Power always computes principal roots:
Powers are not generically inverses of roots:
With approximate numbers, imaginary parts can be generated:
Use
Chop to remove the small imaginary part:
The branch cut makes this function discontinuous:
Its derivative nevertheless generically simplifies to 0:
Machine-precision can give incorrect numerical results on the branch cut:
Machine-number inputs can give arbitrary-precision results:
Powers can be very large:
Some powers are too large for any computer:
Powers can give indeterminate expressions:
The precision of each result is determined by the precision of the zero:
Symbolic powers of 1 are only evaluated when the 1 is an exact or machine-precision number:
Numerical decision procedures with default settings cannot simplify this power:
Machine-precision numerical evaluation is inadequate:
Higher internal precision resolves the result:
Nonrational powers are not absorbed into series:
Power applies element-wise to matrices:
Use
MatrixPower for matrix powers: