HW1_Solutions

pdf

School

University of California, Los Angeles *

*We aren’t endorsed by this school

Course

30B

Subject

Aerospace Engineering

Date

Apr 3, 2024

Type

pdf

Pages

17

Uploaded by JudgeRiver13513

Report
Problem 4.1. Further Exercise 1 Exercise 4.1.FE 1 Is a saddle point an attractor? Justify your answer. For something to be an attractor, it has to attract other trajectories in space. However, for a saddle point, it generally does not. The trajectories can approach the point from an angle, and then almost always run away in the long run. Thus, a saddle point is NOT an attractor. Problem 4.1. Further Exercise 2 Exercise 4.1.FE 2 Does a trajectory that approaches a limit cycle attractor ever reach the attractor? Explain. So let’s say that the trajectory will actually reach the attractor. Let’s think about this... 1. If the trajectory reaches the attractor, that means that the attractor is actually a part of the trajectory. 2. However, we know that all trajectories will approach the attractor. So if #1 is true, it means that the attractor will belong to ALL trajectories. 3. If #2 is true, then the trajectories will cross each other at least at the region of the limit cycle attractor. 4. However, we know that trajectories can NEVER cross. As a result, the limit cycle attractor can never belong to any trajectory. In other words, A trajectory that approaches a limit cycle attractor will NEVER reach the attractor. Problem 4.1. Further Exercise 4 Exercise 4.1.FE 4 Describe jet lag and recovery from it in dynamical terms. The body controls sleep cycle via the circadian rhythm, which oscillates throughout the day. After a long flight, the circadian rhythm witnesses a drastic switch, where the levels of biomarkers do not math with the (expected) normal levels at certain hours. However, the body is not able to adapt immediately since it takes time to control levels of biomarkers in the entire body. Rather, the system slowly adjusts and approaches the stable limit cycle where the circadian rhythm is normally at. 1
Problem 4.1. Further Exercise 6 Exercise 4.1.FE 6 Suppose a 2D system has a stable equilibrium point that is located somewhere outside a limit cycle. Can a trajectory starting inside the limit cycle reach this point? Justify your answer. (Hint: It may help to draw the situation.) As we can see, if a trajectory starts inside the limit cycle, it will try to approach the limit cycle but can never reach it (see Further Exercise 2 ). However, if it can never reach the cycle, it can never cross it to reach the stable equilibrium point on the other side of the cycle. Thus, A trajectory starting inside the limit cycle CANNOT reach this point. Problem 4.2.3 Exercise 4.2.3 Verify this assertion. We are looking to verify the assertion: “This negative feedback model will not oscillate, no matter how steep the feedback.” . To do this, we will need to look at the trajectories and/or time series for the system when it has: (1) a low value of n (not-steep feedback), and (2) a high value of n (steep feedback). There are a few different ways you can accomplish this: animation, interactive, assign-directly, etc. We will show the codes for the most simple method, which is to assign a definite value to n . Recommended Initial Condition: H = 0 . 25 and G = 2 2
1 var( ‘H, G’ ) # define H and G as symbolic CoCalc variables 2 n = 5 # assign a value to n 3 time list = srange( 0 , 100 , 0.1 ) # define the list of time points to obtain solutions 4 5 Hprime(H, G) = 1 / ( 1 + Gˆn ) - 0.2 * H # define H’ 6 Gprime(H, G) = H - 0.2 * G # define G’ 7 8 sol = desolve odeint( [ Hprime(H, G), Gprime(H, G)], dvars = [H, G], ics = [ 0.25 , 2 ], times = time list) # simulate the model 9 10 list plot( sol ) # plot the trajectory Small n Value (e.g. n = 5 ) Large n Value (e.g. n = 50 ) As observed, regardless of the steepness of the feedback (i.e. how large the value of n is), the system always has one stable equilibrium point, instead of a stable oscillation/limit cycle attractor. Problem 4.2.4 Exercise 4.2.4 In the delay differential equation Y ( t ) = 16 Y ( t 2) + 8 Y ( t ), what does Y ( t 2) refer to? What does Y ( t ) refer to? We have: Y ( t ) the value of Y at time t (i.e. right now) Y ( t 2) the value of Y at time t 2 (i.e. two time units prior ) Problem 4.2.5 3
Your preview ends here
Eager to read complete document? Join bartleby learn and gain access to the full version
  • Access to all documents
  • Unlimited textbook solutions
  • 24/7 expert homework help
Exercise 4.2.5 What aspect of the function does V max control? Let’s graph V max · X n 1+ X n for a few different values of V max . We also pick: n = 5. X Ventilation Rate V max = 1 V max = 2 V max = 3 As we observed, the V max controls the maximum value for the Ventilation Rate, and graphically, it controls the height of the function. Problem 4.2.6 Exercise 4.2.6 Let X = 6 16 · X ( t 0 . 2) 5 1 + X ( t 0 . 5) 5 · X Assume that for all t 0, X ( t ) = 0 . 5. Use Euler’s method with a step size of 0.1 to approximate X (0 . 3). A Note on Our Notation: X ( t = 3) ” vs “ X 3 In this case, the first notation means the value of X at t = 3. Meanwhile, the second notation means the value of X on its third iteration . This may or may not mean at t = 3. If the step size is ∆ t = 1, then by the third step, you will reach t = 3, hence X ( t = 3) = X 3 for ∆ t = 1. However, let’s say that ∆ t = 0 . 01, then by the third step, you only reach t = 0 . 03, and clearly X 3 = X ( t = 0 . 03) ̸ = X ( t = 3) for ∆ t ̸ = 1. To reach t = 0 . 3 in step size of ∆ t = 0 . 1, we need to take 3 steps . We also have the following equations: X ( t ) = 6 16 · X ( t 0 . 2) 5 1 + X ( t 0 . 5) 5 · X ( t ) (1) X t new = X t old + X ( t old ) · t (2) 4
Step #0: Find X at t = 0 As t = 0 0: X (0) = 0 . 5 Step #1: Find X at t = 0 . 1 We have t new = 0 . 1 and t old = 0. It is important to remember that we are evaluating X at t old (as seen in Equation (2)). X ( t 0 . 2) X ( t 0 . 5) X X old X (0 0 . 2) = X ( 0 . 2) = 0 . 5 X (0 0 . 5) = X ( 0 . 5) = 0 . 5 X old = 0 . 5 Inputting to Equations (1) and (2), we have: X (0 . 1) = X (0) + X (0) · 0 . 1 = 0 . 5 + " 6 16 · (0 . 5) 5 1 + (0 . 5) 5 · (0 . 5) # · (0 . 1) = 1 . 07576 5
Step #2: Find X at t = 0 . 2 We have t new = 0 . 2 and t old = 0 . 1. It is important to remember that we are evaluating X at t old (as seen in Equation (2)). X ( t 0 . 2) X ( t 0 . 5) X X old X (0 . 1 0 . 2) = X ( 0 . 1) = 0 . 5 X (0 . 1 0 . 5) = X ( 0 . 4) = 0 . 5 X old = 1 . 07576 Inputting to Equations (1) and (2), we have: X (0 . 2) = X (0 . 1) + X (0 . 1) · 0 . 1 = 1 . 07576 + " 6 16 · (0 . 5) 5 1 + (0 . 5) 5 · (1 . 07576) # · (0 . 1) = 1 . 62360 Step #3: Find X at t = 0 . 3 We have t new = 0 . 3 and t old = 0 . 2. It is important to remember that we are evaluating X at t old (as seen in Equation (2)). X ( t 0 . 2) X ( t 0 . 5) X X old X (0 . 2 0 . 2) = X (0) = 0 . 5 X (0 . 2 0 . 5) = X ( 0 . 3) = 0 . 5 X old = 1 . 62360 Inputting to Equations (1) and (2), we have: X (0 . 3) = X (0 . 2) + X (0 . 2) · 0 . 1 = 1 . 62360 + " 6 16 · (0 . 5) 5 1 + (0 . 5) 5 · (1 . 62360) # · (0 . 1) = 2 . 14488 In Summary, we have: Step Starting X n X ( t 0 . 2) X ( t 0 . 5) Compute X Final X n +1 0 - - - - X (0) = 0 . 5 1 X 0 = 0 . 5 0.5 0.5 X = 5 . 75758 X 1 = 1 . 07576 2 X 1 = 1 . 07576 0.5 0.5 X = 5 . 47842 X 2 = 1 . 62360 3 X 2 = 1 . 62360 0.5 0.5 X = 5 . 21280 X 3 = 2 . 14488 Problem 4.2. Further Exercise 1 Exercise 4.2.FE 1 Some people have difficulty maintaining a stable weight. Instead, they gain a lot of weight, go on a diet, lose the weight, but then gain it back. This pattern is sometimes referred to as yo-yo dieting. 6
Your preview ends here
Eager to read complete document? Join bartleby learn and gain access to the full version
  • Access to all documents
  • Unlimited textbook solutions
  • 24/7 expert homework help
a) What kind of feedback loop is involved in this situation? b) Use your understanding of feedback loops and oscillations to suggest what might help such a person to stabilize their weight. Part a: What kind of feedback loop is involved in this situation? In this case we have: Increase in weight decrease/change eating habits (go on a diet) Lost weight In other words, an increase in the value of a variable ultimately leads to its own decrease. Thus, this is a NEGATIVE FEEDBACK LOOP . Schematically, we have: Body Weight Eating Habits + Part b: Use your understanding of feedback loops and oscillations to suggest what might help such a person to stabilize their weight. The oscillations are caused by: 1. High sensitivity (steep reaction by the person): too much weight drastic diet too low weight eat a lot too much weight repeat 2. Time delay by the body because it takes time for body weight to change It is not as feasible to control the time delay since it is a physiological process. Thus, the person can only stabilize their weight by not changing their diet too drastically. Just adjust a little. Let the weight stabilize around a new value. If that is still not the desired weight, then make additional small change to the diet. Problem 4.2. Further Exercise 2 Exercise 4.2.FE 2 While traveling, you find yourself in a hotel room in which using the thermostat leads to large oscillations in the room’s temperature. The thermostat responds to the room’s air temperature by turning on an air conditioner on the other side of the room if 7
the temperature near the thermostat gets too warm. Similarly, when the temperature near the thermostat gets cold, the air conditioner switches off. What could the builder of the hotel have done to prevent the oscillations you are experiencing? Oscillations are caused by: (1) Steep reaction (high sensitivity), and (2) Time delay In this case, the reaction is steep (on/off) but it is not controllable (as it is a part of the mechanical design of the air conditioner to simply be on/off, instead of automatically adjusting how cold it wants the air to be). Thus, we can only control the time delay. To reduce drastic oscillations, we need to reduce time delay. In this case, it can be argued that the time delay is at maximum because the distance between the air conditioner and the thermostat is the greatest. Thus, the distance the air has to travel is also the greatest. To reduce time delay, they can put the thermostat closer to the AC (or AC closer to the thermostat, either way works). Problem 4.2. Further Exercise 4 Exercise 4.2.FE 4 Meerkats are highly social small carnivores that live in southern Africa. They rely on each other to raise their young. Use the following assumptions to model the number of adult meerkats, M , in a population. You can invent parameters as necessary. The per capita rate at which meerkats give birth to babies who survive to adulthood is a steep sigmoid function of the adult population, with higher reproductive success at higher populations. Meerkats die of natural causes at a constant per capita rate d . Meerkats are preyed upon by eagles and jackals. These predators have many other prey, so their population does not depend on the meerkat population. The rate at which jackals prey on meerkats is a nonsigmoid saturating function of the meerkat population. The rate at which eagles prey on meerkats is a sigmoid function of the meerkat population. The sigmoid is not very steep. Please note that for this problem, we will not add the signs to the contributions. We will add the signs when evaluating the inflow/outflow at the very end. Part a: The per capita rate at which meerkats give birth to babies who survive to adulthood is a steep sigmoid function of the adult population, with higher reproductive success at higher populations. For this one, it is up to interpretation, so this answer is acceptable: 8
M M = k · M n ( a 1 ) n + M n M = k · M n ( a 1 ) n + M n · M However, one can argue that the number of adults NOW is dependent on the number of babies born τ years back. And we know that the number of babies born THEN is dependent on the number of adults AT THAT TIME ( t τ ), not at this moment. As a result, the contribution from birth to M is: M M ( t τ ) = k · [ M ( t τ )] n ( a 1 ) n + [ M ( t τ )] n ⇐⇒ M = k · [ M ( t τ )] n ( a 1 ) n + [ M ( t τ )] n · M ( t τ ) (3) We will utilize this latter version of Equation (3). Part b: Meerkats die of natural causes at a constant per capita rate d . The contribution from death to M is: M M = d ⇐⇒ M = d · M (4) Part c: Meerkats are preyed upon by eagles and jackals. These predators have many other prey, so their population does not depend on the meerkat population. See Equations from Part d and e Part d: The rate at which jackals prey on meerkats is a nonsigmoid saturating function of the meerkat population. The contribution from Predation [From Jackals] to M is: M = c 1 · M a 2 + M · J (5) Part e: The rate at which eagles prey on meerkats is a sigmoid function of the meerkat population. The sigmoid is not very steep. The contribution from Predation [From Eagles] to M is: M = c 2 · M s ( a 3 ) s + M s · E (s is small) (6) 9
Your preview ends here
Eager to read complete document? Join bartleby learn and gain access to the full version
  • Access to all documents
  • Unlimited textbook solutions
  • 24/7 expert homework help
Putting Things Together: By combining Equations (3) - (6), we have: M = [Sum of Inputs] [Sum of Outputs] = [Birth] [Death + Predation from Jackals + Predation from Eagles] = [Birth] [Death] [Predation from Jackals] [Predation from Eagles] M ( t ) = k · [ M ( t τ )] n ( a 1 ) n + [ M ( t τ )] n · M ( t τ ) d · M ( t ) c 1 · M ( t ) a 2 + M ( t ) · J ( t ) c 2 · [ M ( t )] s ( a 3 ) s + [ M ( t )] s · E ( t ) for certain constants a 1 , a 2 , a 3 , c 1 , c 2 , n , s and τ Problem 4.2. Further Exercise 5 Exercise 4.2.FE 5 The garibaldi is a large orange fish that lives off the coast of California and Baja California. Use the assumptions below to write a differential equation for the size of an adult garibaldi population. The number of adults joining a population is the number of eggs laid times the fraction that hatch times the fraction that survive to adulthood. Garibaldis lay eggs at a constant per capita rate, b . Because garibaldis sometimes eat their own eggs, the fraction of eggs that hatch is a declining sigmoid function of the adult population. Larval garibaldis float as plankton before becoming adults and joining a population. Thus, the number of individuals joining a population is proportional to the number that hatched six years earlier, with proportionality constant r . Adult garibaldis die at a constant per capita rate d . Please note that for this problem, we will not add the signs to the contributions. We will add the signs when evaluating the inflow/outflow at the very end. Part 1: Contribution from Birth The number of adults joining a population is the number of eggs laid times the fraction that hatch times the fraction that survive to adulthood. Garibaldis lay eggs at a constant per capita rate, b . 10
Because garibaldis sometimes eat their own eggs, the fraction of eggs that hatch is a declining sigmoid function of the adult population. Larval garibaldis float as plankton before becoming adults and joining a population. Thus, the number of individuals joining a population is proportional to the number that hatched six years earlier, with proportionality constant r . We have the following diagram: Number of Eggs ( G 1 ) Hatched Eggs ( G 2 ) Survived to Adulthood ( G 3 ) Time: t 6 Time: t 6 Time: t We know that the number of “new adults joining a population” at time t is dependent on the number of eggs hatched six years earlier (as seen in Bullet #4), which in turn is dependent on the number of eggs produced AT THAT TIME. 11
Let’s start with the following statement: Bullet #1: The number of adults joining a population is the number of eggs laid times the fraction that hatch times the fraction that survive to adulthood. We have our first equation: G = G 3 = (# of eggs) · (Fraction Hatching) · (Fraction Survived) (7) Bullet #2: Garibaldis lay eggs at a constant per capita rate, b We saw from the diagram that the number of eggs is from the time t 6. In turn, the number of eggs laid is dependent on the population of Garibaldis at time t 6 (i.e. per-capita of the population at t 6). Thus, 1 G ( t 6) · G 1 = b ⇐⇒ G 1 = (# of Eggs) = b · G ( t 6) (8) Bullet #3: Because garibaldis sometimes eat their own eggs, the fraction of eggs that hatch is a declining sigmoid function of the adult population. We saw from the diagram that the hatching occurs at time t 6. Thus, (Fraction Hatching) = k a n + [ G ( t 6)] n (9) Note that this is NOT the same as G 2 because G 2 = (Fraction Hatching) · G 1 Note that the numerator does NOT have to be 1 because if G ( t 6) = 0 , it is only required that: k a n = 1 for some constants a , n , and k Bullet #4: Larval garibaldis float as plankton before becoming adults and joining a population. Thus, the number of individuals joining a population is proportional to the number that hatched six years earlier, with proportionality constant r . We have: (Fraction Survived) = r (10) Putting Things Together: Let’s plug in Equations (8) - (10) into Equation (7), we have the contribution from birth to G : G = ( b · G ( t 6) ) · ( k a n + G ( t 6) n ) · ( r ) (11) 12
Your preview ends here
Eager to read complete document? Join bartleby learn and gain access to the full version
  • Access to all documents
  • Unlimited textbook solutions
  • 24/7 expert homework help
Part 2: Contribution from Death We know that: Bullet #5: Adult garibaldis die at a constant per capita rate d . The contribution of death is in real time (i.e. at time t ), thus 1 G ( t ) · G = d ⇐⇒ G = d · G ( t ) (12) Part 3: Putting Things Together By combining Equations (11) and (12), we have: G = [Sum of Inputs] [Sum of Outputs] = [Birth] [Death] G ( t ) = ( b · G ( t 6) ) · ( k a n + G ( t 6) n ) · ( r ) d · G ( t ) for some constants b , a , k , n , r , and d Problem 4.3.1 Exercise 4.3.1 Explain what each term in this model means and why it has the algebraic form (for example, SP 2 ) that it does. We are looking at this model: S = V 0 cSP 2 P = cSP 2 kP We have: V 0 : the production rate of fructose-6-phosphate (from the two-step conversion of glucose not featured in reaction scheme) cSP 2 : the rate at which one molecule of fructose-6-phosphate is lost (and conversely, one molecule of ADP is formed) SP 2 : the probability that ONE molecule of fructose-6-phosphate and TWO molecules of ADP will meet each other AT THE SAME TIME ( P 2 because there are two molecules of ADP) c : the probability constant that tells us how likely one successful encounter between three molecules (1 fructose-6-phosphate + 2 ADPs) will result in a successful forward reaction 13
kP : the rate of degradation of ADP k : the per-capita rate of degradation of ADP Problem 4.3.2 Exercise 4.3.2 Why can h act as a half-saturation density? In other words, what is the consumption rate when N = h , and what does this mean biologically? At N = h , f ( N = h ) = C max · h h + h = C max · h 2 · h = C max 2 Thus, mathematically, we see that at N = h , the rate of consumption of the predator is at half of its maximum rate. Biologically, it can act as a switch between low and high stable values: Above h, an increase in N will has less effect in the rate of consumption. Thus, if N increases at a really fast rate (i.e. big ∆ N ), the predators’ consumption rate cannot keep up (i.e. small f ), thus the number of preys increases to a higher value, as shown in blue below. Below h, an increase in N will cause a big increase in the rate of consumption. Thus if N increases even slightly (i.e. small ∆ N ), the predators can eat a lot more (high increase in consumption rate, i.e. big ∆ f ), thus the number of preys decreases to a lower value, as shown in red below. N f h N f N f 0 14
Problem 4.3.3 Exercise 4.3.3 Find the equilibria for this model using the parameter values in Figure 4.34 . (Hint: Work with the second equation first.) We have the following equations: N = r 1 · N · 1 N k w · N d + N · P (13) P = r 2 · P · 1 j · P N (14) Let’s simplify Equation (13) a little more, by factoring out the N . Equation (13) becomes: N = r 1 · N · 1 N k w · N d + N · P = r 1 · N · 1 N k w d + N · N · P ⇐⇒ N = N · " r 1 · 1 N k w d + N · P # (15) We know that the equilibria of the system occur when N = 0 AND P = 0. Thus, we need to set BOTH Equations (14) and (15) to be equal to 0 (recall that Equation (15) is a variation of Equation (13)). We can follow the hint by solving for P using P = 0 first, and then plug the P value into the equation for N = 0 to find the value for N . Using Equation (14), we have: P = 0 ⇐⇒ r 2 · P · 1 j · P N = 0 ⇐⇒ P = 0 OR 1 j · P N = 0 ⇐⇒ P = 0 OR j · P N = 1 ⇐⇒ P = 0 OR j · P = N ⇐⇒ P = 0 OR P = 1 j · N As we can see, there are two values of P that will make P = 0. Now, let’s use each of them, and find the corresponding N value in the solution using Equation (15). It is also important to note that for P = 0, P must be defined. In other words, N ̸ = 0. Situation #1: P = 0 for P = 0 By plugging P = 0 into Equation (15), Equation (15) becomes: 15
Your preview ends here
Eager to read complete document? Join bartleby learn and gain access to the full version
  • Access to all documents
  • Unlimited textbook solutions
  • 24/7 expert homework help
N = N · " r 1 · 1 N k w d + N · 0 | {z } 0 # = N · r 1 · 1 N k (16) To find N = 0, we then set Equation (16) to be equal to 0. Thus, we have: N = 0 ⇐⇒ N · r 1 · 1 N k = 0 ⇐⇒ N = 0 OR 1 N k = 0 ⇐⇒ N = 0 OR N k = 1 ⇐⇒ N = 0 OR N = k Since N ̸ = 0, we have one equilibrium point: N = k and P = 0 Situation #2: P = 1 j · N for P = 0 By plugging P = 1 j · N into Equation (15), Equation (15) becomes: N = N · " r 1 · 1 N k w d + N · 1 j · N # (17) To find N = 0, we then set Equation (17) to be equal to 0. Thus, we have: N = 0 ⇐⇒ N · " r 1 · 1 N k w d + N · 1 j · N # = 0 ⇐⇒ N = 0 OR r 1 · 1 N k w d + N · 1 j · N = 0 Since we mentioned that N ̸ = 0 for P to be defined, we only will choose the latter case. Let’s solve this equation for N first: r 1 · 1 N k w d + N · 1 j · N = 0 ⇐⇒ r 1 · k N k w · N j · ( d + N ) = 0 ⇐⇒ r 1 · ( k N ) · j · ( d + N ) w · k · N k · j · ( d + N ) = 0 ⇐⇒ r 1 · ( k N ) · j · ( d + N ) w · k · N = 0 16
⇐⇒ r 1 · j · k · d + k · N d · N N 2 w · k · N = 0 ⇐⇒ ( r 1 · j ) · N 2 + ( r 1 · j · k r 1 · j · d w · k ) · N + r 1 · j · k · d = 0 (18) Using real values of r 1 = 1, r 2 = 0 . 1 = 1 10 , k = 7, d = 1, j = 1, and w = 0 . 3 = 3 10 , our Equation (18) becomes: ( 1 · 1) · N 2 + 1 · 1 · 7 1 · 1 · 1 3 10 · 7 · N + 1 · 1 · 7 · 1 = 0 ⇐⇒ − N 2 + 39 10 · N + 7 = 0 ⇐⇒ 10 · N 2 39 · N 70 = 0 This equation has two solutions: N = ( 39) ± p ( 39) 2 4 · (10) · ( 70) 2 · (10) = 39 ± 4321 20 ⇐⇒ N 5 . 2367 OR N ≈ − 1 . 3367 Since 1 . 3367 < 0, N = 1 . 3367 is not within our state space. As a result, N 5 . 2367. Substituting N 5 . 2367 back to equation P = 1 j · N , we have: P = 1 j · N = 1 1 · N = N 5 . 2367 Therefore, our equilibrium point will be: N 5 . 2367 and P 5 . 2367 (you can round down to whole values if you want) In summary, our equilibria are: N = k = 7 and P = 0 N 5 . 2367 and P 5 . 2367 Note: we don’t need to worry about the fraction w · N d + N because since d > 0 and N > 0 , d + N > 0 , i.e. no chance of being equal to 0 17