Any fix to batched optimization?

[TianningGao]

I noticed there's a "batch_size" parameter for experiments.run_one_replication() but in experiments_util.get_acqf(), line 465 - 508, it will trigger some error on purpose if batch size is not 1. I tried to surpass the "batch_size == 1" condition but it triggers error elsewhere.

Is there any fix to enable batched optimization?

[Balandat]

I don't think we ran any batch sizes q > 1 for the paper, so I'm not sure this is expected to work here - @sdaulton?

In general it's not hard to support this, but this is an archival repo, and so we wouldn't make the changes here (and instead in the BoTorch PR here: pytorch/botorch#1533).

[TianningGao]

Got it. Thanks.