AI: Search, Games & Reasoning Mind Map

Every AI: search, games & reasoning visualization on this site on one map. The branches are the main ideas. Each leaf is a page; the grey line under it is the one thing to remember about it. Click a branch to focus on it, click a page to open it.

Recall mode hides the page names. Read the hint, say the name, then click the "?" to check.

1. How the map is organized

The map at a glance: 7 branches, 12 pages.

Ask what the program is looking for:

A move against an opponent? Game search: minimax looks at every reply, MCTS samples them.
Values that obey rules? Constraint satisfaction: backtrack and prune.
A good solution, not proven best? Local search: change one candidate and keep the better one, with tricks to escape local optima.
A plan with rewards, or the truth behind noisy evidence? Sequential decisions for rewards, probabilistic reasoning for evidence.

2. The branches side by side

Branch	Searches over	Gives	Optimal?
Game search	game tree	best move	minimax: yes; MCTS: in the limit
Constraint satisfaction	partial assignments	a valid assignment	finds one if it exists
Local search	complete solutions	a good solution	no guarantee
Sequential decisions	states and actions	a policy	yes, after convergence
Probabilistic reasoning	hidden variables	probabilities or best path	exact here

3. Five facts to remember

Alpha-beta gives the same answer as minimax; good move ordering lets it search about twice as deep.
MCTS needs no evaluation function: random playouts estimate how good a node is.
Hill climbing stops at local optima; random restarts and annealing escape them.
Value iteration needs the transition model; Q-learning learns from experience without it.
Viterbi finds the single most likely hidden sequence in O(T·N²) by dynamic programming.

See all AI: search, games & reasoning pages in the index, or the other mind maps.