1. Could you make it so it weighs 101% success higher than 99.5%? The optimizer will pick three under before five over, and that's kinda annoying.
2. It will put in "ignores +/-" ships if there is an event that lowers stats (like Miracle Worker), if that ship will get closer to 100% than some other options. This seems like a real waste to me. The Phoenix Replica shuttle in particular really messes with a lot of events, and will often enough eat up most or all of the ten or so options you let the program give us.
3. If an assignment gets into negative stats with an event, the optimizer treats it like it was the normal stat, rather than zero like the game does.
Since version 1.0, a new scoring ship formula was added. This formula is not only much simpler but has also proven to be far better at avoiding the problems you've listed. Note that ASO only uses the new formula when you set the desired critical rating for an assignment above 0%. If critical rating is set at 0%, the old scoring formula will still be used.
For those who are interested in the details, both formulas are used to score how well a set of ships matches an assignment's requirements. An example of a ship-set would be the Delta Flyer, Presidio Command Battlecruiser (T6) and U.S.S. Rhode Island.
Under the old ship-set scoring formula, the ship-set scores well if the set is able to match the assignment's requirements. Which means the ship-set must match 3 conditions, namely the assignment's ENG, TAC and SCI requirements. This isn't an easy requirement to meet and with my Fed characters who have 80+ ships, only 2-5 ship-sets actually come close to matching these requirements. And out of this small handful of ship-sets, there are invariably a few that comes in a few points shy of meeting the assignment's requirement. This happens despite of having x3 weighting against "coming in at 99.5%" of the assignment's requirements.
The new formula came about because ASO had added support for hitting specific critical rating values. So instead of having to hit ENG, TAC, SCI targets, there's only one target now - critical rating value. This turned out to offer several huge and unexpected advantages.
The biggest advantage is really that a ship-set has to exceed the assignment requirements before it could even match the critical rating value requirement. Because of this, I could give a huge x10 negative weight to ship-sets that don't even meet the assignment requirements. There are other advantages such as a much simpler formula that mostly only consider critical rating and a much bigger number of ship-sets that meet the simplified requirement. All of which greatly reduces the chance of problems 2 & 3 actually being a real problem.
So all things considered, the new scoring formula turned out to be a much greater boon than I had anticipated. It's getting to the point where I'm seriously considering removing the old scoring formula and just stick with this new one.