In Artificial Intelligence (AI), search refers to the process of exploring possible

solutions to a problem to find the optimal or acceptable one. It is one of the

fundamental concepts in AI and plays a critical role in problem-solving, decision-
making, and reasoning. Search algorithms are used to explore the state space of a
problem, which represents all possible configurations or states that the problem can
have.
Key Search Concepts in AI:
State Space:
1. The state space is a representation of all possible states or configurations of a
problem. Each state represents a possible situation or configuration in the problem
domain.
2. Initial state: The starting point or configuration.
3. Goal state: The desired end configuration.
4. Actions: The set of operations or moves that transition the agent from one state to
another.
Search Tree:
1. A search tree is a graphical representation of the state space, where each node
represents a state, and edges represent actions or transitions between states.
2. The root node represents the initial state, and leaf nodes represent possible end
states.
Search Algorithms: Search algorithms are categorized based on how they
explore the state space:
Uninformed Search (Blind Search): These algorithms search the
state space without any domain-specific knowledge. They
systematically explore the state space without guidance on which
direction is likely to be more fruitful.
1. Examples:
1. Breadth-First Search (BFS): Explores all nodes at the current depth
before moving to the next depth level.
2. Depth-First Search (DFS): Explores as far as possible along a
branch before backtracking.
3. Uniform Cost Search: Expands the node with the lowest cost path
first.
Informed Search (Heuristic Search): These algorithms use domain-
specific knowledge or heuristics to make the search process more
efficient. A heuristic is a rule of thumb or approximation that helps
guide the search.
2. Examples:
1. Greedy Best-First Search: Uses a heuristic to choose the node that
appears to be closest to the goal.
2. A*: Combines cost to reach a node and a heuristic estimate of the
cost to reach the goal from that node.
Heuristic Function:
1. A heuristic is a function that estimates the cost of reaching the goal from the
current state. Heuristics are used to guide search algorithms in informed searches.
2. The heuristic function is typically denoted as h(n), where n is a node in the search
tree.
 3. Admissible heuristic: A heuristic that never overestimates the cost to reach the goal
(i.e., it is optimistic).
Optimal and Completeness:
1. A complete algorithm guarantees finding a solution if one exists.
2. An optimal algorithm guarantees finding the best possible solution, usually the one
with the lowest cost or shortest path.
Search Space:
1. The search space refers to the collection of all possible states or configurations that
could be considered during the search process. It is often large and requires efficient
algorithms to navigate.
Exploration and Exploitation:
1. Exploration refers to the agent searching through unvisited areas of the state space,
potentially discovering new solutions.
2. Exploitation involves refining the current known information to optimize the path
or solution based on what is already known.
Search Strategies:
1. Breadth-First Search (BFS): Explores all nodes at the present depth before going
deeper.
2. Depth-First Search (DFS): Explores a branch fully before backtracking.
3. A*: A popular search algorithm that balances exploration and exploitation, using
both cost and heuristics.
4. Iterative Deepening Depth-First Search: Combines depth-limited search and
breadth-first search, ideal when the solution depth is not known.
Types of Search in AI:
Uninformed Search (Blind Search):
1.
1. Breadth-First Search (BFS): Explores all possible states level by level. Guaranteed to
find the shortest solution in terms of the number of actions.
2. Depth-First Search (DFS): Explores as deep as possible into the state space before
backtracking. Can get stuck in loops or deep branches if not implemented carefully.
3. Uniform Cost Search: Expands the node with the lowest total cost so far.
Informed Search (Heuristic Search):
1. Greedy Best-First Search: Chooses the node that appears closest to the goal based
on a heuristic function. It is fast but does not guarantee optimal solutions.
2. A*: Uses both the cost to reach a node (g(n)) and an estimated cost to reach the
goal (h(n)). It is complete and optimal when using an admissible heuristic.
Local Search:
1. In local search, the algorithm operates within a smaller region of the search space
and iteratively improves the solution.
2. Hill Climbing: A greedy algorithm that moves towards the direction of increasing
value but can get stuck in local optima.
3. Simulated Annealing: A probabilistic technique that allows for exploration by
accepting worse solutions with some probability to avoid local optima.
2.
Adversarial Search:
1. Used in scenarios where multiple agents or players compete, and the outcome
depends on the strategies of both the agent and its opponent.
 2. Minimax Algorithm: Used in two-player games like chess, where one player is
maximizing the score and the other is minimizing it.
3. Alpha-Beta Pruning: An optimization of the minimax algorithm that reduces the
number of nodes evaluated by pruning branches that will not affect the final
decision.
Evolutionary Search:
1. Inspired by biological evolution, these search techniques use processes like
mutation, selection, and crossover to explore the search space.
2. Genetic Algorithms: Search for solutions by simulating the process of natural
selection.
Examples of Search Problems:
 Pathfinding: Finding the shortest path in a maze or a network (e.g., Google Maps routing).
 Game Playing: AI in games like chess, where the agent searches for the best move.
 Puzzle Solving: Solving puzzles like the 8-puzzle or Rubik's Cube using search techniques.
 Optimization Problems: Finding optimal solutions in scheduling, resource allocation, etc.
Summary:
 Search in AI is a critical concept where an agent explores possible states in the problem
domain to find a solution.
 Search algorithms can be categorized into uninformed (blind) and informed (heuristic-based)
approaches, depending on whether they use additional information to guide the search.
 Heuristics play an important role in making search efficient, especially when dealing with
large state spaces.
 Different search strategies are suited to different problem domains, from path finding to
optimization and game playing.

Problem formulation in AI
Problem formulation in AI refers to the process of defining a real-world problem in
such a way that it can be solved by an AI system using algorithms and techniques like
search, reasoning, or optimization. It involves identifying key components of the
problem, such as the initial state, goal state, possible actions, and constraints.
Effective problem formulation is crucial because it transforms an abstract or vague
issue into a structured form that AI techniques can handle.
Components of Problem Formulation in AI:
Initial State:
1. The state from which the agent starts or the starting point of the problem.
2. It represents the situation or configuration at the beginning.
3. Example: In a pathfinding problem, the initial state is the starting location.
Goal State:
1. The desired state or outcome the agent is trying to reach.
2. The goal state represents the solution to the problem.
3. Example: In a puzzle-solving problem, the goal is to reach the solved configuration.
State Space:
1. A collection of all possible states that the problem can be in, from the initial state to
the goal state.
2. It defines the scope of the problem and can be finite or infinite depending on the
complexity.
3. Example: In chess, the state space is all possible board configurations that can occur
during a game.
Actions:
 1. The possible moves or transitions that the agent can take to go from one state to
another.
2. Actions depend on the current state and define the steps the agent can perform to
progress toward the goal.
3. Example: In a robot navigation problem, actions could be moving forward, turning
left, or turning right.
Transition Model:
1. A description of how the agent's actions cause changes in the state.
2. The transition model defines the relationship between the current state, the action
taken, and the resulting state.
3. Example: In a game, moving a piece changes the game state based on predefined
rules.
Path Cost:
1. A numeric value that represents the cost of moving from one state to another.
2. Path cost is used to evaluate the efficiency of a solution, with lower-cost paths being
more desirable.
3. Example: In a GPS routing system, the cost could be the distance traveled, time
taken, or fuel consumed.
Solution:
1. A solution to a problem is a sequence of actions (or a path) that leads from the
initial state to the goal state.
2. It should meet the criteria defined by the problem and be optimal or satisfactory
based on the problem's constraints.
Constraints:
1. Rules or limitations that restrict the available actions or the form of the solution.
2. Constraints can include things like time limits, resource availability, or specific rules
that must be followed.
3. Example: In a scheduling problem, a constraint might be that a particular task
cannot start until another task is finished.
Example of Problem Formulation:
Consider a simple pathfinding problem where an AI agent (like a robot or
navigation system) must find the shortest path from point A to point B on a grid.
1. Initial State: The agent starts at point A (e.g., a specific grid location).
2. Goal State: The agent needs to reach point B (another grid location).
3. State Space: All possible positions on the grid that the agent could occupy.
4. Actions: Move up, down, left, or right on the grid.
5. Transition Model: Moving in one direction takes the agent from one grid cell to another.
6. Path Cost: Each movement could have an associated cost (e.g., moving to a cell might cost 1,
or certain cells might have a higher cost due to obstacles).
7. Constraints: The agent cannot move through obstacles, and it might need to find the
shortest or least costly path.
Importance of Problem Formulation:
1. Efficient Solutions: A well-defined problem allows AI systems to apply search algorithms and
optimization techniques effectively, leading to efficient solutions.
2. Correctness: A clear problem formulation ensures that the AI agent correctly understands
the objective and solves the problem in the right way.
3. Scalability: Proper problem formulation helps AI systems scale to handle larger or more
complex versions of the problem.
 4. Search Space Reduction: By precisely defining the problem, unnecessary or irrelevant states
and actions can be excluded, reducing the complexity of the search space.
Common Types of Problems in AI:
Search Problems:
1. Problems where the agent needs to find a sequence of actions that lead from the
initial state to the goal state.
2. Example: Pathfinding, game playing.
Constraint Satisfaction Problems (CSPs):
1. Problems where the solution must satisfy a set of constraints or rules.
2. Example: Sudoku puzzles, scheduling tasks.
Optimization Problems:
1. Problems where the goal is to find the best solution, often in terms of minimizing or
maximizing some criteria.
2. Example: Traveling Salesman Problem (TSP), where the goal is to minimize the
distance traveled.
Planning Problems:
1. Problems where the agent must create a plan or strategy to achieve its goal, taking
into account future states and consequences.
2. Example: Task scheduling, robot motion planning.
Summary:
Problem formulation is a critical step in AI that involves defining the components of a
problem — including initial state, goal state, actions, state space, and constraints — in
a way that allows AI techniques to solve it effectively. A good problem formulation
leads to efficient, scalable, and accurate solutions by focusing the AI agent's search
and reasoning processes on the relevant parts of the problem.