Trapezoidal method

[WSUPhys]

Hello,
I am working on a program to find the value of the current in a coil. This value satisfies the following equation:
y'=sin(2t)-[(e[sup]y[/sup]-1)/(e[sup]y[/sup]+1)]

which is of the form y'=f(t,y)

I know that in order to solve this I need to use the trapezoidal method to solve a differential equation, the formula is:
y[sub]n+1[/sub]=y[sub]n[/sub]+.5*h(f(t[sub]n[/sub],y[sub]n[/sub])+f(t[sub]n+1[/sub],y[sub]n+1[/sub]) where h=t[sub]n+1[/sub]-t[sub]n[/sub]

I have found examples of the standard trapezoidal method but I do not think they will help because of the difference in the formulas. Can anybody give me some guidance here?

 

[xwb]

Do you need guidance in the math or the programming?

 

[CJDeeds]

the programming, I know how the math behaves but I am unfamiliar as far as how to write this in C++ code.

 

[xwb]

I am totally ignorant as to what is required. Do you need a load of y values given a set of t values or the y and t value when they satisfy some condition like y' is zero. I also don't really know how you would solve something like

yn+1=yn+.5*h(f(tn,yn)+f(tn+1,yn+1)

because yn+1 appears on both sides. Assuming h is really small, why not use the Euler form of

y(n+1) = yn + h * f(tn, yn)

In that case, the same question arises: do you want all the values given a starting t0 and y0 up to a particular value of tn or do you just want it satisfying some criteria.

 

[mikrom]

My answer is here

http://www.tek-tips.com/viewthread.cfm?qid=1710515