This project is dedicated to solving the Sokoban puzzle, represented as a grid containing various elements. The puzzle is loaded from an external file, and the state of the puzzle is constructed based on its configuration.
- Walls (X): Impassable obstacles.
- Goals (.): Target locations where diamonds should be placed.
- Diamonds ($): The items that need to be pushed onto the goals.
- Player (@): Represents the user-controlled entity.
- Completed Diamonds (*): Indicates a diamond successfully placed on a goal.
The aim is to push all diamonds onto goals. For a successful solution:
- Every goal should be occupied by a diamond.
- Diamonds must be on distinct goal coordinates.
The core of this project is the planning process to find a sequence of actions transitioning from the initial state to the goal state. We utilize a Breadth-First Search (BFS) algorithm to explore potential states, guaranteeing the shortest path to the goal.
Each action during BFS exploration is logged. This includes the player's movements and if a diamond is being pushed. Possible moves are represented by:
- 'u': Up
- 'd': Down
- 'l': Left
- 'r': Right
When a diamond is pushed, the move is appended with a 'p'. For example, a solution might look like "u", "lp", "r", "d", "l", "up", indicating both player moves and diamond pushes.
Once the solution is found, it's printed on the console, providing a sequence of actions to solve the Sokoban puzzle. This serves as a guide for our robot's behaviors.