Demonstrations

The demonstrations for this problem follow a pattern similar to that of example 1. Sampling was not used, and so a belief model with no error is presumed. Unlike example 1, this problem has two variables, the user's intend(book-flight-window) and the agent's bel(have-seat). These beliefs occur at different levels of nesting. For example, bel(have-seat) needs to be evaluated at level 2 corresponding with its occurrence the second level of the game tree. it also is evaluated at level 4, corresponding with its occurrence the fourth level of the game tree. For the purpose of simplifying the demonstration by reducing the number of variables, it is assumed that beliefs are mutual, which renders level 2 and level 4 beliefs equal, and reduces the problem to one of two variables. This is a reasonable assumption in the usual circumstances that acts are always observable to both agents and so all the belief revisions that develop their belief sets make mutual updates ( see Section 3.4.7). The following demonstrations present and explain the output of the planner.