State search 🔍 is a problem-solving method in AI and computer science 💻. It involves exploring a space of possible states to find a sequence of actions required to solve a problem 🤔. State search algorithms can be categorized into uninformed and informed search algorithms 🧐. Examples of state search algorithms include depth-first search 🕳️, breadth-first search 🌳, A* search ⭐, NBA* search 🏀, and others 🤔. Concepts introduced in state search have practical applications in game playing 🎮, robotics 🤖, logistics and supply chain management 🚛, automated planning 📆, and others 🤖💡. Understanding state search algorithms is essential for developing intelligent systems that can solve complex problems 🤖💡.
Search for TODO
text in the repository with CTRL+F and replace it with you code written according to it.
- Clone repository: git clone:
git clone <repository url>
- Complete TODOS the exercises
- Commit your changes
git add <path to the changed files> git commit -m <commit message>
- Push changes to your repository main branch
git push -u origin master
The rest will be taken care of automatically. You can check the GRADE.md
file for your grade / test results. Be aware that it may take some time (up to one hour) till this file
This class requires Python with version at least 3.8. Recommended way to run is to create a virtual environment first:
python -m venv state-search
source state-search/bin/activate
Then install required packages:
pip install -r requirements.txt
Finally you can run a solver:
python solve.py -a <algorithm> -p <problem> -h <heuristic> <path_to_instance>
, e.g.
Where <algorithm>
is one of:
- dfsrecursive
- dfsiter
- bfs
- dijkstra
- greedy
- astar
- nbastar
the <heuristic>
depends on the selected <problem>
and can be one of the following (select one from listed in {}
):
- grid_path_finding: { grid_euclidean, grid_diagonal, grid_manhattan }
- n_puzzle: { n_puzzle_tiles_out_of_place, n_puzzle_manhattan }
- rush_hour: { rush_hour_distance_to_exit, rush_hour_blocking_cars, rush_hour_indirect }
- blocks_world: { blocks_world_naive }
- pancake: { pancake_gap, pancake_largest_pancake }
the <problem>
is one of:
- grid_path_finding
- n_puzzle
- rush_hour
- blocks_world
- pancake
and instance
is one of the files placed in <problem>/instances
directory (see project structure below for example).
Example:
python solve.py -p rush_hour -a astar -h rush_hour_indirect problems/rush_hour/instances/81.txt
You can also run a benchmark:
python benchmark.py -p <problem> -t timeout <path_to_instance>
Example:
python benchmark.py -p rush_hour problems/rush_hour/instances/54.txt
If you run script with incorrect arguments, you will get some helpful info ;)
.
├── base # API for problem and solver classes
├── problems # List of defined problems (place to define problems)
│ ├── ...
│ ├── n_puzzle # directory with a problem
│ │ ├── instances # directory with problem instances
│ │ └── ...
│ └── ...
├── solvers # directory with algorithms
│ ├── generic # code shared by several algorithms
│ ├── bfs.py # example of an algorithm
│ └── ...
├── tree # search tree representation
├── utils # various utilities
├── solve.py # solve tool (run as a script)
├── benchmark.py # benchmark tool (run as a script)
└── cli_config.py # configuration of the cli tools (do not touch)
- interactive pathfinding on grids
- definitions of various grid heuristics
- nice blog post about A*
- elegant n-puzzle visualization
- NBA* - papers explaining the algorithm: 1, 2 — requires access via AGH library/other means,