Create a function of n variables

Hello,

I have a few questions concerning the same function. I am trying to recreate the following function:

For a specific n and a=1, I have the following code:

n = 5 ;
syms b ;
x = sym('x', [1 n]) ;
f = sum(x)^(1/b)

I was wondering:

1. Is there a way to create this function without specifying n? That is (without the powers to simplify the question):
2. In my above code, I assumed a=1. If I try to include any power to x, I get an error.

So far I tried:

n = 5 ;
syms b ;
syms a ;
x = sym('x', [1 n]) ;
f=sum(x^a)^(1/b)

& (using a = 2 as a test)

n = 5 ;
syms b ;
x = sym('x', [1 n]) ;
f=sum(x^2)^(1/b)

Both resulted in the following error:

Error using symengine
Not a square matrix.

Error in sym/privBinaryOp (line 1002)
            Csym = mupadmex(op,args{1}.s, args{2}.s, varargin{:});

Error in  ^  (line 325)
        B = privBinaryOp(A, p, 'symobj::mpower');

It's worth mentioning I'm entirely new to Matlab, so I'm still trying to make sense of everything.

Thanks in advance,

Answer 1

[madhan ravi]

.^ instead of ^ % lookup element wise operation

[Michaël de Clercq]

Thank you! I saw .* for multiplication but since I didn't look into it yet, I suppose my brain didn't think about trying .^. Would you happen to know the answer to my first question as well? I'm interested in the general behaviour of the function. For example a=b < 1 results in concavity. Although n would be irrelevant for this, I'd like to be able to prove this in Matlab without having to specify n.

[madhan ravi]

perhaps you want to use symsum()

[Michaël de Clercq]

I looked into it using symsum() first, but that way I couldn't even recreate the function for a specified n, let alone use it to evaluate for an unspecified n. Looking into it again atm.