You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Overkill Heuristic that destroys both Manhattan and Euclidean
Dijkstra
Genetic Algorithm
DL/RL approaches
Q-Learning
Q-Learning with NN and Experience Replay
Q-Learning with LSTM and Experience Replay
A* with NN heuristic
A* with LSTM heuristic
A* with CNN heuristic
Evaluation Metrics
Time
Number of Nodes Expanded
Number of Nodes in the Path
Number of Nodes in the Solution
Number of Nodes in the Maze
Heuristics
Manhattan
Euclidean
NNHeuristic
Implemented
✅
✅
✅
About
nextdataAI: a Python library for solving mazes using classical algorithms (A*, BFS, DFS, Dijkstra, Greedy), reinforcement learning (QLearning, QTable, QNN, QLSTM), and novel pseudo-heuristics. Features include a neural network-based heuristic approximating Manhattan distance (NNHeuristic) and a custom SManhattan heuristic, a dataset gener