Page 90 - DCAP506_ARTIFICIAL_INTELLIGENCE

Artificial Intelligence

Notes 1. Reason forward from the initial states: Being formed the root of the search tree. General
the next level of the tree by locating all the rules whose left sides go with the root node,
and use their right sides to produce the siblings. Repeat the process until a configuration
that goes with the goal state is produced.
2. Reason forward from the goal state(s): Start building a search tree starting with the goal
configuration(s) at the root. Produce the next level of the tree by locating all the rules
whose right sides go with the root node. Use the left sides of the rules to produce the new
nodes. Persist until a node that goes with the begin state is produced.

Did u know? The method of chaining backward from the preferred final state is known as
goal directed reasoning or back tracing.

Assortment of forward reasoning or backward reasoning is based on which direction provides
less branching factor and justifies its reasoning process to the consumer. Most of the search
methods can be utilized to search either forward or backward.

Notes One exception is the means-ends analysis technique which continues by decreasing
differences among current and goal states, at times reasoning forward and at times
backward.

The following are the factors which find out the option of direction for a specific problem:

1. Are there more possible start states on goal states? We would like to shift from the smaller
set of states to the bigger set of states.
2. In which direction is the branching factor (that is, their average number of nodes that can
be reached directly from a single node) greater ? we would like to carry on in the direction
with the lower branching factor.

3. Will the program be inquired to validate its reasoning process to a user? If so, it is imperative
to continue in the direction that matches more closely with the way the user will think.
4. What type of event is going to generate a problem-solving episode? If it is the arrival of a
new factor, forward reasoning makes sends. If it is a query to which a response is preferred,
backward reasoning is more normal.

Self Assessment

Fill in the blanks:
11. A search procedure must locate a path among initial and ............................ states.
12. The method of chaining backward from the preferred final state is known as ...........................

6.4 Matching and Control Knowledge

6.4.1 Matching

How to take out from the whole collection of rules that can be applied at a specified point?

84 LOVELY PROFESSIONAL UNIVERSITY