Answer the following questions.
1.One argument against a traditional approach to AI, where all of the decision making is pre-programmed, is that even simple tasks can be very complicated. As an example lets look at a simple task - cutting a steak.
Assume that you have been give the task of coming up with an algorithm for a T-bone steak cutting machine for disabled persons. You should express the algorithm as a sequence of IF - THEN statements.
T-bone steaks are not uniform in size and have a T shaped bone in them that may or may not visible from both sides. They usually have a layer of fat on the edge of the steak.
Your system should be able to ask the user if they want the fat removed from the edge before eating and what size pieces they desire - small, medium, or large.
The steak cutting machine has a vision system, three arms: one to rotate the plate on the table, one with a knife, and one with a fork, as well as sensors to indicate the amount of pressure placed on the knife and fork.
Your algorithm should begin when the plate is placed on the table (plate has a steak, french fries, and mixed vegetables on it). Your algorithm should end after all of the steak has been cut into the desired sized pieces.
Remember that your algorithm is going to control the machine so it needs to be precise and it needs to be robust enough to handle any unforeseen occurrences (e.g. no steak on the plate).
2.Consider the following set of known facts and rules for a production system.
Facts (known to be true): F,G,H,I,J Rules:
1. K & G & J → L (means that if K and G and J are all true then L is true)
2. F → K
3. H → N
4. N & O → P
5. L & F & Q → X
6. N → Q
7. Q & F → R
8. R & H → S
9. F & N → T
10. R & S & T → V
a. Which rules would fire and in what order if we used forward chaining, assuming the system starts at rule 1 and goes to rule 10 and repeats. What facts do we know after forward chaining has completed?
b. Using backward chaining, prove that X is true. Which rules and in what order, would fire to accomplish this? What facts do we know after backward chaining has completed?
c. Using this example, what have we learned about the applications for forward and backward chaining? Which approach do you think is used more often and why? Give an example of a situation where each would be the best choice.
3.Consider the difference between a standard rule-based expert system and a fuzzy expert system. Then go to the discussion forum and give two examples of applications suited to a standard rule-based expert system and two examples of an application suited to a fuzzy expert system. Explain why your selected system would be the best choice for that application. Which type of system do you think is the most useful in general and why?