...Durfee
\ Supported, in part, by NSF grant IRI-9158473.
```.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
```
...practical
There are, of course, other approximation techniques that increase practicality which rely on various abstractions of the problem [1].
```.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
```
...knowledge
Common knowledge about a fact x means that everybody knows that everybody knows...that everybody knows x.
```.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
```
...themselves.
This is in contrast to Russell's work, where each node has a utility associated with it.
```.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
```
...believed
We can anthropomorphize this advice to say that, it does not help to pretend things are better than they probably are; we must determine exactly where we stand and take the optimal action based on this assessment.
```.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
```
...one.
Test results show that this approximation is, indeed, good enough.
```.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
```
...strategy
In a pure strategy one action is chosen with probability 1, the rest with probability 0.
```.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
```
...times
Some basic calculations, along these lines, are shown in Section 4.1.
```.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
```

Jose M. Vidal
jmvidal@umich.edu
Sun Mar 10 12:52:06 EST 1996