ β + π K ) , and using the Taylor Β§ series expansion about π¦ β : π¦ β + π KR3 = π¦ KR3 = π π¦ β + π K = π π¦ β + π T π¦ β π K + π(π K2 ) Given that π(π¦ β ) = π¦ β Γ π KR3 = π T π¦ β π K + π(π K2 ) Β§ Β§ If we take the linear approximation Γ Linearized map: π KR3 = π T π¦ β π K Β§ Eigenvalue / multiplier: π = π T π¦ β Β§ 7
F IXED POINTS Linearized map: π KR3 = π T π¦ β π K Β§ Solution of linearized map: π 3 = ππ < , π 2 = ππ 3 = π 2 π < β¦.. π K = π K π < Β§ If π = |π T π¦ β | < 1 Γ π K β 0, for π β β, and π¦ β is linearly stable Β§ If π = |π T π¦ β | > 1 Γ π K β β, for π β β, and π¦ β is unstable Β§ The linear stability holds also for the general map Β§ The marginal case π = |π T π¦ β | = 1 doesnβt allow to draw conclusions. Β§ In this case the quadratic term π(π K2 ) determines the stability If π =0, then the fixed point is said superstable Β§ π¦ KR3 = sinπ¦ K Β§ π¦ β = 0 is a fixed point Β§ π = π T π¦ β = 1 , marginal case Β§ Β§ Cobweb Γ Itβs stable! Is it global? For all orbits π¦ K β 0 ? Β§ For any π¦ < , π¦ 3 β [β1,1] since |sinπ¦ 3 | < 1 Β§ 1 Β§ Γ From cobweb we can say itβs global 8
A NOTHER EXAMPLE , LIMITING BEHAVIOR π¦ KR3 = cos π¦ K Β§ Kβl π¦ K ? β¦ by iterating the map (e.g., use calculator!), π¦ K β 0.739.. lim Β§ Solution of trascendental equation π¦ = cosπ¦ Β§ The fixed point 0.739β¦ has π < 0 Γ Damped oscillations Β§ For 0 < π < 1 convergence to a stable fixed point is monotonic Β§ 9
L OGISTIC MAP π¦ KR3 = π π¦ K (1 β π¦ K ) Β§ π¦ K is a dimensionless measure of the population in the π th generation Β§ and π is the intrinsic growth rate (with capacity being limited to 1) Letβs restrict 0 β€ π β€ 4 Β§ Γ The map maps 0,1 β 0,1 Letβs fix π and study the evolution Β§ Β§ Trivially, for small growth rates, π < 1, the population always goes 1 π¦ K extinct, as π¦ K β 0 2 For 1 < π < 3 , population grows Β§ and eventually reaches a non-zero steady state Watch out: this a time series! 10
A PATH TO THE CHAOS β¦ 11
βR EGULAR β BEHAVIOR , PERIODIC ATTRACTORS 12
βR EGULAR β BEHAVIOR , PERIODIC ATTRACTORS 13
T RANSITION TO CHAOTIC BEHAVIOR 14
C HAOS : S ENSITIVITY TO INITIAL CONDITIONS 15
P ERIODS IN THE LOGISTIC MAP Β§ Oscillating about the previous steady state, alternating between small and large populations Period-4 cycle Β§ Β§ Period-2 cycle: Oscillation repeats every two iterations, periodic orbit Period-doubling to cycles appears by increasing π Β§ Β§ They correspond to bifurcations i n phase diagram Β§ Successive bifurcations come faster and faster! Limiting value π K β π l = 3.569946 β¦ Β§ Β§ Geometric convergence, in the limit the distance between successive values shrink to a constant: 16
C HAOS β¦ Β§ π > π l ? Β§ For many values of π , the sequence never settles down to a fixed point or a periodic orbit Β§ Aperiodic, bounded behavior! 17
O RBIT DIAGRAM What happens for larger π ? Sure, more chaos β¦. Even more interesting things! Β§ Orbit diagram: systemβs attractors as a function of π Β§ Β§ Construction: Choose a value of π Β§ Select a random initial condition π¦ < and generate the orbit: lets iterate for ~ 300 cycles to let Β§ the system settle down, then plot the next ~ 300 points from the map iterations Move to an adjacent value of π and repeat, sweeping the π interval Β§ At π β π l = 3.57 the map Β§ becomes chaotic and the attractor changes from a finite to an infinite set of points For π > 3.57 , mixture of order Β§ and chaos, with periodic windows interspersed between clouds of (chaotic) dots 18
O RBIT DIAGRAM 19
C HAOS AND ORDER 20
C HAOS AND ORDER , S ELF - SIMILARITY The large window at π β 3.83 Β§ contains a stable period-3 orbit Β§ Looking at the period-3 window even closer: a copy of the orbit diagram reappears in miniature! Γ Self-similarity 21
Recommend
More recommend