Convert output from symbolic math (sym) to float

My question is similar to this question but I believe it to be more general.
I use Matlab's symbolic math toolbox to solve an equation:

MAZ = 0.5;
MAU = 1.0;
XI = 1.0;
ALPHA = 2.0;

DRG = 0.5;
SRG = 1.0;

PHI = 1 / (2 * MAU);

syms L;

f = 1 - DRG - sqrt(1 + (ALPHA * XI - L / (2 * XI * PHI) ) ^ 2 ) / ...
    sqrt(1 + (ALPHA * XI) ^ 2) + L / (4 * PHI * SRG * sqrt(1 + (ALPHA * XI)^2));

a = solve(f,L,'Real',true);

The answer is:

a =
5^(1/2)/3 + (10*((4*5^(1/2))/25 + 6/25)^(1/2))/3 + 8/3
5^(1/2)/3 - (10*((4*5^(1/2))/25 + 6/25)^(1/2))/3 + 8/3  

How do I automatically convert these expressions for a - which do not contain any symbolic expressions - to floats so that I can access them later in my code?

Matlab's symbolic math toolbox contains a function called double.
This function converts the result of the solve function - which is a sym - to a double:

double(a)

ans =

      5.98921078320145
     0.834834535131742

While double is correct, it is limited to the precision of 64bit floating points. Use vpa if a higher precision is needed:

>> vpa(a)

ans =

  5.9892107832014511063435699584181
 0.83483453513174202459587915406938

>> digits(99)

>> vpa(a)

ans =

  5.98921078320145110634356995841813862213621375395128614574627036653958858547362556353272837962692249
 0.834834535131742024595879154069378868157531819123064337100994463734092031618244369410214559292265698