Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!hellgate.utah.edu!caen!sdd.hp.com!think.com!mintaka!snorkelwacker.mit.edu!shelby!agate!ucbvax!vax.oxford.ac.uk!POPX
From: POPX@vax.oxford.ac.uk (Jocelyn Paine)
Newsgroups: comp.ai
Subject: Games for programmed players
Message-ID: <9102062219.AA18294@ucbvax.Berkeley.EDU>
Date: 6 Feb 91 16:00:00 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Lines: 63

Newsgroups: comp.ai
Subject: Games for programmed players
Summary:
Expires:
Sender:
Reply-To: popx@vax.ox.ac.uk (Jocelyn Paine)
Followup-To:
Distribution: comp.ai, comp.simulation
Organization: Experimental Psychology, Oxford University, UK.
Keywords:

The AI magazine published, in Autumn 1983, an article called "Knowledge
Programming in Loops", by Mark Stefik, Daniel Bobrow, Sanjay Mittal and
Lynn Conway. This described an experimental course at Xerox Parc about
building expert systems in Loops.

The course was centered round a game called "Truckin'" in which the
players drove lorries round a board populated with shops. To win, you
had to get to the end of the board, having bought and sold from the
shops in such a way as to make more profit then any other player. Or to
be exact: your _player_ had to do this. _Your_ job was to program your
player, using expert system techniques, so as to make it win.

Things were made more interesting by features designed to cause goal
conflicts. For example, if your lorry drove quickly over a bumpy road,
it might damage its load; but if the load was perishable, it would go
off if not delivered quickly. There were highwaymen, robbing from
lorries that passed too slowly; and weighing stations, fining lorries
which went too fast or were too heavy.

Ineptly programmed lorries would reveal themselves in various ways:

    "A player may be racing to Alice's Restaurant [the final stop]. One
    move before the game ends it is unable to resist a business
    'opportunity' and doesn't make it to Alice's.

    A player may go to the closest place to sell goods, even if that's
    the City Dump, which unfortunately pays a 'negative price'.

    A player may become focussed on a tight producer/consumer loop,
    making money faster than any other player on the board. If it is
    programmed to only buy fuel from stations along its route, but there
    is no petrol station in the tight loop, the team will watch
    anxiously as the fuel gauge drops lower and lower.

    A player may try to park next to Alice's Restaurant near the end of
    the game, even if it happens to be the Union hall, which confiscates
    all goods and cash."

All this comes from the original paper, and I've described it in case
anyone wants to take up the idea for teaching.


I would like to ask whether anyone else has designed such games, or if
anyone knows references to them. I'll summarise replies to the net.
Somebody mentioned a game called "Robot Wars" (_not_ Core Wars, which is
the one in which machine code programs fight it out) but I've been
unable to trace it.

I'd also like to ask for the E-mail address of the authors, so that I
can ask for more details about the rules of Truckin'.

                                    Jocelyn Paine ( POPX @ UK.AC.OX.VAX )