Summary: Small notes on missionaries---try 4 \Lambda
John McCarthy Eyal Amir
April 1, 1996
Here are some principles of this try. When this memo is debugged it will be glued
into the earlier ``Missionaries and Cannibals: Making it Elaboration Tolerant''.
1. Using situation calculus and not asserting that situations that agree on a
specific set of fluents are equal allows greater elaboration tolerance than
the state based formalisms of the previous sections. In that formalism,
two states were equal if the missionaries, cannibals and the boat were on
the same bank. In the situation calculus formalism, new objects such as
the oars can be added.
2. Groups of missionaries or cannibals will be identified with the sets of their
members. I think we may lose some elaboration tolerance from this, but
I don't have a definite reason for believing it.
3. When an action succeeds, the result of trying the action is equal to the
result of doing the action. In missionaries and cannibals we use tryg, since
who tries is not really identified, and we can suppose it is some god.
4. The places are lb and rb. In this try we do not take into account the fact
that a river is long. There is water between lb and rb.