1 / 16

Solving Problems by Searching

Solving Problems by Searching. 제 4 주 강의. Problem solving agent. What to do by finding sequences of actions that lead to desirable states the current state (environment), applicable operations, nest states Find a sequence of states maximizing the performance measure.

sernesto
Télécharger la présentation

Solving Problems by Searching

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Solving Problems by Searching 제4주 강의

  2. Problem solving agent • What to do by finding sequences of actions that lead to desirable states • the current state (environment), applicable operations, nest states • Find a sequence of states maximizing the performance measure

  3. Problem solving • Goal formulation • Problem formulation :: the process of deciding what actions and states to consider • Missionary-carnival problem, 청소하는 로봇, 교육하는 시스템 등에 적용 • Sequence, solution

  4. Formulating Problems • Amounts of knowledge that an agent can have concerning its actions and the state that it is in  depends on the how the agent is connected to its environment through its percepts and actions

  5. Knowledge and problem type • Single-state problem ::: 에이전트는 세상을 정확히 인지하고, 행위의 결과도 정확히 안다. • Multiple-state problem ::: 행위의 결과는 정확히 알지만, 세상(환경)을 인식하는 데 한계가 있다; 또는 행위의 결과를 정확히 예측하지 못할 때; Murphy’s law

  6. Knowledge and problem type (cont.) • Contingency problem :::  Exact prediction is impossible  Solving this problem requires sensing during the execution phase;  Keep eyes open while walking or driving  Agents need to act before founding a guaranteed plan  Interleaving of search and execution

  7. Knowledge and problem type (cont.) • Exploration problem  No information about the effects of its actions  a strange country with no map  The agent must experiment, gradually discovering what its action and what sorts of states exists  Reinforcement learning

  8. Well-defined problems and solutions Initial state operation new state Successor function operation new state select a state goal test • state space  path, solution, path cost (g)

  9. Measuring problem-solving performance • Effectiveness of a search  Does it find a solution ?  Is it a good solution ?  What is the search cost associated with the time and memory to find a solution ?

  10. Example Problems • The 8(16)-puzzle problems 

  11. 문제 정의(8 puzzle) • <States, Operators, Goal test, Path cost> • Solving as search ::: Breadth first, Depth-first • Heuristics • 16 puzzle

  12. 8 Queens problem • Goal test ::: 8 queens on board, none attacked • Path cost ::: zero • States ::: any arrangement of 0 to 8 queens on board • Operators ::: add a queen to any square • Branch and Bound, Dynamic Programming • AI techniques

  13. Cryptarithmetic • FORTY 29785 + TEN 850 + TEN 850 -------- ------ SIXTY 31418 (F=2, O=9, R=7, etc.)

  14. The vacuum world • 복사하여 제공 후 설명 • Robot arm problem과 비교하여, 그림 3.7)처럼 결과 나오게 상태도를 그린다 (숙제)

  15. Other toy problems • Missionary and cannibals • Water jug

  16. Real world problems • Route finding • Travelling salesperson problem • VLSI design • Robot navigation • Assembly sequencing

More Related