Tile-Slider

🕵️ Introduction

Ever wondered how computers play games? In this challenge train model to solve the game Tile-Slider, a replica of the game Match the Tiles.

Understand with code! Here is getting started code for you.😄

🕹 Game Description

The Game board is a grid with some obstacles that cannot be passed through. There are some source tiles (coloured squares with holes in the centre) and corresponding destination positions (denoted by the coloured circles). Our target is to move all source tiles to their respective destination positions. Moves allowed are up, down, left, right (by swiping in the original game, by printing out a letter U, L, D, or R in here). When you play one move, say you choose to swipe down, then all source tiles will move as far down as they can i.e will stop only when there is no way down or they have hit some obstacle or another tile (and that tile cannot go any more down too). Note that only the source tiles can move and not the destination tiles.

💾 Dataset

The dataset comprises 2000 text files. Each file contains a grid, source, and destination points. You need to output a set of moves for each test case. Every source should reach its respective destination tile with the output set of moves. Lesser the number of moves required to reach the destination better the score is.
Input format 
First-line contains two integers N, M where N represents the number of rows in the grid and M number of columns in the grid. Following N lines contain M characters each where each character is either. or #  where # signifies obstacles which can not be passed through and . free space. Following that will be an integer S representing the number of source tiles. Then below S lines, each will be of format source_row source_column destination_row destination_column. It is guaranteed that in all test files any source/destination position does not collide with obstacles. Note that 0 based indexing is followed.

Example:

📁 Files

The dataset comprises 2000 files having input format as mentioned above. These are of varying grid sizes.

Following files are available in the resources section:

• dataset.zip - (2000 text files) Each text file has an input grid, sources, and destination points in the format as described above.

🚀 Submission

• Submit a CSV file containing headers [filename, moves].
• Move length corresponding to any file should not be greater than 1000.
• This file should contain solutions for all 2000 puzzles.
• Each row of files will contain a test filename (under filename column) and a string containing a set of moves (under moves column).
• Sample submission format available at sample_submission.csv in the resources section.

🖊 Evaluation Criteria

The score will be calculated as the sum of squares of the number of moves required to solve puzzles (which are solved correctly) and for the puzzles which are not solved by your output set of moves, some large number is added to the score.

🎮 Simulator

We offer a python code to visualize this problem in PyGame. You can use this simulator code to visualize this gameplay for a given input grid and set of moves.
Get the code here. 

📱 Contact

• Animesh Sinha
• Kanish Anand