`OP_POW` not defined about homotopycontinuation.jl HOT 5 CLOSED

Ah yes the square root is the problem, but this is actually exactly the case that I wanted to support in the first place. I will try to take a look soon :)

from homotopycontinuation.jl.

The code path is only hit if the exponent is not an integer, so OP_POW_INT is most likely wrong. What type of exponents do you have?

When implementing the procedure, I wanted to deal with general exponents initially but I abandoned this approach since it would only be allowed in very specific circumstances. However, I think just supporting rational numbers or floating point values should be easily feasible.

from homotopycontinuation.jl.

Many thanks for responding so quickly. I have managed to make a minimum working example:

using HomotopyContinuation

@var x
@var a b

vars = HomotopyContinuation.Variable[x]
para = HomotopyContinuation.Variable[a, b]

pol = x*sqrt((a^2 - b^2))
eq = Expression[pol]

sys = System(eq, variables = vars, parameters = para)

solve(sys, start_system=:total_degree, target_parameters=Number[0.99, 0.98])

To problem seems to be the sqrt((a^2 - b^2)) in the expression, i.e., we have an exponent of 1/2.

from homotopycontinuation.jl.

I have tested out the new PR and it solves my original problem I experienced in HarmonicBalance.jl. Thank you!

from homotopycontinuation.jl.

Happy to hear that :) Just tagged a new release (2.8.3) that contains the fix

from homotopycontinuation.jl.