----------------------------------------

From: at354bro@cdf.utoronto.ca (Andrew Brown)
Subject: PS 3
Date: 2 Apr 97 15:04:14 GMT

There are still a few more things that people have been
asking about.

1. The "expected time-average number of tankers "in the queues,
means the time average length of each queue. QTABLE has a function
which will compute this for you. The "expected average in-port
residence time" of each tanker is the time that the tanker spends
in the system, from the point when it arrives until it is dropped
off in the harbour by the tug.

2. If there are still situations in the system which you think are
ambiguous under the rules for the various protocols, then state in
your report what the situation is and what you implemented.

3. Some additional comments about the hard copy hand-in. If you haven't
been able to get all three policies working properly then please state
at the beginning of the report, which parts you got working and which
you did not. This will make it easier to evaluate what you've done.
So, here's what you should have in your hardcopy hand-in:

        -A statement of what works and what doesn't

        -A statement of the relative goodness of the three operating
        policies, with reference to your simulations. If you've got
        everything working and you've got data from all three simulations,
        then you should also compute comparison statistics between the
        baseline and P1, and the baseline and P2. If you are trying to
        argue that one policy is better than another then it's important
        to be able to say whether the difference is "statistically
        significant". But only after everything else is complete.

        -A print out of three tables (one for each policy) showing
        the estimates and confidence intervals of each statistic. If
        you've gotten as far as computing the relative comparisons of
        the alternatives then give those tables as well.

        -A print out of the code for your three simulations

----------------------------------------

From: at354bro@cdf.utoronto.ca (Andrew Brown)
Subject: Re: PS 3
Date: Date: 2 Apr 97 20:18:27 GMT

>-A print out of three tables (one for each policy) showing
>the estimates and confidence intervals of each statistic.

When I said one table for each policy I just meant the print out
of the numbers for each simulation. I was not referring to the TABLE
data structure in CSIM. Sorry if I caused anyone distress.

Here's an example:

TABLE 1: Baseline Protocol

The expected portion of time that the tug is:
Idle:                           ### +/- ###
Traveling without a tanker:     ### +/- ###
Engaged:                        ### +/- ###

The expected portion of the time that each berth is:
Unoccupied:                     ### +/- ###
Occupied and not loading:       ### +/- ###
Loading:                        ### +/- ###

etc...


TABLE 2: Protocol 1

The expected portion of time that the tug is:
Idle:                           ### +/- ###
Traveling without a tanker:     ### +/- ###
Engaged:                        ### +/- ###

The expected portion of the time that each berth is:
Unoccupied:                     ### +/- ###
Occupied and not loading:       ### +/- ###
Loading:                        ### +/- ###

...some more nice stats and confidence intervals...


TABLE 3: Protocol 2

The expected portion of time that the tug is:

...more of same...


And then if you've done comparison stats of the Baseline vs. P1
and Baseline vs. P2, include those in your report too.

----------------------------------------