
Title: Directional Field

Series: Ordinary Differential Equations

YouTubeTitle: Ordinary Differential Equations  Part 3  Directional Field

Bright video: https://youtu.be/Gs125mL8O0

Dark video: https://youtu.be/_ZzEUQI3I28

Quiz: Test your knowledge

PrintPDF: Download printable PDF version

Thumbnail (bright): Download PNG

Timestamps

Subtitle in English
1 00:00:00,571 –> 00:00:06,006 Hello and welcome back to the video series about ODEs
2 00:00:06,314 –> 00:00:10,578 and in today’s part 3 we will talk about the directional field.
3 00:00:10,778 –> 00:00:16,886 This one will help us here at the beginning of this topic to visualize ODEs.
4 00:00:17,257 –> 00:00:23,204 Indeed it’s a vector field that can be sketched such that we understand the problem.
5 00:00:23,404 –> 00:00:26,228 Ok, but before we start with that, you already know,
6 00:00:26,428 –> 00:00:31,982 first I want to thank all the nice people who support this channel on Steady via Paypal or by other means.
7 00:00:32,414 –> 00:00:39,625 In fact you should know only because of your support it’s possible to create such videos about mathematics
8 00:00:39,825 –> 00:00:45,723 and as a thank you and as a bonus you can download the PDF version and the quiz for this video.
9 00:00:46,300 –> 00:00:51,066 Ok, then let’s start by telling again what we mean by an ODE.
10 00:00:51,266 –> 00:01:00,245 So you might remember if we write x dot (the first derivative of x) is equal to a function w, then we have an ODE.
11 00:01:00,614 –> 00:01:06,888 Moreover this function w has the independent variable t as an input and x itself.
12 00:01:07,357 –> 00:01:15,265 In addition you should know that x can be a vector. So we can have a whole system of ODEs.
13 00:01:15,465 –> 00:01:24,051 However to be more precise you have already learned that an ODE of this form is called explicit and of first order.
14 00:01:24,600 –> 00:01:32,428 This means that we only have the first derivative in the whole equation and the first derivative is also alone on the lefthand side.
15 00:01:32,743 –> 00:01:39,747 However, I can already tell you that in the next video we will show that this is not really a restriction.
16 00:01:40,243 –> 00:01:45,924 Indeed, we will see that we can simplify that even more without losing generality
17 00:01:46,124 –> 00:01:50,268 and in order to see that let’s first look at some examples here
18 00:01:50,829 –> 00:01:56,437 and the first one here should be the ODE x dot is equal to lambda times x.
19 00:01:56,743 –> 00:02:03,945 So a very common example, but what we should see here is that the function w doesn’t depend on t.
20 00:02:04,145 –> 00:02:09,769 Hence the independent variable, the time variable, does not occur in the equation.
21 00:02:09,969 –> 00:02:15,787 Therefore we introduce a new name for such ODEs. They are called autonomous.
22 00:02:16,286 –> 00:02:21,619 On the contrary we can immediately think of an example which is not autonomous.
23 00:02:22,229 –> 00:02:25,415 Namely x dot is equal to t.
24 00:02:25,700 –> 00:02:30,561 There the first variable t will change the value of the function w.
25 00:02:31,200 –> 00:02:35,983 Hence this one is an ODE, but not an autonomous ODE
26 00:02:36,329 –> 00:02:40,871 and of course this notion can also be used for a system of ODEs.
27 00:02:40,986 –> 00:02:43,525 For example in 2dimensions.
28 00:02:44,100 –> 00:02:48,708 So here x dot is equal to (x_2, x_1).
29 00:02:49,300 –> 00:02:56,415 So you see, there is no t on the righthand side here. Hence the whole system is also called autonomous.
30 00:02:56,971 –> 00:03:03,744 Indeed, exactly these systems will be the interesting ones so let’s consider them for the rest of the video
31 00:03:04,214 –> 00:03:08,005 and of course we have to put this into a new definition.
32 00:03:08,386 –> 00:03:12,942 In fact we want to keep it short and just call it an autonomous system.
33 00:03:13,614 –> 00:03:20,457 So it means we have a system of ODEs and the function w on the righthand side does not depend on t.
34 00:03:20,786 –> 00:03:25,644 Therefore we can just write it as a function w only depending on x.
35 00:03:26,129 –> 00:03:34,945 Moreover you know we have a whole system, so this can be a vector equation. Which means the function v maps R^n to R^n.
36 00:03:35,471 –> 00:03:41,024 However in general you know v can have a domain so it maps U into R^n,
37 00:03:41,486 –> 00:03:45,864 but of course U should then be a subset of R^n as well
38 00:03:46,357 –> 00:03:50,556 and indeed often we will choose U as an open set.
39 00:03:50,756 –> 00:03:58,716 Moreover I would also say that all the times you see that v is at least a continuous function defined on U.
40 00:03:59,100 –> 00:04:04,640 Ok, but now the point is, such an autonomous system can be nicely visualized
41 00:04:04,840 –> 00:04:08,629 and this is done by the so called directional field.
42 00:04:09,086 –> 00:04:15,331 I would say this is easy to understand, because it’s simply a plot of our function v.
43 00:04:15,614 –> 00:04:20,870 So you could say this is a vector field which visualizes the function v.
44 00:04:21,070 –> 00:04:26,232 Therefore you would start with the domain U. Which is a subset of R^n.
45 00:04:26,432 –> 00:04:30,333 So for example it could be the whole R^2.
46 00:04:30,400 –> 00:04:36,267 Hence here let’s visualize an example where v maps R^2 into R^2.
47 00:04:36,886 –> 00:04:44,937 So if you choose any point p in this coordinate system, the resulting value of v will also be a vector
48 00:04:45,300 –> 00:04:50,911 and this vector we can simply visualize as an arrow starting at p.
49 00:04:51,111 –> 00:04:55,552 Now, this is a common way to explain a vector function like v
50 00:04:55,752 –> 00:05:02,331 and now of course the important fact here is that you could do it at any other point in this coordinate system
51 00:05:03,114 –> 00:05:07,052 and then you simply have a picture with a lot of arrows.
52 00:05:07,800 –> 00:05:14,761 However we now call it the directional field of the ODE, because it explains what solutions have to do.
53 00:05:14,961 –> 00:05:24,790 Namely the ODE, the whole system of ODEs says that at a given point this vector is exactly the derivative of the solution.
54 00:05:25,400 –> 00:05:31,179 Indeed this fits nicely in with the pictures of the orbits we had in the last video.
55 00:05:31,379 –> 00:05:36,961 Please note, an orbit would be the image of a solution here in our domain.
56 00:05:37,557 –> 00:05:47,257 However, now we know this can only be a solution of our system if at each point the vector we see is a tangent for this curve.
57 00:05:47,900 –> 00:05:54,849 Simply because at each point the vectors here tell us in which direction the solution should flow.
58 00:05:55,049 –> 00:05:59,671 In other words if we plot the directional field of this vector function v,
59 00:05:59,757 –> 00:06:04,171 we can already see what the solutions can do and what they can’t do.
60 00:06:04,714 –> 00:06:12,508 Therefore this whole thing here is already very helpful to see the behaviour of possible solutions of the ODE,
61 00:06:12,708 –> 00:06:17,075 but of course this whole procedure here, we should explain with examples
62 00:06:17,657 –> 00:06:21,354 and first let’s start with a 1dimensional example.
63 00:06:21,554 –> 00:06:25,163 So x dot should be equal to the sine of x.
64 00:06:25,457 –> 00:06:31,305 So not too complicated, but we also realize that we don’t see all the solutions immediately.
65 00:06:31,714 –> 00:06:37,521 However we can recognize that v is defined on the whole real number line
66 00:06:37,721 –> 00:06:42,904 and this means that our directional field is now just a 1dimensional picture.
67 00:06:43,471 –> 00:06:48,944 Hence for the arrows we want to draw in there are only 2 possible directions.
68 00:06:49,614 –> 00:06:56,343 To see what happens here let’s look at some important points like 0, pi/2, pi and so on.
69 00:06:56,543 –> 00:07:01,709 Of course we choose exactly these numbers, because of the sine function.
70 00:07:01,909 –> 00:07:08,542 So for example if we put pi/2 into our function v, we get out 1.
71 00:07:08,742 –> 00:07:13,757 In other words we have an arrow of length 1 that points to the right.
72 00:07:14,329 –> 00:07:24,341 On the other hand if we choose 3/2 times pi and put that into our function v, we get out 1. So an arrow that points to the left.
73 00:07:24,929 –> 00:07:31,508 So you see this is what we could do for all points on the real number line and then we get the whole directional field.
74 00:07:32,171 –> 00:07:40,891 So as another example if you choose a small, positive number here on the right of 0, we also get an arrow that points to the right.
75 00:07:41,543 –> 00:07:48,081 However please note at the point 0 and the point pi, we simply get out 0.
76 00:07:48,281 –> 00:07:53,046 This means at these points, there is no arrow pointing anywhere,
77 00:07:53,486 –> 00:07:59,029 but exactly that is immediately an important information for the solutions.
78 00:07:59,586 –> 00:08:05,100 So for example you could take a solution alpha(t) that is always 0.
79 00:08:05,600 –> 00:08:09,996 So it means alpha(t) vanishes no matter which t you put in
80 00:08:10,196 –> 00:08:14,151 and now we can simply check that this is indeed a solution.
81 00:08:14,351 –> 00:08:19,228 This is not hard to do. You simply put alpha into the ODE.
82 00:08:19,428 –> 00:08:24,718 So we want that alphadot(t) is equal to sin(alpha(t)).
83 00:08:25,857 –> 00:08:35,824 So this is what a solution has to fulfill for all t and here we don’t have a problem, because the derivative is 0 and sin(0) is also 0.
84 00:08:36,571 –> 00:08:40,596 So here we simply have 0 = 0.
85 00:08:41,100 –> 00:08:48,927 Now, this is important to remember. This always works if you find a point in the directional field that has no arrow at all
86 00:08:49,400 –> 00:08:54,006 and therefore we can immediately give a second solution alpha(t).
87 00:08:54,486 –> 00:08:58,651 We simply choose it as the constant function will value pi.
88 00:08:58,851 –> 00:09:06,799 Indeed, with the same reasoning this is also a solution, because the derivative is 0 and sin(pi) is 0
89 00:09:07,314 –> 00:09:11,591 and the important fact is that this claim holds for all t in R.
90 00:09:12,429 –> 00:09:19,130 In other words you see here that we have solutions that stay with their orbit just on one point
91 00:09:19,657 –> 00:09:24,467 and for this reason these points in the picture are called stationary points.
92 00:09:25,200 –> 00:09:30,406 However the directional field does not only tell us about the constant solutions here.
93 00:09:31,214 –> 00:09:37,499 Moreover you should see that we also get information about the solution that lies in here.
94 00:09:37,699 –> 00:09:42,717 In particular we can look at a solution that hits the value pi/2.
95 00:09:42,917 –> 00:09:51,686 There we know if we now increase t, we have to go to the right with the solution, because this is the direction of the arrow.
96 00:09:52,057 –> 00:09:57,362 Indeed, this is the direction of all the arrows in the interval 0 to pi.
97 00:09:58,057 –> 00:10:05,816 Therefore we can directly conclude that a solution inside this interval has to be monotonically increasing.
98 00:10:06,186 –> 00:10:14,054 Moreover we also see that we can’t exceed the value pi in the end, because pi is a stationary point.
99 00:10:14,829 –> 00:10:20,244 This means if we go with t to infinity, we will approach this value pi.
100 00:10:20,514 –> 00:10:27,326 So this is very nice. Without knowing the explicit solution, we can already say something about its behaviour.
101 00:10:28,043 –> 00:10:30,920 So this is the big advantage of the directional field.
102 00:10:31,120 –> 00:10:34,483 It helps you immediately with understanding solutions.
103 00:10:34,683 –> 00:10:40,736 However maybe the whole thing is more interesting if we look at a higher dimensional example.
104 00:10:41,586 –> 00:10:46,883 So I would say let’s take a 2dimensional one and a similar one to the last video.
105 00:10:47,486 –> 00:10:53,306 So x_1 dot should be x_2 and x_2 dot should be x_1.
106 00:10:53,829 –> 00:10:58,395 So here our function v goes from R^2 into R^2.
107 00:10:58,929 –> 00:11:04,103 Hence there the directional field can be visualized in a 2dimensional plane.
108 00:11:04,757 –> 00:11:10,634 Each point in the plane get’s mapped to a vector which is given by (x_2, x_1).
109 00:11:11,186 –> 00:11:16,247 Therefore let’s draw the plane and let’s visualize this directional field.
110 00:11:16,814 –> 00:11:24,203 Indeed, we see that this whole picture is not so complicated, because essentially it’s just flipping the arrows.
111 00:11:24,957 –> 00:11:30,686 Therefore the arrows also get longer if we get further away from the origin.
112 00:11:31,300 –> 00:11:37,478 Of course now we could fill in the whole picture and what you should see is such a rotational field.
113 00:11:38,014 –> 00:11:42,129 Now, if you don’t want to draw it by hand, you can also use a computer.
114 00:11:42,178 –> 00:11:45,753 For example Python will make this nice picture.
115 00:11:46,529 –> 00:11:52,354 There please note, I have shortened the arrows such that you can see this nice flow here.
116 00:11:52,886 –> 00:12:00,121 In other words, if we have a solution, it will go around the origin. Namely in the positive direction.
117 00:12:00,643 –> 00:12:05,641 That simply because we now know, a solution has to follow the arrows.
118 00:12:06,243 –> 00:12:11,285 Hence in our example here, the orbits will look like circles.
119 00:12:12,029 –> 00:12:18,899 So the result is the same as before. We can already sketch solutions with the help of the directional fields.
120 00:12:19,357 –> 00:12:25,665 Moreover we also see that we have a stationary solution, a stationary point in the picture.
121 00:12:26,286 –> 00:12:29,902 Indeed, there is only one and it is the origin.
122 00:12:30,729 –> 00:12:36,967 Ok, so in summary you see, the directional field helps us a lot to visualize solutions.
123 00:12:37,486 –> 00:12:44,320 However some questions remain. What does it tell us about the existence and uniqueness of solutions?
124 00:12:45,071 –> 00:12:49,447 And I tell you, these questions we will answer with the next videos.
125 00:12:49,971 –> 00:12:54,200 So I really hope we meet again and have a nice day. Bye!