CSCI-561 – Fall 2019 – Foundations of Artificial Intelligence
Due October 21, 2019 23:59:59
Image from cardboard_box @ stackexchange
This is a programming assignment. You will be provided sample inputs and outputs (see below).
Please understand that the goal of the samples is to check that you can correctly parse the
problem definitions and generate a correctly formatted output. The samples are very simple and
it should not be assumed that if your program works on the samples it will work on all test cases.
There will be more complex test cases and it is your task to make sure that your program will
work correctly on any valid input. You are encouraged to try your own test cases to check how
your program would behave in some complex special case that you might think of. Since each
homework is checked via an automated A.I. script, your output should match the specified
format exactly. Failure to do so will most certainly cost some points. The output format is simple
and examples are provided. You should upload and test your code on vocareum.com, and you
will submit it there. You may use any of the programming languages provided by vocareum.com.
Your code will be tested as follows: Your program should not require any command-line
argument. It should read a text file called “input.txt” in the current directory that contains a
problem definition. It should write a file “output.txt” with your solution to the same current
directory. Format for input.txt and output.txt is specified below. End-of-line character is LF (since
vocareum is a Unix system and follows the Unix convention).
Note that if your code does not compile, or somehow fails to load and parse input.txt, or writes
an incorrectly formatted output.txt, or no output.txt at all, or OuTpUt.TxT, you will get zero
points. Anything you write to stdout or stderr will be ignored and is ok to leave in the code you
submit (but it will likely slow you down). Please test your program with the provided sample files
to avoid any problem.
Academic Honesty and Integrity
All homework material is checked vigorously for dishonesty using several methods. All detected
violations of academic honesty are forwarded to the Office of Student Judicial Affairs. To be safe
you are urged to err on the side of caution. Do not copy work from another student or off the
web. Keep in mind that sanctions for dishonesty are reflected in your permanent record and can
negatively impact your future success. As a general guide:
Do not copy code or written material from another student. Even single lines of code
should not be copied.
Do not collaborate on this assignment. The assignment is to be solved individually.
Do not copy code off the web. This is easier to detect than you may think.
Do not share any custom test cases you may create to check your program’s behavior in
more complex scenarios than the simplistic ones considered below.
Do not copy code from past students. We keep copies of past work to check for this. Even
though this problem differs from those of previous years, do not try to copy from
homeworks of previous years.
Do not ask on piazza how to implement some function for this homework, or how to
calculate something needed for this homework.
Do not post code on piazza asking whether or not it is correct. This is a violation of
academic integrity because it biases other students who may read your post.
Do not post test cases on piazza asking for what the correct solution should be.
Do ask the professor or TAs if you are unsure about whether certain actions constitute
dishonesty. It is better to be safe than sorry.
In this project, we will play the game of Halma, an adversarial game with some similarities to
checkers. The game uses a 16×16 checkered gameboard. Each player starts with 19 game pieces
clustered in diagonally opposite corners of the board. To win the game, a player needs to
transfer all of their pieces from their starting corner to the opposite corner, into the positions
that were initially occupied by the opponent. Note that this original rule of the game is subject
to spoiling, as a player may choose to not move some pieces at all, thereby preventing the
opponent from occupying those locations. Note that the spoiling player cannot win either
(because some pieces remain in their original corner and thus cannot be used to occupy all
positions in the opposite corner). Here, to prevent spoiling, we modify the goal of the game to
be to occupy all of the opponent’s starting positions which the opponent is not still occupying.
See http://www.cyningstan.com/post/922/unspoiling-halma for more about this rule
In more details (from https://en.wikipedia.org/wiki/Halma):
Setup for two players:
Note: we only consider the two-player variant here; this game can also be played by four players
but we will not explore this here.
– Simple wooden pawn-style playing pieces, often called “Halma pawns.”
– The board consists of a grid of 16×16 squares.
– Each player’s camp consists of a cluster of adjacent squares in one corner of the board.
These camps are delineated on the board.
– For two-player games, each player’s camp is a cluster of 19 squares. The camps are in
– Each player has a set of pieces in a distinct color, of the same number as squares in each
– The game starts with each player’s camp filled by pieces of their own color.
The initial setup is shown below for black and white players. We will always use this exact initial
setup in this homework.
We first describe the typical play for humans. We will then describe some minor modifications
for how we will play this game with artificial agents.
– Create the initial board setup according to the above description.
– Players randomly determine who will move first.
– Pieces can move in eight possible directions (orthogonally and diagonally).
– Each player’s turn consists of moving a single piece of one’s own color in one of the
o One move to an empty square:
§ Move the piece to an empty square that is adjacent to the piece’s original
position (with 8-adjacency).
§ This move ends the play for this player’s turn.
o One or more jumps over adjacent pieces:
§ An adjacent piece of any color can be jumped if there is an empty square
on the directly opposite side of that piece.
§ Place the piece in the empty square on the opposite side of the jumped
§ The piece that was jumped over is unaffected and remains on the board.
§ After any jump, one may make further jumps using the same piece, or end
the play for this turn.
§ In a sequence of jumps, a piece may jump several times over the same
– Once a piece has reached the opposing camp, a play cannot result in that piece leaving
– If the current play results in having every square of the opposing camp that is not already
occupied by the opponent to be occupied by one’s own pieces, the acting player wins.
Otherwise, play proceeds to the other player.
Below we show examples of valid moves (in green) and invalid moves (in red). At left, the isolated
white piece can move to any of its empty 8 neighbors. At right, the central white piece can jump
over one adjacent piece if there is an empty cell on the other side. After one jump is executed,
possibly several other valid jumps can follow with the same piece and be combined in one move;
this is shown in the sequence of jumps that start with a down-right jump for the central piece.
Note that additional valid moves exist that are not shown (e.g., the central white piece could
move to some adjacent empty location).
Note the invalid moves: red arrow going left: cannot jump over one or more empty spaces plus
one or more pieces. Red arrow going left-down: cannot jump over one or more pieces plus one
or more empty spaces. Red arrow going down: cannot jump over more than one piece.
Playing with agents
In this homework, your agent will play against another agent, either created by the TAs, or
created by another student in the class. For grading, we will use two scenarios:
1) Single move: your agent will be given in input.txt a board configuration, a color to play,
and some number of seconds of allowed time to play one move. Your agent should return
in output.txt the chosen move(s), before the given play time has expired. Play time is
measured as total CPU time used by your agent on all CPU threads it may spawn (so,
parallelizing your agent will not get you any free time). Your agent will play 10 single
moves, each worth one point. If your agent returns an illegal move, a badly formatted
output.txt, or does not return before its time is up, it will lose the point for that move.
2) Play against reference agent: your agent will then play 9 full games against a simple
minimax agent with no alpha-beta pruning, created by the TAs. There will be a limited
total amount of play time available to your agent for the whole game (e.g., 100 seconds),
so you should think about how to best use it throughout the game. This total amount of
time will vary from game to game. Your agent must play correctly (no illegal moves, etc)
and beat the reference minimax agent to receive 10 points per game. Your agent will be
given the first move on 5 of the 9 games. In case of a draw, the agent with more remaining
play time wins.
Note that we make a difference between single moves and playing full games because in single
moves it is advisable to use all remaining play time for that move. While playing games, however,
you should think about how to divide your remaining play time across possibly many moves
throughout the game.
In addition to grading, we will run a competition where your agent plays against agents created
by the other students in the class. This will not affect grade. But the top agents will be referred
to a contact at Google for an informal introduction. There will also be a prize for the grand winner.
Agent vs agent games:
Playing against another agent will be organized as follows (both when your agent plays against
the reference minimax agent, or against another student’s agent):
A master game playing agent will be implemented by the grading team. This agent will:
– Create the initial board setup according to the above description.
– Randomly assign a player color (black or white) to your agent.
– When playing against the reference minimax, you will get the opening move. Otherwise
who plays first will be chosen randomly.
– Then, in sequence, until the game is over:
o The master game playing agent will create an input.txt file which lets your agent
know the current board configuration, which color your agent should play, and
how much total play time your agent has left. More details on the exact format of
input.txt are given below.
o We will then run your agent. Your agent should read input.txt in the current
directory, decide on a move, and create an output.txt file that describes the move
(details below). Your time will be measured (total CPU time). If your agent does
not return before your time is over, it will be killed and it loses the game.
o Your playing time remaining will be updated by subtracting the time taken by your
agent on this move. If time left reaches zero or negative, your agent loses the
o The validity of your move will be checked. If the format of output.txt is incorrect
or your move is invalid according to the rules of the game, your agent loses the
o Your move will be executed by the master game playing agent. This will update
the game board to a new configuration.
o The master game playing agent will check for a game-over condition. If so, the
winning agent will be declared the winner of this game.
o The master game playing agent will then present the updated board to the
opponent agent and let that agent make one move (with same rules as just
described for your agent; the only difference is that the opponent plays the other
color and has its own time counter).
Input and output file formats:
Input: The file input.txt in the current directory of your program will be formatted as follows:
First line: A string SINGLE or GAME to let you know whether you are playing a single move
(and can use all of the available time for it) of playing a full game with potentially
many moves (in which case you should strategically decide how to best allocate
your time across moves).
Second line: A string BLACK or WHITE indicating which color you play. The colors will always be
organized on the board as follows:
(black starts in the top-left corner and white in the bottom-right).
Third line: A strictly positive floating point number indicating the amount of total play time
remaining for your agent.
Next 16 lines: Description of the game board, with 16 lines of 16 symbols each:
§ W for a grid cell occupied by a white piece
§ B for a grid cell occupied by a black piece
§ . (a dot) for an empty grid cell
In this example, your agent plays a single move as white color and has 100.0 seconds. The
board configuration is just the one from the start of the game.
Output: The file output.txt which your program creates in the current directory should be
formatted as follows:
1 or more lines: Describing your move(s). There are two possible types of moves (see above):
E FROM_X,FROM_Y TO_X,TO_Y – your agent moves one of your pieces from location
FROM_X, FROM_Y to adjacent empty location TO_X, TO_Y. We will again use zero-based,
horizontal-first, start at the top-left indexing in the board, as in homework 1. So, location
0,0 is the top-left corner of the board; location 15,0 of the top-right corner; location 0,15
is the bottom-left corner, and location 15,15 the bottom-right corner. As explained above,
TO_X,TO_Y should be adjacent to FROM_X,FROM_Y (8-connected) and should be empty.
If you make such a move, you can only make one per turn.
J FROM_X,FROM_Y TO_X,TO_Y – your agent moves one of your pieces from location
FROM_X,FROM_Y to empty location TO_X,TO_Y by jumping over a piece in between. You
can make several such jumps using the same piece, as explained above, and should write
out one jump per line in output.txt.
For example, output.txt may contain:
E 11,15 10,15
This moves the leftmost white piece on the bottom row of the board to the left. The resulting
board would look like this, given the above input.txt:
Or it could contain:
J 12,15 10,13
which is a jump along the left-up diagonal for a bottom piece. Note that here we have only one
jump, but there could be several (one per line in output.txt) if the board configuration allows
(see example 3 below). The board would look like this after this jump, given the above input.txt:
Notes and hints:
– Please name your program “homework.xxx” where ‘xxx’ is the extension for the
programming language you choose (“py” for python, “cpp” for C++, and “java” for Java).
If you are using C++11, then the name of your file should be “homework11.cpp” and if
you are using python3 then the name of your file should be “homework3.py”.
– The board you will be given as input will always have nineteen W letters, nineteen B
letters, and the rest will be . (empty space).
– Likely (but not guaranteed), total play time will be 5 minutes (300.0 seconds) when
playing against another agent, and 30.0 seconds for single moves.
– Play time used on each move is the total combined CPU time as measured by the Unix
time command. This command measures pure computation time used by your agent,
and discards time taken by the operating system, disk I/O, program loading, etc. Beware
that it cumulates time spent in any threads spawned by your agent (so if you run 4 threads
and use 400% CPU for 10 seconds, this will count as using 40 seconds of allocated time).
– If your agent runs for more than its given play time (in input.txt) + 10 seconds (grace
period), it will be killed and will lose the single move or the game.
– The grace period is only so that we do not kill your agent prematurely, and you should
not plan on using it. You should aim to write out your output before your allocated time
has passed. The actual play time taken by your agent will be subtracted after your agent
returns, and you will lose if your agent ends up exceeding its allocated play time.
– You need to think and strategize how to best use your allocated time. In particular, you
need to decide on how deep to carry your search, on each move. In some cases, your
agent might be given only a very short amount of time (e.g., 5 seconds, or even 0.01
seconds), for example towards the end of a game. Your agent should be prepared for that
and return a quick decision to avoid losing by running over time. There is no lower bound
on the amount of play time that will be given in input.txt except that it will always be >0.
– To help you with figuring out the speed of the computer that your agent runs on, you are
allowed to also provide a second program called calibrate.xxx (same extension
conventions as for homework.xxx). This is optional. If one is present, we will run your
calibrate program once (and only once) before we run your agent for grading or against
another agent. You can use calibrate to, e.g., measure how long it takes to expand some
fixed number of search nodes. You can then save this into a single file called
calibration.txt in the current directory. When your agent runs during grading or during a
game, it could then read calibration.txt in addition to reading input.txt, and use the data
from calibration.txt to strategize about search depth or other factors.
– You need to think hard about how to design your eval function (which gives a value to a
board when it is not game over yet).
– You are allowed to maintain persistent data across moves during a game, by writing such
data to a single file called playdata.txt in the current directory. Before a new game starts,
the master game playing agent will delete any playdata.txt file. So on your first move, this
file will not exist and you should be prepared for that. Then, you can write some data to
that file at the end of a move, and read that file back at the beginning of the next move.
We expect that simple content in a format of your choice would be sufficient, but if you
want to save/load complex data structures, have a look at things like boost::serialization
or the C++11, header-only cereal library at https://github.com/USCiLab/cereal (written
by our former students).
– Note that there is an advantage in this game for the player that gets the first move. We
give your agent that advantage when playing against the reference minimax agent by
giving your agent the first move for 5 of the 9 games. But, in the other 4 games, your
agent should still be “smarter” than plain minimax and should still be able to win. In the
competition, we will play two agents against each other for an even number of games,
and advance both agents to the next round of the competition if they both win half of the
games. If an agent loses more than half of the games, it will be eliminated and only the
other agent will move to the next round of the competition. We may end up with several
equivalent winners of the competition.
For this input.txt:
one possible correct output.txt is:
E 4,3 3,2
For this input.txt:
one possible correct output.txt is:
J 4,3 2,1
For this input.txt:
one possible correct output.txt is:
J 9,9 11,9
J 11,9 11,11
J 11,11 13,13
CSCI-561 Fall 2019
Homework 2 addendum
Following discussion on Piazza about how to best prevent spoiling, the rules are updated as
follows (see thread 334 and others on piazza):
1. Moving pieces
a) Players cannot make a move that starts outside their own camp and causes one of their
pieces to end up in their own camp.
b) If a player has at least one piece left in their own camp, they have to
• Move a piece out of their camp (i.e. at the end of the whole move the piece ends up
outside of their camp).
• If that’s not possible, move a piece in their camp further away from the corner of their
own camp ([0,0] or [15,15] respectively).
Only if the player does not have any pieces left in their camp or none of the two alternatives
above are possible are they free to move pieces outside of their camp.
Note: To move “further away”, you should simply move so that you either move further away
horizontally (while not moving closer vertically), or vertically (while not moving closer
horizontally), or both.
2. Winning the game (no real change here, just a clarification)
A player wins if they are the first to fill out all the space in the opposite camp that’s not
occupied by the opposite player’s pieces using at least one of their pieces.
What you should do:
If you already have some code that plays according to the original rules, likely somewhere you
have a function that lists or loops over all the moves that your agent could take given the
current state of the board. You should modify this function slightly to remove the possible
moves that would violate the new rules above. We expect that no change should be necessary
to your core minimax and alpha-beta code, although this may depend on your exact
Time is extended to October 25, 2019 at 11:59:59pm
To not penalize students who finish by the original date, a bonus of 5% will be given to
submissions made before the original due date of October 21, 2019 at 11:59:59pm
本网站支持淘宝 支付宝 微信支付 paypal等等交易。如果不放心可以用淘宝交易！
E-mail: [email protected] 微信:itcsdx