THIS INFORMATION IS PROVISIONAL. File hand1.pdf contain the latest information about this course.

Sequential and parallel algorithms for numerical and combinatorial problems will be discussed. The use of randomization in the solution of algorithmic problems will be explored. Applications to be considered will include string matching, polynomials and FFT algorithms, sorting networks, algebraic computations and primality testing and factoring, matrix operations, randomized algorithms for sorting and selection, and data compression. Search-indexing data structures (inverted lists) will be examined. Web ranking algorithms such as Google's PageRank, and Kleinberg's HITS will also be discussed. Topics in computational geometry will be examined.

INSTRUCTOR:	Alex Gerbessiotis	E-MAIL:	alg667@cs.njit.edu
OFFICE:	GITC 4213, 4th floor	TEL:	(973)-596-3244
OFFICE HOURS:	Tue 4:00-5:30pm, Thu 4:00-5:30pm
OFFICE HOURS:	By appointment some other time on Tue/Wed/Thu
CLASS HOURS:	Thu 6-9:05pm, Room: TIERN 114

Course Web Page: http://www.cs.njit.edu/~alexg/courses/cs667/index.html

1.2 Course Administration

Prerequisites

CS 610 or an undergraduate course on algorithms equivalent to CS 435.

Textbook

T.C.Cormen, C.E.Leiserson, R.L.Rivest, and C. Stein. ``Introduction to Algorithms'', 3rd edition, MIT Press, ISBN-10 : 0262033844 (ISBN-13: 978-0262033848). We abbreviate it in class as CLRS. Note that a 2nd edition is also available; Differences between the two are minor, and mainly on page numbers.

CourseWork:

Five homeworks HW1 to HW5 each one worth 200 points. The best 4 of 5 only count, so you can miss one. Programming problems related to the material covered in class will be given; you can substitute points gained from programming for regular points. Bonus points can also be gained.

Exam:

A take-home final exam (you can also call it a required homework) that will be posted at 6pm on Fri May 3, 2013 and due by 6pm on Thu May 9, 2013. It is worth 200 points (one homework worth of points).

Grading scheme:

1000 points is the target maximum collected from the best 4 of 5 homeworks (800 points max) plus the take-home final (200 points). You are to be given at least a C if you collect at least 500 points, at least 850 or more points are needed for an A. If more points are given, scale accordingly.

Extra work

Bonus points, when (and if) assigned in the 5 homeworks.

Due Dates

Written homeworks are due by the beginning of a class. Programs must be received by email by NOON the same day.

Late work

25% per day starting from 6:01pm for homeworks and 12:01 for programming assignments. No exceptions (only 4 of the 5 count towards the final grade). No emailing of late homeworks! No late exams.

The following describes a tentative list of topics that is intended to be covered in class. The code Ti refers to a topic. A topic may spread over one or more lectures. The code ALi refers to the ACM Computing Curricula 2001 topic description code. In parentheses, we provide an approximate number of hours per topic. Hour coverage may change depending on demand and circumstances (eg. weather).

AL3 : Fundamental Computing Algorithms  (15 hours)
AL8 : Advanced Algorithmic Analysis     (13 hours)
AL6 : Complexity                        ( 3 hours)
AL9 : Cryptographic Algorithms          ( 5 hours)
AL11: Parallel algorithms               ( 6 hours)



Topics  to be covered


T1 : AL3  : Representation of Polynomials, Polynomial evaluation with and without preprocessing.  
            Lower bounds for polynomial evaluation. Exponentiation. Polynomial algorithms for multiplication (Karatsuba's Algorithm). 
T2 : AL8  : Polynomial Interpolation and Convolution. DFT and FFT and FFT implementations. Applications to information hiding.
T3 :AL8/9 : Arithmetic algorithms (eg. Euclid's algorithm) and the bit model, solution of modular equations, 
            Chinese Remainder Theorem, Power raising through repeated squaring, primality testing (Miller-Rabin algorithm), 
            Pollard's rho method for integer factorization.  Public key cryptography and the RSA cryptosystem.
T4 : AL11 : Comparison networks, zero-one principle for sorting, bitonic and odd-even merge sort. 
T5 : AL11 : The PRAM model for parallel computation. Parallel algorithms for arithmetic problems and sorting. Parallel prefix.
T6 : AL3  : Matrix operations, Strassen's algorithm for matrix multiplication, boolean matrix multiplication, 
            solution of linear equations and matrix inversion. Complexity Results.
T7 : AL3  : Algorithms for string matching. Rabin-Karp. Knuth-Morris-Pratt. Boyer-Moore.
T8 : AL9  : Randomization and its application: Sorting, Selection and Number Theory 
T9 : AL3  : Data Compression (Huffman coding).
T10: AL8  : A quick overview on hashing.
T11: AL6  : Complexity classes (P,NP,RP,BPP,ZPP)
T12: AL10 : Geometric Algorithms
T13: AL8  : Google's PageRank and Kleinberg's HITS algorithms; Inverted Indexes.




Any modifications or deviations from the posted dates, will be done in
consultation with the attending students and will be posted on the course Web-page.
It is imperative that students check the Course Web-page regularly and frequently.





Course  Policies







Grading
Written work will be graded for conciseness and
correctness. Use formal arguments. Be brief and to the point.
Label solutions with problem/subproblem number clearly.
Programming problems will be graded based on test instances
decided by the grader on a test platform of
his choice (most likely, an AFS machine). 
Do not expect partial credit if your code fails to run on all test instances, and
you do not provide a bug report. 
Written Work
Handwritten or typed solutions must be readable, clear, concise and complete for 
the homeworks.  DO NOT USE pencils to write down your solutions; if you decide to 
use a pencil do not complain about grading. 
Programs
Code must be ANSI C/C++  compliant and 
compile on the test platform/compiler; for Java the AFS environment should
compile your code.
For Java, do not use packages that are not available on AFS.
Extension policies
Discussed on page 1. Note that 4 out of 5 homeworks only count.
Grade questions
Check the marks in a written work and
report errors promptly. Make sure you report such problems to the
grader or the instructor within two weeks from receipt but no later than the Reading Day. 
Final Grade
The final grade is decided based on a  0 to
1000 point performance with an adjustment made based on programming assignment 
performance. A student who collects at least 500 points should expect a passing
grade (C). At least 850 or more points may be needed for an A.
Collaboration 
Students who turn in solutions (programming or otherwise) that 
are derived from solution outlines of past assignments/homeworks,
were obtained through the Internet, or are
a product of another student's work, risk severe punishment, 
as outlined by the University. 
The work you turn in MUST BE your own personal work, composed and written 
by you. If you talk a problem with a fellow student 
cite this clearly in your homework (name the fellow student before the 
solution of the problem in question). Your work will then be 
compared to the other student's work to verify that your solution  was
written by you and reflect your own personal effort. 
If you don't report it, it will be considered a violation of the course rules.
You are not allowed to exchange code for the programming part of a 
homework. 
Mobile Devices
Mobile phones/devices and/or laptops/notebooks MUST BE SWITCHED OFF 
during class or be made noise-free.
Email/SPAM
Send email from an NJIT email address. NJIT spam filters
or us will filter other email address origins. Do not send course email 
to the instructor's email address unless there is a good reason 
(eg. you don't want the grader to read the email). 
Include CS 667 in the subject line then. 




The NJIT Honor Code will be upheld; any violations will be 
brought to the immediate attention of the Dean of Students.
Read this handout carefully!