ARTIFICIAL INTELLIGENCE
Uploaded by
subrata baruiARTIFICIAL INTELLIGENCE
Uploaded by
subrata baruiARTIFICIAL INTELLIGENCE
Title – Comparative Study Among Different Search Techniques
Used In Artificial Intelligence.
Abstract :- Solving a the basis of their merits and
problem mean looking for a demerits to make it easy to
solution, which is best among choose a technique for a
others. Finding a solution to a particular problem.
problem in Computer Science
Introduction :- Search
and Artificial Intelligence is algorithms are used for a
often thought as a process of multitude of AI tasks,one of
search through the space of them being the path finding. The
possible solutions. On the other area of search in AI is very much
hand in Engineering and connected to real life problem
Mathematics it is thought as a solving. AI has investigated
process of optimization i.e. to search methods that allow one
find a best solution or an to solve path planning problems
in large domains. Having
optimal solution for a problem.
formulated problems, we need to
These reduce search space and
solve them and it is done by
improve its efficiency. At each searching through the state
and every step of search,it space during this process. Most
select which have the least of the researches on search
futility. In this paper ,We methods have studied how to
categorize the different AI solve one-shot path- planning
search and optimization problems. Search is mostly a
techniques in a tabular form on repetitive process, therefore,
many AI systems re-plan from problem space as a whole or to
scratch to solve the path only some states. It may be
planning problem independently. available a priori or only after a
II. Problem Statement :- node has been expanded. On
There are many search and such basis we can categorize
optimization algorithms in
different search techniques as
Artificial Intelligence, the popular
shown in Fig1
ones being Uninformed Search,
Heuristic Search and
Evolutionary algorithms etc.
Although a lot of research work
is done on individual algorithm
but not enough research is done
on the comparison of these
algorithms under different Uninformed search
problems. This is essential algorithms for problem
considering the fact that these solving are a central topic
algorithms behave differently or of classical computer
perform differently for different science by Horowitz and
problems. By analyzing how an Sahni 1978,operations
algorithm performs under a research by Drefus by
certain problem, the 1969.Uninformed search
shortcomings of the algorithm strategies use only that
can be found out and problem by information which is
finding out the pros and cons of available in the problem
the tested algorithm. definition.
Iv. Comparison Of
III. Ai Solution Search Different Search
Techniques & Algorithms :- Algorithms:-
Search problems can be The output of problem
classified by the amount of solving algorithm is either
information that is available to failure or solution. Some
the search process. Such algorithms might get stuck
information might relate to the in an infinite loop and
never return an output. The comparison among
When determining which different search algorithms
search algorithm is by these factors is shown
appropriate for a problem by the table 1:
space, it is necessary to
derive and compare
general attributes of each
algorithm .We will
evaluate algorithm’s
performance in four ways.
A. Completeness: - Is that
algorithm guaranteed to Where
find a solution when there
is one? This column is a d =depth of solution with
Boolean indicator of in search tree
whether or not the search b = branching factor of
algorithm is exhaustive. search tree
n =subset of b for which
B. Optimality: - Does that algorithm will actually
strategy find the optimal process.
solution? This column In table 1, the attributes
indicates that whether or create a basis for decision
not the solution found will making. Each of the
always be the optimal algorithms discussed
solution. contains
weak and strong
D. Space Complexity: - attributes.
How much memory is
V. Conclusion:-It is the
needed to perform the
search? This column is the function of the problem
order of complexity space to weight the trade-
memory requirements of offs between the
algorithm also expressed algorithms and determines
as a function. which algorithm provides
the best solution . It can
be seen from the table the search algorithms are
that the time estimate more distributed than the
from all the searches are time estimates. In many
similar. The three cases a search algorithm
exceptions are the will approach a problem
Bidirectional, Beam and Breadth First or Depth
Generate and Test first.
searches. The main reason
Vi. References :-
that the Bidirectional
search has a lesser time 1. Artificial Intelligence, K,
estimate is because it is Knight, E. Rich, S.B. Nair,
3rd Edition TMH
simultaneously working
from both ends of the 2. A classical approach to
problem looking for a Artificial Intelligence, M.C.
common intermediate Trivedi, 2nd Edition,
node. The Beam search Khanna Publishing
has a time estimate of House, New Delhi
O(nd) as opposed to the
3. Introduction to Artificial
morecommon O(bd). It is
Intelligence & Expert
because the Beam Search Systems, D.W. Patterson,
is modified A* Search that PHI
examines on the best n 4. Artificial Intelligence A
branches at any node. It Modern Approach, Stuart
speeds up processing, but Russel, Peter Norvig,
at the cost of assuming Pearson
that a suboptimal node 5. Computational
will never need to be Intelligence, D. Poole, Alan
travelled to reach the goal Mackworth, and Randy
Goebe, lOUP
state. If that is the case
the solution to the search 6. Logic & Prolog
will never be found. The Programming, Saroj
memory requirement of
Kaushik, New Age
International
7. Expert Systems
principle and
programming, J.C.
Giarranto, Cengage
Learing.
