It is not possible to encode the exploit as you would encode a payload. The reason is that the data that is part of an exploit is directly processed by the vulnerable application (in this case Flash). The function inside the Flash binary which process the exploit data won't decode it first. In case of payloads we can encode it because when the payload gets executed, we already have code execution privileges on the box so we can first decode the payload and then execute it.
There are a lot of ways through which you can bypass the signature based IDS rules. I am not sure if you have access to the actual signature code (most probably the regex) but if you have, you can study it and see what it is looking for. In my past experience with Cisco ASA IDS, the signatures were so dump you can practically replace ' OR '1'='1' --
with ' OR '1000'='1000' --
and it would work.