Path Finding Problem Zolbayar Magsar Chanmann Lim Yihan Xu 

Introduction • Solve Path Finding problem • Find the routing path on a grid from a given square A to destination B using: - Uninformed search algorithms (Breadth First, Depth First) - Heuristic search algorithms ( Best First, A*) - Local Search algorithms (Hill-climbing, Simulated annealing) • Make use of Manhattan distance heuristics for the informed search and scoring function for local search. 

Motivation Seeing search in action so AI students can understand search better! 

Problems to be studied • Search Algorithms • Visualization 

Existing Work Implemented in NodeJS (Raphaël) 

Our Implementation We implemented in Java (Swing) adding local searches 

Detailed Aspects Algorithm Used • Breadth First Search • Depth First Search • Greedy Best First Search • A* Search • Hill Climbing Search • Simulated Annealing 

Algorithm - Uninformed Search Breadth First Search Depth First Search 

Best First Search A* Search Algorithm - Heuristic Search 

Algorithm - Local Search Hill-Climbing Search Simulated Annealing 

Cooling schedule • Initial Temperature : 1000 • Temperature decrease function: Linear Multiplicative Cooling 

Detailed Aspects Running Performance Execution Time Comparison 0 1 2 3 4 5 Sample 1 2 3 4 5 6 7 8 9 10 Breadth First Depth First Best First A* Hill-Climbing Simulated Annealing 

Detailed Aspects Running Performance Path Cost Comparison 0 100 200 300 400 500 600 700 Sample 1 2 3 4 5 6 7 8 9 10 Breadth First Depth First Best First A* Hill-Climbing Simulated Annealing 

Detailed Aspects Running Performance Expanded Nodes Comparison 0 100 200 300 400 500 600 Sample 1 2 3 4 5 6 7 8 9 10 Breadth First Depth First Best First A* Hill-Climbing Simulated Annealing 

Conclusion • Breadth First Search is guaranteed to find the goal, the Depth First can reach the goal only in finite state space. • Greedy Best First Search finds suboptimal path, and A* Search is guaranteed to find the shortest path if the heuristic is never larger than the true distance. In the implementation, we use Manhattan Heuristic. • Hill-Climbing Search and Simulated Annealing are not even guaranteed to find a solution, have linear time/space advantage. • A* is a good choice for most pathfinding needs. 

Q&A Thank You