If i was in charge (which i am not

), i would not separate things into too many categories. Someone has to judge them all, which is hardly doable for 10 cathegories with 3 winners each. Also, this would mean 30 winners, how many entries do we expect?!
I'd go for this:
Categories:
(1 vs 1) competitive maps
(2 vs 2) competitive maps
(3 vs 3)/(4 vs 4) big team game maps with less of a focus on competitiveness
Each category has one winner and one runner up.
Judgement
Step 1)
Community vote for the top 5 entries of each cathegory via polls.
This way we would make sure that the top entries are well received by the community and the fun aspect is good.
Step 2)
A qualified judgement board with deep knowledge of the worldbuilder tool and mapping in general will have the final judge on the top 5 and pick the winners.
Technical aspects like optimization and detailing can only really be judged by qualified people. This way it is made sure that the winner maps are both fun to play and also high quality from the technical point of view