Yes.
Of course. Remember, though, that efficiency doesn't mean saving an extra byte of memory here or there, or making your code as tight and cryptic as possible. Efficiency means things such as choosing an O(n) algorithm over an O(n^2) algorithm. Don't sacrifice program clarity for optimization.
Probably not. I just made up the numbers.
No, despite what it says on the assignment
No. Use real numbers, and don't round. The examples in the tutorial notes and the textbook use integers for the sake of simplicity.
The latter.
Yes! You won't be able to meaningfully compare the results otherwise. Make a new instance of each of the random number generators at the beginning of each algorithm.
You could as this is conceptually what is happening; in practice a real scheduler probably would not actually take the job off the CPU and put it back in the queue if the queue were empty.
After one second or when a process finishes, whichever is sooner. Read your textbook for more details.
http://www.cs.tpu.ee/~jaagup/uk/ds/chp1/chapter1.html
http://www.strath.ac.uk/Departments/DMEM/MSRG/simulate.html
http://www.labmed.umn.edu/~michael/des.html
No.
You can if you want to.
Adopt an efficient strategy.
In the past this term may have been used somewhat ambiguously; the tutorial notes have since been updated to use the term interarrival time -- i.e. the time between arrivals of subsequent processes. The actual point in time a process arrives in the system is more accurately called the arrival instant.
No. In fact, you probably shouldn't.
The namespace size is the number of distinct pages a process has. It is an upper bound on the number of distinct pages in the reference string.
Assume you have sufficient memory to allocate 64 frames.
There is a typo. The text "If the IFT is greater than or equal to the IFT" should read "If the IFT is greater than or equal to the threshold". This error is not duplicated in the Week 8 tutorial notes.
It means that for some algorithms, you may be able to pass through the reference string once, generating some special data structure, and then use that data structure to quickly and efficiently determine the page fault rate. (Then again, maybe not. This is for you to determine.)
For input, see the Java FAQ at http://www.afu.com/. For formatted output, use java.text.NumberFormat, or use some third-party I/O library. One such library is at http://www.horstmann.com/corejava/Format.java; it emulates C-style printf() calls.
No. It's just a sequence of page references separated by white space.
Pick some tie-breaking rule and document it. Don't just pick a victim at random.