matlab - Why is isequaln(f, simplify(f)) false? -

I have a function f as defined in this form:

  Symfun symfun (c + t - (2 * (t - 2 ^ (1/2) * (a * t) ^ (1/2)) * (a - t / 2)) / ( 2 * A, however, check the similarity with the simplified version yield:  
  isequaln (f, simplify (f)) ans = 0  
 
 Is not expected to return the exact equivalent 
 

uneven >>

Probably use to test There is no suitable function for equality. sym / isequaln :

For help with montage (A, B), iff A and B are equally true of NaNs

As the guideline has been given to @ benoit1.1, the documentation is not completely clear, but, as it is numeric equivalent, the function ("the same" Despite the unfortunate use of the term) seems to be a test for simplified functions There is an identity for the unassumed function, but it is not an analogy, according to the documentation, can keep in mind the sym / isequaln account and possibly simple relations, but it is a change internally and It does not simplify, it is almost certainly because simplification is a computationally expensive process, so it's probably some well-defined (but well documented) tests Was a design decision to join.

Disadvantages

So, how can you move forward? In addition to sym / isequaln , parallels and inequalities are tested in a more consistent way in the symbolic math toolbox:

  syms catf = symfun (c + t - (2 * (T-2 ^ (1/2) * (one * t) ^ (1/2)) * (a - t / 2)) / (2 * a - t), [cat]) g = Simplified However, in R2014b it returns  false  and the following warning: 
 
  Warning: can not be proved 'c + t - ((2 * t - 2 * 2 ^ (1/2) * (a * t) ^ (1/2)) * (a - t / 2)) / (2 * a - t) == c + 2 ^ (1/2) * (A * t) ^ (1/2) '& gt; / Apps / MTB 2014 b.app/toolbox/symbolic/symbolic/symengine.p>-symengine 56 on sym.is always at 38  
 
 You should be aware that the symbolic variable is considered to be complex by default And  is always   c ,  a  try to prove this connection to  all possible combinations  Is doing  T . Simplification is also a malicious process - From the documentation: 
 
 
 The simplicity of mathematical expression is not a clearly defined subject, there is no universal idea because any form of expression is simple. The form of a mathematical expression which is the simplest of a problem can be complex or may be unsuitable for another problem. 
 
 
 If you look closely at your basic equation, you will see that a uniqueity has been removed in R2014b in  g (1,1,2)  Returns the value of  3 , but evaluates  f (1,1,2) : 
 
  in the muppad command Error: Error using Mupdammex: Zero Division [_power] Evaluation: _symans_32_15992 ...  Use of assumptions   
 If you know something about your symbolic variable, then avoid this by applying the assumptions , For example  
  symfun symfun (c + t - (2 * (t - 2 ^ (1/2) * (a * t) ^ (1/2)) * (A - T / 2)) / (2 * a - t), [cat]) g = simplify (f) value (t & 2; a2) elawe (f == g)  
 
 Which now returns  true , by using the example of Benoit 1,  sys ab; Always returns directly  true  ((a + b) ^ 2 = a ^ 2 + 2 * a * b + b ^ 2) .