Passing parameters in boundary value problem using BVP4C

For a system of boundary value problems I need to pass a parameter as a coefficient. Let's assume that I have defined the problem as:

solinit = bvpinit(linspace(0,10,50),[0 10]);
    sol = bvp4c(@twoode,@twobc,solinit,a);
    x = linspace(0,4);
    y = deval(sol,x);
   
    function dydx = twoode(x,y,a)
        dydx = [ y(2); -a*exp(-x)+x*exp(-x)];
    end

    function res = twobc(ya,yb,a)
        res = [ya(1); yb(1)-10*exp(-10)];
    end

Where a is the parameter, and my first try was to add a to all the functions involved, but I receive the "Not enough input arguments" error. Is there a way to include the parameter a in the problem?

Answer 1

[Star Strider]

Define ‘a’ in your workspace, then use this bvp4c call:

sol = bvp4c(@(x,y)twoode(x,y,a),@(ya,yb)twobc(ya,yb,a),solinit);

You are passing the extra parameter correctly, however bvp4c does not need to know about it, so only show bvp4c the arguments it wants, thus:

@(x,y)twoode(x,y,a)

and:

@(ya,yb)twobc(ya,yb,a)

The anonymous function construction (that you wrote correctly) will pick up ‘a’ from your workspace and pass it to your functions.

With this change (and specifying ‘a’), your code ran without error with this bvp4c call when I tested it.

[Saeid]

Thanks, like always, your suggestions were a great help.

[Star Strider]

As always, my pleasure!

I very much appreciate your compliment!

[Christian Chamberlayne]

This is super helpful. I simularly was stuck with the same issue. Thank you!

Answer 2

[Stephan]

See here: passing extra parameters. I suggest to either use the anonymous functions or the nested functions approach. Both is easy to apply.

[Saeid]

Thanks Stephan! I just did and it worked!