A Theory of Heuristic Information in Game-Tree Search by Chun-Hung Tzeng PDF

By Chun-Hung Tzeng

ISBN-10: 3642613683

ISBN-13: 9783642613685

ISBN-10: 3642648126

ISBN-13: 9783642648120

Searching is a crucial method in so much AI platforms, specifically in these AI creation structures which includes a world database, a collection of construction ideas, and a regulate process. as a result of intractability of uninformed seek strategies, using heuristic details is important in so much looking strategies of AI structures. this crucial thought of heuristic informatioD is the crucial subject of this booklet. We first use the 8-puzzle and the sport tic-tac-toe (noughts and crosses) as examples to assist our dialogue. The 8-puzzle comprises 8 numbered movable tiles set in a three x three body. One cellphone of the body is empty in order that it truly is attainable to maneuver an adjoining numbered tile into the empty cellphone. Given tile configurations, preliminary and aim, an 8-puzzle challenge contains altering the preliminary configuration into the target configuration, as illustrated in Fig. 1.1. an answer to this challenge is a chain of strikes best from the preliminary configuration to the objective configuration, and an optimum answer is an answer having the smallest variety of strikes. now not all difficulties have recommendations; for instance, in Fig. 1.1, challenge 1 has many options whereas challenge 2 has no resolution at all.

Show description

Read or Download A Theory of Heuristic Information in Game-Tree Search PDF

Best theory books

Read e-book online Computer Aided Systems Theory – EUROCAST 2007: 11th PDF

The concept that of solid as machine Aided structures thought was once brought by way of F. Pichler within the past due Eighties to surround computer-theoretical and sensible advancements as instruments for problem-solving in process technology. It was once regarded as the 3rd of 3 elements (the different being CAD and CAM) that jointly offer a whole photo of the trail from laptop and structures sciences to useful advancements in technology and engineering.

i-Smooth Analysis: Theory and Applications by Arkadii V. Kim PDF

The version introduces a brand new type of invariant derivatives  and indicates their relationships with different derivatives, resembling the Sobolev generalized spinoff and the generalized by-product of the distribution conception. it is a new path in arithmetic.   i-Smooth research is the department of sensible research that considers the idea and purposes of the invariant derivatives of services and functionals.

Get Group Theory of Chemical Elements: Structure and Properties PDF

During this monograph, group-theoretical ways are used to construct a process of hadrons and qualitatively describe the houses of chemicals. This serves as a supplement to numerically and nearly remedy the many-electron Schrödinger equation, with a view to comprehend the habit of chemical parts.

Extra info for A Theory of Heuristic Information in Game-Tree Search

Example text

Suppose that Fi is a Borel subfield generated by the partition of a game space corresponding to a heuristic search. Then we treat Xi =EFJX) as an estimator of X. If {F;} is an increasing sequence of Borel subfields, then {Xi' F;} becomes a martingale. This means that X i + 1 is a more precise estimator of X than is Xi. Such a model for estimating random variables is discussed fully in Chap. 7. 5 Probabilistic Game Models and Game Values In this chapter we introduce a formal definition of a probabilistic game model for further theoretical study.

Thus, E. = 1, that is, the probability that MAX wins the game starting from a node of height n if both players use minimax optimal strategies. Since M 0 is the value at a terminal node, its expected value is the probability of 1 at the terminal node: Eo=p· Let n > 0 be an even integer. Then, M. is the minimax value of a MAX node; M. = 0 if and only if the minimax value of each successor of the node is zero. _1)b. That is, if n is even. Now let n > 0 be an odd integer. Then, M. is the minimax value of a MIN node; 46 Probabilistic Game Models and Game Values M n = 1 if and only if the minimax value of each successor of the node is one.

4 Average Propagation The conventional minimaxing process is the right way to combine values at a node if those values are actual minimax values. It is also the best way to combine values ifthe player's opinions ofthe values of previously analyzed positions do not change on later moves, that is, if those values can be treated as actual values. However, real game-playing programs reanalyze positions after each move is made and usually come up with slightly different opinions on the later analyses.

Download PDF sample

A Theory of Heuristic Information in Game-Tree Search by Chun-Hung Tzeng

by Steven

Rated 4.73 of 5 – based on 38 votes