CPSC 322 Prologue to Counterfeit consciousness.


73 views
Uploaded on:
Category: People / Lifestyle
Description
c) MAX has a static board assessment work that profits. greater qualities if a board is good to MAX. d) the assessment capacity shows signs of improvement as the amusement gets ...
Transcripts
Slide 1

CPSC 322 Introduction to Artificial Intelligence October 25, 2004

Slide 2

Things... Task 3 was posted Friday night Midterm exam 2 is one week from today Bring me your first midterms before the second midterm on the off chance that you need to be retested on issue 3 Jessica Hodgins talk, Thursday, October 28, 1-2pm, MacLeod 214

Slide 3

Game hunt (otherwise called antagonistic inquiry) has these segments: move (or load up) generator static load up assessment work (this is the heuristic part - it doesn\'t create advances or look - it\'s static) minimax calculation to on the other hand proliferate minima and maxima upward from "base"

Slide 4

Minimax calculation Start with the accompanying: a) there are two players, MAX and MIN b) it\'s MAX\'s swing to move c) MAX has a static load up assessment work that profits bigger values if a load up is great to MAX d) the assessment capacity improves as the amusement gets closer to an objective state (else why try to produce the diversion space?) e) MAX trusts that MIN\'s assessment capacity is no better than MAX\'s (if that is not valid, then MAX should in any event abstain from wagering cash on this amusement)

Slide 5

Minimax calculation Generate the amusement tree to the same number of levels (handles) that time and space constraints permit. The top level is called MAX (as in it\'s currently MAX\'s turn to move), the following level is called MIN, the following level is MAX, and so on. Apply the assessment capacity to all the terminal (leaf) states/sheets to get "goodness" values Use those terminal board qualities to decide the qualities to be assigned to the prompt guardians: a) if the guardian is at a MIN level, then the worth is the base of the estimations of its kids b) if the guardian is at a MAX level, then the worth is the most extreme of the estimations of its youngsters Keep proliferating values upward as in step 3 When the qualities achieve the highest point of the amusement tree, MAX picks the move showed by the most noteworthy worth

Slide 6

1997 Deep Blue Garry Kasparov essential insights: indispensable measurements: • 200,000,000 moves for every second • 3 moves for every second • 480 custom chess-playing chips • meat

Slide 7

Puzzles, recreations, and AI Puzzles and two-player tabletop games have since quite a while ago served as a lab for analyses with heuristic hunt Solving riddles and having recommends knowledge on the influence of a human player, however people don\'t do this stuff the same way that pursuit methodology do Electronic Arts likely won\'t employ you exclusively on your dominance of minimax inquiry (yet it couldn\'t do any harm)

Slide 8

Puzzles, diversions, AI Still, heuristic hunt is a straightforward yet helpful device, and the minimax amusement playing methodology is plainly effective You presumably don\'t completely comprehend these hunt strategies until you\'ve needed to execute them...especially diversion inquiry Let\'s take a gander at only one more amusement...

Slide 9

Oska • two players • four pieces each

Slide 10

Oska • two players • four pieces each

Slide 11

Oska • two players • four pieces each • piece development:

Slide 12

Oska • two players • four pieces each • piece development: • piece might be propelled one space on the inclining

Slide 13

Oska • two players • four pieces each • piece development: • piece might be advanced one space on the corner to corner

Slide 14

Oska • two players • four pieces each • piece development: • piece might be propelled one space on the askew

Slide 15

Oska • two players • four pieces each • piece development: • piece might be propelled one space on the slanting

Slide 16

Oska • two players • four pieces each • piece development: • piece might be advanced one space on the corner to corner • piece may hop forward on the askew over a rival\'s piece to a void space, in this manner catching the rival\'s piece

Slide 17

Oska • two players • four pieces each • piece development: • piece might be propelled one space on the askew • piece may bounce forward on the inclining over a rival\'s piece to a vacant space, in this way catching the adversary\'s piece

Slide 18

Oska • two players • four pieces each • piece development: • piece might be propelled one space on the askew • piece may hop forward on the corner to corner over a rival\'s piece to an unfilled space, accordingly catching the rival\'s piece

Slide 19

Oska • two players • four pieces each • piece development: • piece might be propelled one space on the askew • piece may hop forward on the corner to corner over a rival\'s piece to a vacant space, therefore catching the rival\'s piece

Slide 20

Oska • two players • four pieces each • piece development: • piece might be advanced one space on the slanting • piece may hop forward on the corner to corner over a rival\'s piece to a vacant space, along these lines catching the rival\'s piece

Slide 21

Oska • two players • four pieces each • piece development: • piece might be propelled one space on the inclining • piece may hop forward on the inclining over a rival\'s piece to a vacant space, in this way catching the rival\'s piece • you don\'t need to catch on the off chance that you would prefer not to

Slide 22

Oska • two players • four pieces each • piece development: • piece might be advanced one space on the corner to corner • piece may hop forward on the inclining over a rival\'s piece to a vacant space, consequently catching the rival\'s piece • you don\'t need to catch in the event that you would prefer not to

Slide 23

Oska • two players • four pieces each • piece development: • piece might be propelled one space on the corner to corner • piece may hop forward on the corner to corner over a rival\'s piece to a vacant space, subsequently catching the rival\'s piece • you don\'t need to catch on the off chance that you would prefer not to

Slide 24

Oska • two players • four pieces each • piece development: • piece might be advanced one space on the askew • piece may hop forward on the inclining over a rival\'s piece to a vacant space, in this manner catching the rival\'s piece • you don\'t need to catch on the off chance that you would prefer not to • different hops not allowed

Slide 25

Oska • two players • four pieces each • piece development: • piece might be advanced one space on the askew • piece may hop forward on the askew over a rival\'s piece to a vacant space, hence catching the rival\'s piece • you don\'t need to catch in the event that you would prefer not to • numerous hops not allowed

Slide 26

Oska • two players • four pieces each • piece development: • piece might be advanced one space on the slanting • piece may hop forward on the slanting over a rival\'s piece to a vacant space, in this way catching the rival\'s piece • you don\'t need to catch in the event that you would prefer not to • various hops not allowed

Slide 27

Oska • two players • four pieces each • piece development: • piece might be advanced one space on the inclining • piece may hop forward on the corner to corner over a rival\'s piece to a vacant space, in this way catching the rival\'s piece • you don\'t need to catch in the event that you would prefer not to • numerous hops not allowed

Slide 28

Oska • two players • four pieces each • piece development: • piece might be propelled one space on the slanting • piece may hop forward on the askew over a rival\'s piece to a vacant space, in this manner catching the rival\'s piece • you don\'t need to catch in the event that you would prefer not to • various hops not allowed

Slide 29

Oska • two players • four pieces each • piece development: • piece might be advanced one space on the askew • piece may hop forward on the slanting over a rival\'s piece to a vacant space, in this manner catching the rival\'s piece • you don\'t need to catch in the event that you would prefer not to • numerous hops not allowed

Slide 30

Oska • two players • four pieces each • piece development: • piece might be propelled one space on the slanting • piece may hop forward on the inclining over a rival\'s piece to a vacant space, in this way catching the rival\'s piece • you don\'t need to catch on the off chance that you would prefer not to • numerous hops not allowed • if a player can\'t make a lawful proceed onward a turn, the player loses turn and rival moves once more

Slide 31

Oska • how to win

Slide 32

Oska • how to win • catch all your rival\'s pieces

Slide 33

Oska • how to win • catch all your rival\'s pieces

Slide 34

Oska • how to win • catch all your rival\'s pieces

Slide 35

Oska • how to win • catch all your rival\'s pieces

Slide 36

Oska • how to win dark wins • catch all your

Recommended
View more...