What is "total bias" in SimulatedAnnealingSampler?
There are two parts to this question
1) In the documentation for SimulatedAnnealingSampler, what is meant by "total bias"? It is mentioned in the context of the "beta_range" in the documentation.
There is only one value for bias for a single node in the graph, right? For a `beta_schedule_type='custom'` it's not clear to me how I should set the beta_range based on this explanation: "Default range is set based on the total bias associated with each node."
2) I imagine there must be some way to get the largest bias from a BQM, but I'm not able to find a command for it in the documentation. Can someone point me towards the relevant docs for this?
Comments
Hello,
1) What is "total bias" referring to in the beta_range documentation?
Take a look at this part of the code where the default beta_range is generated:
https://github.com/dwavesystems/dwave-samplers/blob/1.2.0/dwave/samplers/sa/sampler.py#L497-L498
2) How to get the largest bias:
The bias is referring to the linear and quadratic biases.
To get the largest bias in a BQM you can use the following snippet:
It might be best to get the largest unsigned bias:
Alternatively, depending on your problem, if you are entering the biases into a BQM yourself, you could keep track of the largest bias as you enter the values.
I hope this was helpful! Please let us know if you have any follow-up questions.
Please sign in to leave a comment.