Skip to content
learn
  1. Learn
  2. Artificial Intelligence
  3. Intelligent Agents and Problem Solving by Search

AI-M01

Intelligent Agents and Problem Solving by Search

How an AI decides what to do: agents, state spaces, and the search algorithms behind route planners, game engines and solvers.

8 lessonsabout 105 min

Before any learning, an AI needs a way to decide. This module frames intelligence as an agent choosing actions in an environment, then shows how many problems, from sliding puzzles to road trips to Sudoku, become a search for a path through a space of states.

You will watch breadth-first and depth-first search spread across a grid, see how A* uses a good guess to explore a fraction of the map, escape local peaks with simulated annealing, play tic-tac-toe against a minimax agent, follow the Monte Carlo tree search behind AlphaGo, and solve puzzles by constraint propagation. These algorithms still run inside maps, games, schedulers and the planning loops of modern AI agents.

Lessons

  1. AI-01.01 Agents and environments An agent perceives its environment through sensors and acts on it through actuators. Beginner 10 min
  2. AI-01.02 Problem formulation and state spaces Many problems can be framed as search: define the states, the actions available in each state, what each action leads to, a goal test and a cost for each step. Beginner 10 min
  3. AI-01.03 Uninformed search: BFS, DFS, uniform cost Uninformed search explores without knowing where the goal is. Beginner 15 min
  4. AI-01.04 Informed search and A* Informed search uses a heuristic, an estimate of the remaining cost to the goal. Intermediate 16 min
  5. AI-01.05 Local search and optimization Local search keeps a single current solution and improves it step by step, which suits huge problems where only the final answer matters. Intermediate 12 min
  6. AI-01.06 Adversarial search: minimax and alpha-beta In two-player games, minimax assumes both players play optimally: you pick the move that maximizes your outcome assuming the opponent then minimizes it, searching a game tree. Intermediate 15 min
  7. AI-01.07 Monte Carlo Tree Search and AlphaGo Monte Carlo Tree Search builds a search tree selectively: it selects a promising path, expands a new node, simulates a game to the end and backs up the result, balancing exploration and exploitation with the UCT formula. Advanced 14 min
  8. AI-01.08 Constraint satisfaction problems A constraint satisfaction problem assigns values to variables so that all constraints hold, as in Sudoku, map coloring or timetabling. Intermediate 12 min

Start with lesson 1