each agent has a list of preferences - things they want to do.
each element on these lists is weighted.
the set of agents (abe and barb) have to decide what they will do as a group, based on the set of abe's preferences and barb's preferences.
the simplest case is binary - abe wants either A or !A, barb wants either A or !A, and you apply some logic gate that decides to do either A or !A based on whatever rule (but what rule should you use, in this case?)
but in a more complex case, abe wants 50% A, 30% B, 20% C. barb wants 40% B, 40% !B, and 20% D.
if they want multiple things, should you try to do all of the things, or only the things they most strongly want to do? if they want to do things that contradict each other, should you want to do neither thing, or both, or roll the dice, or what?