The keys EX1, EX2... correspond to the entries in the syllabus. Some exercises may have their own links, others appear here directly. RN4 means Chapter 4 of Russell and Norvig.
In general it's a good idea to go through all the exercises in RN for the current chapter to make sure you feel comfortable with them. On the other hand, sometimes the problems are not directly from the chapter, and may require a little outside reading. If you are frustrated by a problem, see the Prof. or a TA. A list of exercises in parentheses like this: (1.9 -- 1.13) means work any one of the exercises in the list. [Square Brackets] mean "work any two". You can do extra problems if you like. Make sure we know which ones they are. If there is a lot of prose in your answers (as opposed to math), consider using a text-processor! Write mathematics correctly -- if in doubt see the Homework and writing tools helper page.
EX1: RN1, RN2, RN3
RN1: (1.9 -- 1.13) that is do ONE of 1.9 -- 1.13.
RN2: 2.1, 2.2, 2.3, (2.10, 2.11), 2.12 (5 points per answer).
Read all the Quake (or Robocup) documentation you can (see links on syllabus page). For this exercise what is important is how we can reprogram the "monsters" to be agents under our control. Quagents have limited ability to do meaningful actions: (mainly moving and picking up things). With sensing we can do more: sense other players, proximity, distance, identification of objects, touching, etc. Make PEAS descriptions of two different sorts of quagents you would like to work with during the semester. We're looking for good ideas for quagent abilities, tasks, and personalities. (10 points for the PEAS)
RN3: 3.1, 3.3, 3.6, 3.7(a,b), 3.7(c,d), (3.12, 3.13).
Write a short paper (say 1 to 3 pages) on the usefulness of state-space problem solving for the quagents you wrote PEAS descriptions for. Problems 3.15 (3.16) and 3.19 might be related (more or less). (15 points for the short paper).
EX2: RN4, RN6
RN4: (4.2, 4.7), 4.3, 4.9, 4.11, 4.14.
Do 4.17, only with no implementation. This style of navigation is a very important one in real-world robotics (check out p. 921 of text). (5 points per answer).
Check out this self-referential aptitude test! . Lots of points extra credit for a program to solve it. (Strictly optional, but 25 points for this!).
Good idea: look over the Chapter 5 questions in RN.
RN6: 6.1, (6.2, 6.7), 6.3 (5 points per answer).
Write a short paper (1-3 pages, or generally "as long as necessary, as short as possible"). on the relevance of game-playing techniques to quagents engaged in some sort of Quake-like competitive activity like shooting or scavenging. (15 points for the short paper).
EX3: RN7, RN8, RN9, RN10
RN7: 7.4 (work any 2 parts), 7.6, 7.8 (work any 4 parts), 7.9.
RN8: (8.2, 8.3, 8.4) (work any 1), 8.6 (work any 3 parts), (8.7, 8.8, 8,9) (work any 1), (8.10, 8.11, 8.12), [8.13, 8.14, 8.15, 8.16] (work any 2).
To see what test questions might be like, check out this AI Classic: An Assignment from 1999 .
RN9: (9.3, 9.4), (9.9, 9.11), 9.19.
For the Quagents whose PEAS descriptions you wrote in EX1, or for a new Quagent you can now imagine thanks to your greater sophistication, give the PEAS description and then axiomatize the performance measure, environment, actuators and sensors. That is, describe how the world and intelligent agent work in first order predicate calculus. You may want to use situational calculus for the actuators at least. A complete axiomatization is unrealistic, but do enough axioms in each (P,E,A,S) category that we can tell you have the idea. Your (pretend) goal is that you can simulate the relevant aspects of the agent's universe (not the graphics!) using a theorem prover and your axioms.
EX4: RN22, RN23
RN22: 22.1, 22.2, 22.3, 22.5, 22.6, 22.7, 22.8, 22.9, (22.12, 22.13), 22.14.
22.10 and 22.11 seem related to programming project.
RN23: 23.6, 23.8, (23.9, 23.10).
EX5: RN11, RN12
RN11: 11,1, 11.3, (11.5, 11.6, 11.7), 11.13.
RN12: 12.1, 12.9, [12.19, 12.20, 12.21].
Short paper (this one may be more like 2-5 pages I'm thinking): I'm interested in whether it makes sense to apply AI planning to intelligent agents. Give your current favorite PEAS description and then consider how best to give the Quagent a planning capability. That is: What problems can planning be applied to? What planning functionalities make sense for your agent? What sort of planner do you need? Can you get such a planner from somewhere (like UCPOP or its descendents free off the web). Is it practical?
EX6: RN13, RN17, RN21
Here's another swine of a problem set...we gotta suck it up here!
RN13: (13.1, 13.2), 13.3, 13.5, 13.6a,c,d, (13.9, 13.10), (13.11, 13.13, 13.15, 13.16), 13.19.
RN17: 17.1, 17.2, (17.4, 17.9, 17.10, 17.11, 17.12, 17.13).
RN21: (21.2, 21.4), (21.5, 21.8), (21.9, 21.10).
EX7: RN20, RN24, RN25
RN20: 20.11, (20.12, 20.13), (20.19, 20.20). RN24: 24.1, 24.2, 24.3, 24.4, 24.5, 24.6, (24.7, 24.8), 24.10.
RN25: Choose one of the following three: (25.1, 25.3, 25.6.)
Here's a Hint for Exercise 25.1 .
Schedule and Syllabus
242 Home
HonestyThis page is maintained by CB.
Last update: 11.16.04.