Demo 01

Symbolic AI And Search: Can machines act intelligent by searching?

Early AI treated intelligence as finding answers in spaces with explicit rules. This chapter shows why search is powerful and why it encounters combinatorial explosion as the state space grows.

Teaching interaction

Search Tree Walkthrough

Why did early AI rely on search, and why did it encounter combinatorial explosion?

Current Expanded Frontier Final path
Frontier size 3
Expanded total 1
Frontier peak 3

Expansion order: Start

Frontier order: A → B → C

Combinatorial growth: With branching factor b and depth d, a uniform search layer grows on the order of b^d. b=3, d=2, b^d = 3^2 = 9.

BFS expands every possibility level by level

Breadth-first search does not guess a direction immediately. It expands one level at a time, which is reliable but quickly creates many states.

Step 1 / 8
Learning goals
  • Understand how search can systematically find solutions in a space with explicit rules.
  • Compare BFS, DFS, and A* by expansion order, frontier size, and path cost.
  • See why combinatorial explosion appears as the state space grows.
Simplification note

This demo runs deterministic BFS, DFS, and A* on a fixed small tree. Every edge costs 1, and the heuristic values are teaching estimates.

Observation guide

Search excels with explicit rules but fears state growth

The earlier problem was that machines still lacked an executable method for problem solving. Search addressed path finding and planning, but it did not solve knowledge representation, heuristic quality, or explosive scale. Its follow-on influence includes planning, game trees, A*, and search ideas in reinforcement learning.

References