CS 482/682: ARTIFICIAL INTELLIGENCE
FALL 2009
Homework 4
Problem 1 - Exploration with POMDPs:
In this problem, we will explore the navigation of discrete grid worlds using POMDPs. While you
do not have to actually solve the POMDPs, if you are interested i
Lecture Notes for AI (CS 482/682)
Fall 2002, Lecture 10
Instructor: Kostas Bekris
Based on the notes by Thanasis Krontiris
September 24, 2008
1
Review of Propositional Logic
1.1
Propositional (Boolean) Logic
(not) : negation
(and) : conjunction
(or) :
Lecture Notes for AI (CS 482/682)
Fall 2009, Lecture 08
Instructor: Kostas Bekris
Based on the notes by Patrick Mahoney
September 17, 2008
1
Review of Search
In search problems the objective is to find a goal state given the state-space, a start state and
Lecture Notes for AI (CS-482)
Fall 2009, Lecture 11
Instructor: Kostas Bekris
Based on the notes by Khai Duc Du
September 29, 2009
1
Adversarial Search for Games
Two-player zero-sum games are the focus of the adversarial search techniques introduced in th
Lecture Notes for AI (CS 482/682)
Fall 2009, Lecture 09
Instructor: Kostas Bekris
September 22, 2009
1
Introduction
In the previous lecture we discussed Backtracking, a classic search algorithm for CSPs (Constraint Satisfaction
Problem). One tool to repre
Lecture Notes for AI (CS 482-682)
Fall 2009, Lecture 13
Instructor: Kostas Bekris
Based on the notes by Edgar Portigal
October 6, 2009
1
Review: Probability Theory
Axioms of probability (Kolmogorovs axioms)
0 <= P (a) <= 1, a
P (true) = 1, P (f alse) =
Lecture Notes for AI (CS 482/682)
Fall 2009, Lecture 05
Instructor: Kostas Bekris
September 8, 2009
1
Overview
In the previous lectures we explored uninformed search techniques, i.e., algorithms that search the state
space of a problem for a solution and
Lecture Notes for AI (CS 482/682)
Fall 2009, Lecture 06
Instructor: Kostas Bekris
Based on the notes by Gareth Ferneyhough
September 10, 2009
1
Local Search Algorithms
The search algorithms discussed so far have exponential time and/or space complexities
Lecture Notes for AI (CS 482/682)
Fall 2009, Lecture 02
Based on the notes by Joshua Hegie
Instructor: Kostas Bekris
August 27, 2009
1
AI Problems
Four components (P.E.A.S) form an AI problem:
1. a performance measure,
2. the environment where an intellig
Lecture Notes for AI (CS 482/682)
Fall 2009, Lecture 04
Instructor: Kostas Bekris
Based on the notes of Benjamin Seelbinder
September 3, 2009
1
Comparing Search Strategies
The table below illustrates the comparison between three basic search techniques. T
Lecture Notes for AI (CS 482/682)
Fall 2009, Lecture 12
Instructor: Kostas Bekris
October 1, 2009
1
Adversarial Search
The type of search used in multi-agent games to determine the optimal strategy against opponents is called
adversarial search. In classi
Lecture Notes for AI (CS 482/682)
Fall 2009, Lecture 07
Instructor: Kostas Bekris
Based on the notes by Valjean Bud Clark
September 15, 2009
1
Local Search Algorithms
Local search has two distinctive features: (i) it stores only a small number of states a
Lecture Notes for AI (CS-482)
Fall 2008, Lecture 03
Arthur LeVesque
September 3, 2008
1 Introduction
Many computational challenges can be solved by searching a state space. For example, shortest-path problems
on graphs require searching the space of all p