Error:undefined variable in trapezoidal integration

Ok this is driving me crazy because it seems so simple. All i need to do is integrate from t=0 and t=3. I have the following function and the comments are what I put into the command window. I need to integrate the ydot vector but once I put it in the function is says that ydot is not defined like what.

[Stephen23]

Original question by Yo mama retrieved from Google Cache:

"Error:undefined variable in trapezoidal integration"

Ok this is driving me crazy because it seems so simple. All i need to do is integrate from t=0 and t=3. I have the following function and the comments are what I put into the command window. I need to integrate the ydot vector but once I put it in the function is says that ydot is not defined like what.

[Rena Berman]

(Answers Dev) Restored edit

Answer 1

[Brian Iwatake]

If you "clear all", then ydot is not there anymore, right?

[Yo mama]

I am the dumbest human alive lol

[Brian Iwatake]

No, you're not. I have done worse. Thanks for introducing me to <f = @(x) g(x)> notation.

Now, I think you have a logic error. <f = @(x) ydot> does not make sense to me because you have defined ydot to be a constant row vector with no dependence on x, I think.

In your "trapezoidal" function, <F = Fun(x)> will always return "ydot", no matter what "x" is, if I am interpreting your code correctly. Please post what you have figured out.

[Yo mama]

so the ydot vector I got was the result after evaluating 2 ODE's using 4th Order Runge Rutta method. Hence it I needed to plot a distance line and a velocity line. So I got all those y dot values from that. My x vector was technically time it was 0 secons to 3 seconds with step size 0.1, and I needed to evaluate what the distance was at 3 seconds. So I was basically integrating a vector with those terms of integration. Honestly, I was using a program we had in the book. I mean I think that makes sense.

[Yo mama]

Like It was a user defined function so like in the command window id put f=@(x)(ydot);

%disp('Value')

%I=trapezoidal(f,0,3,0.1)

[Brian Iwatake]

Hi, I'm still missing something. I didn't have your "ydot", so for my "ydot", I created a simple arbitrary ramp function, ydot = (0 : 2/3 : 100), which is just some big vector - I didn't want to run out of data points.

I = trapezoidal(f, 0, 3, 0.1) was your original function call.

try doubling the step size:

I = trapezoidal(f, 0, 3, 0.2)

For my "ydot" ramp function, the integration result is exactly half. This is not correct.

For your "ydot" data, do you get a reasonable answer for both trapezoidal(f, 0, 3, 0.1) and trapezoidal(f, 0, 3, 0.2)?

Does your result make sense to you?

As you make the step size "h" smaller, your integration result converges to the exact answer. For example, the integration result (usually) shouldn't double every time you cut "h" in half.

[Yo mama]

1. my step size was fixed for the question being asked, the result of my integration was compared to my runge-kutta result for the distance at 3 seconds and I got a percent error of less than one percent - so I think it was spot on, my code only worked when I ran a Sys2ODEsRK4 function as well in another window. My step size needed to be 0.1 explicitely.