Algorithm idea

In large search spaces, especially those used in natural language processing, robotics, or game playing, it is often impractical to perform exhaustive search (like BFS or A*) due to time and memory constraints.

Beam Search was introduced as a memory-efficient approximation of Best-First Search. It sacrifices completeness and optimality in exchange for speed and reduced resource usage, making it suitable for real-time systems or applications requiring fast decision.

Unlike using heuristic function with additional g(n) denotes the actual cost from the start node to n node as A∗ search, Beam Search is a varation of Best-First Search keeping the top k most promising nodes at each level of the search tree, where k is called the beam width

Guide

Drag the START(😊) and TARGET(🏠) icons to change their positions, and click on the blank nodes to add Walls.

Start Node
Target Node
Wall
Visited
Shortest Path