Semester project for COMP472.

The main file for this project is candycrisis.py

To run this on the lab machines at Concordia (Linux) you will need to run the following command: module load python/2.7.10 The generate_configs.py file will randomly generate problems for levels 1-4. For the demo we probably do not need to run this as we will be given inputs. The candycrisis.py file will execute the problem solving. After selecting automatic mode (by pressing B), you will need to select a level for the inputs. The levels have to be between 1-4, otherwise the program will crash :-) The input files associated with the levels must be passed as input{level}.txt. Note that this is singular, not plural (ex: "input1.txt"). You will need to run the program for each level you wish to get an output for.

The output of each level will be saved to a file called output{level}.txt, again, singular. After each run of the program, the "full" solution (i.e. the board after each move) for each puzzle will be saved under the /moves folder. Note that this will be overwritten after each run. So check it before running another challenge if you want to see :-)

Novice: 10 seconds for 50 puzzles

• Current best: Heuristic 3: 0.969126939774 secs

Apprentice: 12 seconds for 50 puzzles

• Current best: Heurstic 3: 2.86068296432 secs

Expert: 30 seconds for 30 puzzles

• Current best: Heuristic 3: 2.81453013423 secs

Master: 20 seconds for 10 puzzles

• Current best: Heuristic 3: 5.00174188614 secs

Automatic Mode (e.g. implementation of the search, output file generation, programming quality…)

• Refactor current code to let user choose between manual and automatic mode
• For automatic mode, output intermediate screens to file rather than on-screen
  • Perhaps save them in memory and only output at the end of the game ?
• Save & output the total number of moves it took to solve all the puzzles of the input file
• Choose heuristic to implement

Manual entries + Functionality of the rules of the game (e.g. visual trace after each move, detection of illegal moves, programming quality…)

Your program will read a series of initial configurations from a text file, where each line will represent an initial configuration. Each candy type will be represented by a unique character (‘r’, ‘b’, ‘w’, ‘y’, ‘g’ or ‘p’) and the empty slot will be indicated by the character ‘e’. For example, the line: e r r r r r b w b b w y b r y represents the first puzzle given on the previous page (empty at position A, Reese’s Pieces at positions B, C, D, E, F, a Bazooka Bubble Gum at G, a Walnetto at H, ...)

Your program should be able to run in manual mode and in automatic mode. This means that you should be able to run your program with:

1. Manual entry for the player, i.e. a human indicating the moves by hand. Note that if a human enters an illegal move, your program should give a warning and allow the user to enter a new move.
2. Automatic mode, i.e. your “AI” solving and indicating the moves. Note that if your program generates an illegal move, the entire puzzle solution will be considered wrong.

After each move, your program must display the new configuration of the candy box.

Your program must output: 1- A set of puzzle solutions: Your program must produce an output file, which contains, for each puzzle of the input file: a) the sequence of moves necessary to solve the puzzle, b) the time (real time in milliseconds in the lab’s desktops) required to solve the puzzle c) In addition, at the end of the output file, you must indicate: d) the total number of moves it took to solve all the puzzles of the input file 2- A visual trace: In addition to the puzzle solutions, for each puzzle, your program must also display on the screen (or in another output file) the configuration of the candy box after each move.