4/21/2021

The problems for which can not be solved through precise models, methods
and algorithms as well as involve uncertainty are classified as AI problems.
Two major approaches to solve AI problems:
I. One is using Hard-Computing Techniques which are classical AI problem solving
methods based on symbolism and search with an aid of symbolic logic.
II. The other uses soft-computing techniques- which consist of five major
components namely Fuzzy Logic, Neural Networks, Probabilistic Reasoning,
Genetic Algorithm and Chaos Theory.
In this course we focus on the first approach for solving AI problems
This involves the following four important tasks:
1. Define the Problem Precisely: This definition must include the precise
specification of what the initial situation will be as well as what final
situations constitute the acceptable solution to the problem.
2. Analyze the Problem : A few very important features can have immense
impact on the appropriateness of the various possible techniques for
solving the problem.
3. Isolate and Represent the task knowledge that is necessary to solve the
problem
4. Choose the best problem solving technique(s) and apply it(them) to the
particular problem
An Example of Build a Program That Could Play Chess
Before start writing a program that can play chess, we need to specify the
following:
1. The starting position of the chess board
2. The rules that defines the legal moves
3. The board positions that represent a win for one side or other

Goal: The program not only play a legal game of chess, but also winning the game if
possible
 Formal Definition
(Representing Chess Playing Problem from programming point of view)
1. The Starting Position: it can be described by 8x8 array where each position
contains a symbol at appropriate place same as in the official chess opening
board position.
2. The Goal Position: Any board position that the opponent can not make any
legal move or his/her king is under attack.
3. Legal moves: These provide the ways of getting from the starting state to goal
state
Describes the starting position and the goal position can be done very
easily, but the real difficulty lies in how to describe the rules for legal moves.
 One Way of Describing the Legal Moves
This left side serves as a pattern to be matched against the This right sides reflects the change in the board after the
current board position application of the legal moves.
 Difficulties in This Legal Move Representation
If we write rules in this manner, there will be 10120 rules equal to the number
of possible board positions. This creates the following implementation difficulties
I. Nobody can supply the complete set of rules
- It takes too much time ( more than life span of our earth)
- It can not done without mistake or repetition
II. No program can handle this huge number of rules
- It is not possible for searching applicable rules for the current board position
- It is not possible to Store 10120 rules in computer memory
 What is State Space?
Background Knowledge
Definition of a State: A situation in the world described by some facts is a
state.
A state can be instantaneous or it can last for some time (non-instantaneous).
The state changes(transitions) can be either discrete or continuous.
In physics: A physical state is a set of measurable physical parameters that is
fully describing some part of the universe.
In computer science: A process state is a set of values that are currently
stored in registers, buffers and stack. So process state is a snap shot of a
process in time.
4/21/2021
 In Artificial Intelligence: A problem state includes everything important
about a situation in one bundle (that is everything necessary to reason
about it). It usually means a list of facts.. There may be lot of facts and
we often include only those facts that are necessary for reason about in
solving the AI problems.

Two tricks to simplify the state description:

1. Record only facts relevant to the very specific problem.

2. Compressing the facts.

4/21/2021
 State Space Search as Solving the AI Problem
• We can view the solving of an AI problem as moving around the state space that
begins with a starting state and ends with one of the goal state.
• The set of all possible states of a problem constitute problem universe.
• Now the problem of playing chess can be defined as a problem of moving
around in a state space, where each state corresponds to a legal position of the
chess board.
• chess can be played by starting at an starting/initial state, using a set of legal
rules to move from one state to another and attempt to end up in one of the
goal/final state.

4/21/2021
 Structure State Space Representation corresponds to the
Structure of Problem Solving

• State space representation allows the formal definition of a problem as the

need to convert some given situation into some desired situation using a
set of permissible operation

• State space representation also permits us to define the process of

solving a particular problem as a combination of known techniques and
the search.

4/21/2021
 An Illustrative Example
Water Jug Problem: There are two water jugs, one is 4-liter capacity and the
other is 3-liter capacity. Neither has any measuring marks on it. There is a
water pump that can be used to fill the jugs with water. How can we get 2
liters of water in the 4-liter jug?

 How to make a good representation of a state for this problem?

 How to describe a set of state transition rules?

4/21/2021
Representation of the (Water Jug) Problem State
The problem state can be represented as an ordered pairs of integers (a, b)
that represent the quantities of water in 4-liter and 3-liter jugs respectively.
Where,
a represents the quantity of water in 4-liter jug and takes one of the values 0, 1, 2, 3, 4.
b represents the quantity of water in 3-liter jug and takes one of the values 0, 1, 2, 3.

Starting(or Initial) State can be represented by (0,0).

Goal(or End) State is represented by (2,x).

4/21/2021
 State Transition Rules or Production Rules
Before describing the rules/operators, we need to make few assumptions which are not
explicitly mention in the informal problem statement. The unstated assumptions for the water jug
problem are:
a. We can fill water in a jug using the pump.
b. We can pour water out of a jug onto the ground.
c. We can transfer water from one jug to another.
d. There no other measuring device available.
✓ Left side of the rule represent the current state.
✓ Right side of the rule represent the state that reflect the changes in the current state by
application of the rule.
✓ Group of similar branches into categories called operators or production rules. It is important to
note that the pre-condition of the rule has to be satisfied by the current state.

4/21/2021
 Production Rules for Water Jug Problem
Filling the water jug up to its full capacity
1. (a, b) → (4, b) Fill 4-liter jug
If a<4 (Pre-Condition)

2. (a, b) → (a, 3) Fill 3-liter jug

If b<3
Emptying the water jug completely
3. (a, b) → (0, b) Empty the 4-liter jug on the ground
If a>0

4. (a, b) → (a, 0) Empty the 3-liter jug on the ground

If b>0

4/21/2021
 Transferring the water from one jug to another
5. (a, b) → (4, b-(4-a)) Pour water from 3 liter jug into 4 liter jug until it becomes full.
If (a+b) ≥ 4 and b>0

6. (a, b) → (a-(3-b), 3) Pour water from 4 liter jug into 3 liter jug until it becomes full.
If (a+b) ≥ 3 and a>0

7. (a, b) → (a+b, 0) Pour all the water from 3 liter jug into 4 liter jug.
If (a+b) ≤ 4 and b>0

8. (a, b) → (0, a+b) Pour all the water from 4 liter jug into 3 liter jug.
If (a+b) ≤3 and a>0

4/21/2021
 Pour some quantity of water(d) out of the water jug
9. (a, b) → (a-d, b) Pour some quantity of water(d) out of 4-liter jug
If a>0

10. (a, b) → (a, b-d) Pour some quantity of water(d) out of 3-liter jug
If b>0

Special Case Rules That Represent the Problem Solving Knowledge

11. (0, 2) → (2, 0) Transfer 2 liter of water from the 3-liter jug into the 4-liter jug

12. (a, 2) → (0, 2) Empty the 4-liter jug

4/21/2021
 One of the solutions to Water Jug Problem
(0,0) Initial State
 Rule-2
(0,3)
 Rule- 7
(3,0)
 Rule-2
Intermediate States (3,3) State Transition Rules /Operators
 Rule-5 or
(4,2) Production Rules
 Rule-4
(0,2)
 Rule-11
(2,0) Goal State
4/21/2021
Translating Informal Description of an AI Problem
into Formal Description
1. Define the Problem State Space-the problem state contains al possible configuration of
relevant objects
2. Specify one or more states within state space that describe possible situation from which
the problem solving process may start. These states are called initial states
3. Specify one or more stats that would be acceptable as a solution to the problem. These
states are called goal state.
4. Specify a set of rules that describe the action available
➢ What are all the unstated assumptions that are hidden in the informal problem
description.
➢ How general should the rules be made?
➢ How much work required to solve the problem should be pre-computed and
represented in the rules.
 Production System
A production system consists of
1. A Set of Rules each consisting of left side that determines the applicability of the rule
and the right side that describe the operation to be performed, if the rule is applied.
2. One or more Knowledge/Databases that contain whatever information is appropriate
for the particular task.
3. Control Strategy that specifies the order in which the rules will compared to the
databases and way of resolving conflicts that arise several rules match at once.
4. A Rule Applier.
***********************************************
❖ Heuristic is a rule of thumb, a piece of information that can provide useful guidance to
a problem solver, but which is not guaranteed to be applicable or beneficial in all
situations.
❖ Heuristic Search is a search using heuristic knowledge to focus the exploration of the
state space on more promising path.
***********************************************
4/21/2021