Presentation on theme: "Search 5 주강의. Generating action sequences Generating a new set of states expanding the state Search strategy ::: the choice of which state to expand."— Presentation transcript:
Generating action sequences Generating a new set of states expanding the state Search strategy ::: the choice of which state to expand first Search tree, Search graph Data structures for search trees : Queue, Stack
Search Strategies Completeness ::: guaranteeing to find a solution where there is one Time complexity Space complexity Optimality ::: to find the high quality solution ???
숙제 ( 기간 :: 10 월 24 일까지 ) Missionary-cannibals 문제에 대해 (1) depth-first search 로는 답을 구하지 못할 수 있음을 보여라 (2) depth-first search 에 같은 sate 가 다시 나타나면 제거하는 기능이 있으면 답을 구할 수 있다. 이를 프로그램하라. (3) breadth-first search 로 구현하라. (4) iterative deepening 방법으로 구현하라. (5) (2), (3), (4) 에 대해 space 와 수행속도를 비교하라
Comparing Search Strategies Criterion Breadth- first Uniform -cost Depth- first Depth- Limited Iterative Deepening Bidirect ional Timebdbd b d bmbm blbl bdbd B d/2 Spacebdbd bdbd bmblbdB d/2 Optimalyes no yes Completeyes noYes, If ≥d yes B: branching factor, d: depth of solution, m: maximum depth, l: depth limit
Constraint Satisfaction Search The states are defined by the values of a set of variables and the goal test specifies a set of constraints that must the values must obey Cryptarithmetic, 8 queens problem, VLSI design In CSP, the goal test is decomposed into a set of constraints on variables rather than being a blackbox
Search in CSP Depth-first search on a CSP wastes time searching when constraints have already been violated Backtracking search Forward checking ::: Each time a variable is instantiated, forward checking deletes from the domains of the as-yet-uninstantiated variables all of those values that conflict with the variables assigned Constraints propagation