Misurata University

Faculty of Information Technology

Cs411 -Artificial Intelligence

LECTURE 3
Solving Problems by Searching

Instructor : Yousef Aburawi

Spring 2022/2023
 Outline
1-2

 Problem-solving agents.
 Search problems and solutions.
 Example problems.
 Formulating problems.
 Basic search algorithms.
 Problem-Solving Agents
1-3

 When the correct action to take is not immediately obvious, an agent may
need to plan ahead: to consider a sequence of actions that form a path to a
goal state.
 Such an agent is called a problem-solving agent, and the computational
process it undertakes is called search.
 In this chapter, we consider only the simplest environments: episodic, single
agent, fully observable, deterministic, static, discrete, and known.
 Describing one kind of goal-based agent called a problem-solving agent.
 Problem-Solving Agents (2)
1-4

 Four-phase problem-solving process:

 Goal formulation: What are the successful world states.
 Problem formulation: What actions and states to reach the goal.
 Search: Determine the possible sequence of actions that lead to the states of
known values and then choosing the best sequence that reaches the goal.
 Execution: Give the solution perform the actions.
 The state space can be represented as a graph in which the vertices are states
and the directed edges between them are actions, (as shown next slide).
 e.g. Romania
1-5
 e.g. Romania (2)
1-6

 On holiday in Romania; currently in Arad.

 Flight leaves tomorrow from Bucharest.
 Formulate goal:
 Be in Bucharest.
 Formulate problem:
 States: various cities.
 Actions: drive between cities.
 Find solution:
 Sequence of cities; e.g. Arad, Sibiu, Fagaras, Bucharest, … .
 Search Problems and Solutions
1-7

 A search problem can be defined formally as follows:

 State space: A set of possible states that the environment can be in.
 The initial state that the agent starts in.
 A set of one or more goal states.
 The actions available to the agent.
 A transition model, which describes what each action does.
 An action cost function.
 A sequence of actions forms a path.
 A solution is a path from the initial state to a goal state.
 An optimal solution has the lowest path cost among all Solutions.
 Example Problems
1-8

 A standardized problem is intended to illustrate or exercise various

problem-solving methods.
 It can be given a concise, exact description and hence is suitable as a
benchmark for researchers to compare the performance of algorithms.
 A real-world problem, such as robot navigation, is one whose solutions
people actually use, and whose formulation is idiosyncratic, not
standardized, because.
 for example, each robot has different sensors that produce different data.
 Problem Formulation
1-9

 A problem is defined by:

 An initial state, e.g. Arad.
 Successor function: S(X)= set of action-state pairs.
 e.g. S(Arad)={<Arad  Zerind, Zerind>,…}.
 initial state + successor function = state space.
 Goal test, can be
 Explicit : e.g. x=‘at Bucharest’.
 Implicit : e.g. checkmate(x).
 Path cost (additive).
 e.g. sum of distances, number of actions executed, …
 c(x,a,y) is the step cost, assumed to be >= 0.
 Problem Formulation (2)
1-10

 A solution is a sequence of actions from initial to goal state.

 Optimal solution has the lowest path cost.
e.g. Vacuum World
 States??
 8 states: 2 locations with or without dirt.
 Initial state??
 Any state can be initial.
 Actions??
 {Left, Right, Suck}.
 Goal test??
 Check whether squares are clean.
 Path cost??
 Number of actions to reach goal.

11
e.g. 8-Puzzle
 States??
 The location of each tile.
 Initial state??
 Any state can be initial.
 Actions??
 {Left, Right, Up, Down}.
 Goal test??
 Check whether goal configuration is reached.
 Path cost??
 Number of actions to reach goal.
12
 Search Algorithms
1-13

 A search algorithm takes a search problem as input and returns a solution, or an

indication of failure.
 We will study algorithms that superimpose a search tree over the state-space graph,
forming various paths from the initial state, trying to find a path that reaches a goal
state.
 Each node in the search tree corresponds to a state in the state space and the edges
in the search tree correspond to actions.
 The root of the tree corresponds to the initial state of the problem.
State Space Vs. Search Tree

 A state is a (representation of) a physical configuration.

 The state space describes the (possibly infinite) set of states in the world, and
the actions that allow transitions from one state to another.
 The search tree describes paths between these states, reaching towards the
goal.
 The search tree may have multiple paths to (and thus multiple nodes for) any given
state, but each node in the tree has a unique path back to the root (as in all trees).

14
Search Tree

 A node is a data structure belong to a search tree.

 A node has a parent, children, … , and includes path cost, depth, …
 Here node= <state, parent-node, action, path-cost, depth>
 FRINGE = contains generated nodes which are not yet expanded.
 White nodes with black outline.

15
 Search Tree (2)
1-16

 The root node of the search tree is at the

initial state, Arad.
 We can expand the node, by considering
the available ACTIONS for that state,
using the RESULT function to see where
those actions lead to, and generating a
new node (called a child node or
successor node) for each of the resulting
states.
 Each child node has Arad as its parent
node.
 e.g. Romania
1-17
 Graph Tree
1-18
 Uninformed Search Strategies
1-19

 (blind search) = use only information available in problem definition.

 When strategies can determine whether one non-goal state is better than another
informed search.
 Categories defined by expansion algorithm:
 Breadth-first search.
 Uniform-cost search.
 Depth-first search.
 Depth-limited search.
 Iterative deepening search.
 Bidirectional search.
Breadth-First Search

 When all actions have the same cost, an appropriate strategy is breadth-first
search.
 The root node is expanded first, then all the successors of the root node are
expanded next, then their successors, and so on.
 Also called best-First-Search.
 Implementation: fringe is a FIFO queue.

20
Breadth-First Search on a Simple Binary Tree

21
 Uniform-Cost Search
1-23

 When actions have different costs, an obvious choice is to use best-first search where
the evaluation function is the cost of the path from the root to the current node.
 This is called Dijkstra’s algorithm by the theoretical computer science community,
and uniform-cost search by the AI community.
 Extension of BF-search: Expand node with lowest path cost.
 Implementation: fringe = queue ordered by path cost.
 UC-search is the same as BF-search when all step-costs are equal.
Depth-First Search

 Expands the deepest node in the frontier first.

 Implementation: fringe is a LIFO queue (=stack).
 Depth-first search is not cost-optimal; it returns the first solution it finds, even
if it is not cheapest.

24
Depth-First Search on a Simple Binary Tree

 A dozen steps (left to

right, top to bottom) in
the progress of a depth-
first search on a binary
tree from start state A to
goal M.

25
 Depth-Limited Search
1-26

 Depth-Limited search is Depth-First search with depth limit .

 i.e. nodes at depth have no successors.
 Problem knowledge can be used.
 Solves the infinite-path problem.
 If then incompleteness results.
 If then not optimal.
 Time complexity:
 Space complexity:
 Iterative Deepening Search
1-27

 A general strategy to find best depth limit l.

 Goals is found at depth d, the depth of the shallowest goal-node.
 Iterative deepening search solves the problem of picking a good value for by trying
all values: first 0, then 1, then 2, and so on—until either a solution is found, or the depth-
limited search returns the failure value.
 Combines benefits of Depth-First and Breadth-First search.
Iterative Deepening Search on a Simple Binary Tree

28
Iterative Deepening Search on a Simple Binary Tree (2)

29
 Bidirectional search

 bidirectional search simultaneously searches forward from the initial state and
backwards from the goal state(s), hoping that the two searches will meet.
 The motivation is:
 Check whether the node belongs to the other fringe before expansion.
 Space complexity is the most significant weakness.
 Complete and optimal if both searches are BF.
 The predecessor of each node should be efficiently computable.
 When actions are easily reversible.

30
 Summary
1-31

 Search algorithms introduced that an agent can use to select action sequences in a
wide variety of environments - as long as they are episodic, single-agent, fully
observable, deterministic, static, discrete, and completely known.
 There are tradeoffs to be made between the amount of time the search takes, the
amount of memory available, and the quality of the solution.
 Before an agent can start searching, a well-defined problem must be formulated.
 A problem consists of five parts: the initial state, a set of actions, a transition model
describing the results of those actions, a set of goal states, and an action cost
function.
 Summary (2)
1-32

 The environment of the problem is represented by a state space graph. A path

through the state space (a sequence of actions) from the initial state to a goal state is
a solution.
 Search algorithms generally treat states and actions as atomic, without any internal
structure.
 Search algorithms are judged on the basis of completeness, cost optimality, time
complexity, and space complexity.
 Uninformed search methods have access only to the problem definition. Algorithms
build a search tree in an attempt to find a solution. Algorithms differ based on which
node they expand first:
 Summary (3)
1-33

 BEST-FIRST SEARCH selects nodes for expansion using to an evaluation function.

 BREADTH-FIRST SEARCH expands the shallowest nodes first; it is complete, optimal for
unit action costs, but has exponential space complexity.
 UNIFORM-COST SEARCH expands the node with lowest path cost, , and is optimal for
general action costs.
 DEPTH-FIRST SEARCH expands the deepest unexpanded node first. It is neither complete
nor optimal, but has linear space complexity. Depth-limited search adds a depth bound.
 ITERATIVE DEEPENING SEARCH calls depth-first search with increasing depth limits until a
goal is found. It is complete when full cycle checking is done, optimal for unit action costs,
has time complexity comparable to breadth-first search, and has linear space complexity.
 Summary (4)
1-34

 BIDIRECTIONAL SEARCH expands two frontiers, one around the initial state and one
around the goal, stopping when the two frontiers meet.
 Readings
1-35

 Chapters 3 of Textbox.
The End