CLASSIFICATION OF SEARCHING

Document Sample
CLASSIFICATION OF SEARCHING Powered By Docstoc
					CLASSIFICATION OF SEARCHING

Searching concept had been developed and many emerging methods. Can generally be classified
as follows:

                  Class                                       Nama

     Any path-Uninformed               Depth-First

                                       Breadth-First

     Any path – Informed               Best-First

     Optimal - Uninformed              Uniform- Cost

     Optimal – Informed                A*

The simplest class-any path uninformed algorithm. Included in this class is the depth-first
algorithm and breadth-first. The second algorithm will find all the nodes in the search tree in a
certain order (regardless of / is independent of goal) and stop when it finds the path to find the
goal state first.
The next class is any informed-path algorithm. The main key of this algorithm is to perform
certain calculations that determine the selection of paths in order to obtain basic goal state faster.
Next is the class uninformed algorithms, optimal. This method ensures finding the best path from
the calculated amount of weight to the path but does not use any information in the graph.
Last class is informed algorithms, which guarantee the optimal path is obtained by using the best
available heuristic information in the graph. This method will get a faster track than the
uninformed method.
Searching strategy will use a base of common searching algorithms. The basic idea is to keep the
composition of (Q) of the node (as a partial path), then take one node of Q, if it reaches a goal
then it will stop otherwise would connect the path to the next node and add to the Q. As a note if
it reaches a state that has ever visited it no longer needs to be added to the Q and Q Enter again
to more than one occasion. This is to avoid looping or cycle even though the nodes are
connected.


Of the general algorithm, the emerging problems or questions. Which node will be chosen and
the extent to which the election is precisely determine the path to goal state. The selection of the
next node will make the search strategy. As an example of deep first search will use a selection
strategy to always look for the deepest node of the search tree. This algorithm will take a
descendant as the next node. While other algorithms breadth first search will retrieve nodes with
the same level in the search tree. The point is that the search method uses a particular rule in a
decision to be taken the next node.
Furthermore, to study the process of searching, introduced two new terminology that is visited
node and expanded nodes. Node is visited if a state or if the node is called a path visited state is
reached and added to the Q. So if there is a state located in the position where Q is the state
visited. An expanded state M is said to be developed if a path from Q after being in that state.
These conditions made the descendant of M is visited and the line is added to Q.
Above simple searching algorithm uses the concept Visited List (a list of previously visited
state). The list will record all the nodes are added to Q. This recording is intended to be able to
know whether the node to be visited have been visited. By knowing the status of this looping can
be avoided.
The process avoids using the visited list looping means to avoid the expansion of state more than
once. The basic concept of the idea was not to conduct the search process on the path toward the
goal to more than one occasion. With this basic concept will save jobs and save a lot of processes
in the search path to goal state.
Concept implementation visited list to consider the problem of computing speed. For cases with
a large size would make the burden of providing a space to store information visited list. One
solution is to simply give an indication of status (flags) to determine whether a particular state
have been visited.
Another concept which can be developed is the use of heuristic value. Heuristic means the
instructions. Is something that can petrify even though there is no guarantee to be able to work.
Heuristic function is to provide assistance in directing the search, but no guarantee can reduce
the shortest path. However, this concept can be obtained by the acceleration in achieving the goal
state.
If you can estimate the distance from current node to the goal state it will simplify or speed up
the search process. However this is not easy. Best First algorithm (or greedy) is a heuristic
method (informed) that uses the value of a heuristic function which is defined to help the search.
This function does not guarantee reaching a goal state, but can guide the search to obtain a goal
state with a faster time.

                         Class                             Nama

                Any path-Uninformed        Depth-First

                                           Breadth-First

                Any path – Informed        Best-First

                Optimal - Uninformed       Uniform- Cost

                Optimal – Informed         A*

				
DOCUMENT INFO
Shared By:
Stats:
views:23
posted:10/8/2012
language:English
pages:3
Description: The simplest class-any path uninformed algorithm. Included in this class is the depth-first algorithm and breadth-first. The second algorithm will find all the nodes in the search tree in a certain order (regardless of / is independent of goal) and stop when it finds the path to find the goal state first.