So I am wondering if the following algorithm can work.

1) Cast each PC member's choice into a linear utility as follows: YES for a paper means weight C_1, MAYBE means weight C_2, and NO means weight EPSILON, where C_1 > C_2 > EPSILON >=0.

2) Compute the market equilibrium using the Eisenberg-Gale program (or using a combinatorial algorithm by Devanur et al).

3) Use the market equilibrium as a probability distribution for randomly assigning papers among PC members.