School of Computing and Information Systems
comp10002 Foundations of Algorithms
Semester 2, 2021
Assignment 2
Learning Outcomes
In this project, you will demonstrate your understanding of dynamic memory and linked data structures (Chapter
10) and extend your program design, testing, and debugging skills. You will also learn about Artificial Intelligence
and tree search algorithms, and implement a simple algorithm for playing checkers.
Checkers
Checkers, or draughts, is a strategy board game played by two players. There are many variants of checkers.
For a guide to checker’s families and rules, see https://www.fmjd.org/downloads/Checkers_families_
and_rules.pdf. Your task is to implement a program that reads, prints, and plays our variant of the game.
(d) Legal captures
Figure 1: Example board configurations, moves, and captures.
Setup. An 8x8 chessboard with 12 black and 12 white pieces initially positioned as shown in Figure 1a.
Gameplay. Each player plays all pieces of the same color. Black open the game by making a move, then white
make a move, and then players alternate their turns. In a single turn, the player either makes a move or capture.
For example, the arrow in Figure 1a denotes an opening move of the black piece from cell G6 to cell F5.
Moves. A piece may move to an empty cell diagonally forward (toward the opponent; north for black and south
for white) one square. The arrows in Figure 1b show all the legal moves of black and white pieces.
Towers. When a piece reaches the furthest row (the top row for black and the bottom row for white), it becomes
a tower (a pile of pieces). The only move of the white piece at cell D7 in Figure 1b promotes it to the tower. A
tower may move to an empty cell diagonally, both forward and backward, one square. The arrows in Figure 1c
show all the legal moves of black and white towers.
Captures. To capture a piece or a tower of the opponent, a piece or a tower of the player jumps over it and lands
in a straight diagonal line on the other side. This landing cell must be empty. When a piece or tower is captured,
it is removed from the board. Only one piece or tower may be captured in a single jump, and, in our variant of
the game, only one jump is allowed in a single turn. Hence, if another capture is available after the first jump, it
cannot be taken in this turn. Also, in our variant of the game, if a player can make a move or a capture, they may
decide which of the two to complete. A piece always jumps forward (toward the opponent), while a tower can
jump forward and backward. The arrows in Figure 1d show all the legal captures for both players.
Game end. A player wins the game if it is the opponent’s turn and they cannot take action, move or capture,
either because no their pieces and towers are left on the board or because no legal move or capture is possible.
Input Data
Your program should read input from stdin and write output to stdout. The input should list actions, one per
line, starting from the initial setup and a black move. The input of moves and captures can be followed by a
single command character, either ‘A’ or ‘P’. Action should be specified as a pair of the source cell and the target
cell, separated by the minus character ‘-’. For example, “G6-F5” specifies the move from cell G6 to cell F5.
1
The following file test1.txt uses eleven lines to specify ten actions, followed by the ‘A’ command.
Stage 0 – Reading, Analyzing, and Printing Input Data (8/20 marks)
The first version of your program should read input and print the initial setup and all legal actions. The first 42
lines that your program should generate for the test1.txt input file are listed below in the two left columns.
1 BOARD SIZE: 8x8
2 #BLACK PIECES: 12
3 #WHITE PIECES: 12
4 A B C D E F G H
Lines 1–21 report the board configuration and specify the initial setup from Figure 1a. We use ‘b’ and ‘w’
characters to denote black and white pieces, respectively. Then, lines 22–42 print the first move specified in the
first line of the input. The output of each action starts with the delimiter line of 37 ‘=’ characters; see line 22.
The next two lines print information about the action taken and the cost of the board; see lines 23 and 24. The
cost of a board is computed as b + 3B − w − 3W, where b, w, B, and W are, respectively, the number of black
pieces, white pieces, black towers, and white towers on the board; that is, a tower costs three pieces. Then, your
program should print the board configuration that results from the action. The complete output your program
should generate in Stage 0 for the test1.txt input file is provided in the test1-out.txt output file.
If an illegal action is encountered in the input, your program should select and print one of the following six
error messages. Your program should terminate immediately after printing the error message.
1 ERROR: Source cell is outside of the board.
2 ERROR: Target cell is outside of the board.
3 ERROR: Source cell is empty.
4 ERROR: Target cell is not empty.
5 ERROR: Source cell holds opponent’s piece/tower.
6 ERROR: Illegal action.
The conditions for the errors are self-explanatory and should be evaluated in the order the messages are listed.
Only the first encountered error should be reported. For example, if line 2 of the test1.txt file is updated to
state “G2-A8”, line 43 of the corresponding output should report: “ERROR: Target cell is not empty.”
Stage 1 – Compute and Print Next Action (16/20 marks)
If the ‘A’ command follows the input actions (see line 11 in the test1.txt file), your program should compute
and print the information about the next Action of the player with the turn. All of the Stage 0 output should be
retained. To compute the next action, your program should implement the minimax decision rule for the tree
depth of three. Figure 2 exemplifies the rule for the board configuration in Figure 3a and the turn of black.
First, the tree of all reachable board configurations starting from the current configuration (level 0) and of the
requested depth is constructed; if the same board is reachable multiple times in the same tree, the corresponding
tree node must be replicated. For example, black can make two moves in Figure 3a: the tower at A6 can move
to B5 (Figure 3b) and the piece at C8 can move to D7 (Figure 3c); see level 1 in Figure 2. The tree in Figure 2
explicitly shows nodes that refer to 15 out of all 30 board configurations in the minimax tree of depth three for
the black turn and the board from Figure 3a. The labels of the nodes refer to the corresponding boards shown
in Figure 3. For instance, nodes with labels (f)–(h) at level 2 of the tree refer to the boards in Figures 3f to 3h,
respectively, which are all the boards white can reach by making moves and captures in the board in Figure 3c.
Figure 2: A minimax tree.
Second, the cost of all leaf boards is
computed; see the nodes highlighted with
a gray background (level 3). For example,
six board configurations at level 3 of the
tree can be reached from the board in Figure
3d. The boards in Figures 3i and 3j
have the cost of 3, while the four boards
reachable via all moves of the tower at cell
B5 in board (d), not shown in the figure,
have the cost of 2. Intuitively, a positive
cost suggests that black win, a negative
cost tells that white win, and the magnitude of the cost indicates the advantage of one player over the other.
Third, for each possible action of the player, we check all possible actions of the opponent, and choose the
next action of the player to be the first action on the path from the root of the tree toward a leaf node for which the
player maximizes their gain while the opponent (considered rational) aims to minimize the player’s gain; see the
red path in Figure 2. Black take actions in the boards at level 2 of the tree. At this level, black aim to maximize
the gain by choosing an action toward a board of the highest cost. This is cost 3 for board (d) toward boards
(i) and (j), 1 for (e) toward (k), 0 for (f) toward (m), 1 for (g) toward (n), and 1 for (h) toward (o); the arrows
from level 3 to level 2 encode the cost selections and node labels at level 2 encode propagated costs. White take
actions in the boards at level 1 of the tree. White aim to maximize their gain, which translates into minimization
of the gain by black. Thus, at every board at level 1, white choose an action toward a board at level 2 with the
lowest cost propagated from level 3. This is cost 1 for board (b) toward (e), and 0 for (c) toward (f); again, see
the arrows from level 2 to level 1 and the costs as node labels at level 1. Finally, to maximize their gain, black
pick the next action in the game to be the one that leads to the board with the highest propagated cost at level 1.
This is move “A6-B5” toward board (b) on the path to (k); assuming white are rational and play “B7-A8” next.
To compute the next action for white, the order of max- and min-levels must be flipped. If several children
of the root have the same propagated maximum/minimum cost, the action to the left-most such child must be
chosen. To construct the children, both for black and white turns, the board is traversed in row-major order
(o)
Figure 3: Board configurations for possible evolutions of an example checkers endspiel.
3
and for each encountered piece or tower all possible actions are explored, starting from the north-east direction
and proceeding clockwise. Children should be added from left to right in the order they are constructed. Each
computed action should be printed with the “***” marker. Lines 1–21 in the right column of the listing in the
Stage 0 description show the output for the example computed action. Black and white towers are denoted by ‘B’
and ‘W’ characters, respectively. Boards in which black or white cannot take an action cost INT MIN and INT MAX,
respectively, defined in the
library. If the next action does not exist and, thus, cannot be computed,
the player loses, and the corresponding message is printed on a new line (“BLACK WIN!” or “WHITE WIN!”).
Stage 2 – Machines Game (20/20 marks)
If the ‘P’ command follows the input actions, your program should Play ten next actions or all actions until the
end of the game, whatever comes first. The game should continue from the board that results after processing
the Stage 0 input. If the game ends within the next ten turns (including the last turn when no action is possible),
the winner should be reported. The computation of actions and winner reporting should follow the Stage 1 rules.
Important...
The outputs generated by your program should be exactly the same as the sample outputs for the corresponding
inputs. Use malloc and dynamic data structures of your choice to implement the minimax rule for computing
the next action. Before your program terminates, all the malloc’ed memory must be free’d.
Boring But Important...
This project is worth 20% of your final mark, and is due at 6:00pm on Friday 15 October.
Submissions that are made after the deadline will incur penalty marks at the rate of two marks per day
or part day late. Students seeking extensions for medical or other “outside my control” reasons should email
ammoffat@unimelb.edu.au as soon as possible after those circumstances arise. If you attend a GP or other
health care service as a result of illness, be sure to obtain a letter from them that describes your illness and their
recommendation for treatment. Suitable documentation should be attached to all extension requests.
You need to submit your program for assessment via the LMS Assignment 2 page. Submission is not
possible through grok. There is a pre-submission test link also provided on the LMS page, so you can try your
program out in the test environment. It does not submit your program for marking either. Note that this
pre-test service will likely overload and fail on the assignment due date, and if that does happen, it will not be a
basis for extension requests. Plan to start early and to finish early!!
Multiple submissions may be made; only the last submission that you make before the deadline will be
marked. If you make any late submission at all, your on-time submissions will be ignored, and if you have not
been granted an extension, the late penalty will be applied. A rubric explaining the marking expectations is
linked from the LMS, and you should study it carefully. Marks will be available on the LMS approximately two
weeks after submissions close, and feedback will be mailed to your student email account.
Academic Honesty: You may discuss your work during your workshop, and with others in the class, but what
gets typed into your program must be individual work, not copied from anyone else. So, do not give hard copy
or soft copy of your work to anyone else; do not have any “accidents” that allow others to access your work;
and do not ask others to give you their programs “just so that I can take a look and get some ideas, I won’t
copy, honest”. The best way to help your friends in this regard is to say a very firm “no” if they ask to see
your program, pointing out that your “no”, and their acceptance of that decision, are the only way to preserve
your friendship. See https://academicintegrity.unimelb.edu.au for more information. Note also that
solicitation of solutions via posts to online forums, whether or not there is payment involved, is also Academic
Misconduct. In the past students have had their enrolment terminated for such behavior.
The FAQ page contains a link to a program skeleton that includes an Authorship Declaration that you must
“sign” and include at the top of your submitted program. Marks will be deducted (see the rubric linked from
the FAQ page) if you do not include the declaration, or do not sign it, or do not comply with its expectations. A
sophisticated program that undertakes deep structural analysis of C code identifying regions of similarity will
be run over all submissions. Students whose programs are identified as containing significant overlaps will
have substantial mark penalties applied, or be referred to the Student Center for possible disciplinary action.
Nor should you post your code to any public location (github, codeshare.io, etc) while the assignment is
active or prior to the release of the assignment marks.
© The University of Melbourne, 2021. The project was prepared by Artem Polyvyanyy, artem.polyvyanyy@unimelb.edu.au.