GameSrch - BACKTRACKING (Continued) Note that Backtracking...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon
BACKTRACKING (Continued) Note that Backtracking algorithms run in a depth first search fashion. Refer to the call sequence of the bit-string printing algorithm's recursion tree. Sometimes some nodes and their descendants are not explored ( pruned ) because we know that they will not produce a better value than what we have already achieved on a previously explored leaf. Refer to the 0-1 Knapsack algorithm, where a node crosses the Knapsack weight-limit (or, where the upper bound for the profit is lower than the best profit made on a leaf before the current node is explored in the algorithm) is pruned. Game-search Algorithms Computer games playing against adversary (human beings) often deploy some search algorithm: called adversarial search . Tic-tac-toe is an example. [DRAW] The problem is, ( Input :) given a board position, ( Output :) to generate a good move. Objective is to win the game . Board positions are evaluated by an evaluation function for their goodness toward wining the game. Each node in the search tree is a board position . Children-nodes are valid board positions that can be generated at the next step. In its search for the next move a computer has to go all the way down to the terminal board positions (leaf ), and percolate the values (win/draw/loss ) up to the children of its current board (in order to decide which child to move to). However, at every alternate level what the search algorithm looks into is actually to be made by the opponent (human). Thus, it needs an honest adversary simulator. Algorithm bi-recursively calls each other.
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
While computer tries to maximize its evaluation function human-simulator would try to minimize it (assuming values for win >0, draw=0, and loss<0).
Background image of page 2
Image of page 3
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 02/10/2012 for the course CSE 5211 taught by Professor Dmitra during the Spring '12 term at FIT.

Page1 / 8

GameSrch - BACKTRACKING (Continued) Note that Backtracking...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online