1. (Due January 29) Prepare a proposal for a class project.
2. (Due January 29) Solve the following constraint optimization problem: minimize the expression (x − 1)2 + (y − 1)2 under the constraint x2 + y2 =1.
3. (Due February 3) Write a program that, given two arrays x = (x1, ..., xn) and y = (y1, ..., yn), uses the Least Squares method that estimates m and b for which yi ~ m * xi + b for all i. Test your program, first by using the exact values yi = m * xi + b, and then by adding a small random noise to the values yi. Submit a printout of the program and the printout of the testing results.
Reminder: The Least Squares method leads to the following formula for estimating m: m = (x * y − x * y) / (x2 - (x)2), where:
4. (Due February 5) Write two programs that, given two arrays x = (x1, ..., xn) and y = (y1, ..., yn), estimates the parameters of, correspondingly, the exponential law y = A * exp(k * x) and power law y = A * xα (by reducing the estimation problem to the case of a linear dependence).
By February 12: please test your programs to make sure that it is correct, and deliver the testing results. To test your program, you need to first pick some A and k (or, for power law, A and α), pick any xi, generate corresponding yi, then apply your mail program and check that you get the same values A and k (or A and α) back:
5. (Due February 12) Write a program that, given three integers m, e, and n, computes me mod n, by using the corresponding part of the RSA algorithm.
6. (Due February 17) Write a program that, given two numbers a and b, computes their greatest common divisor.
7. (Due February 17) Write a program that, given two numbers φ(n) and e for which gcd(e, φ(n)) = 1, returns the value d for which d * e = 1 mod φ(n).
8. (Due February 17) Show, step by step, how you can parallelize multiplication of two 2 x 2 matrices if you have enough processors.
9. (Due February 19) Combine your RSA-related programs into a single package consisting of three methods:
10. (Due February 26; extra credit if turned in by February 24) Write a program that uses fuzzy methodology that we discussed in class -- based on the handout. This program should, given a difference in temperatures Δt, compute the appropriate control value u. Make this program modular: use separate methods
11. (Due March 19; extra credit if turned in by March 17) Write a program that simulates a simple 3-layer back-propagation neural network and how this network can learn.
12. (Due March 31) Report on the progress of your project.
13. (Due March 31) Sign for amazon.com cloud, learn how to use it.
14. (Due April 9) Estimate the costs and decide whether it is beneficial to sign a contract for T = 3 years. The cost of buying a unit of computations on a year-by-year basis is c0 = 1, the contract offer a discount price c1 = 0.89, the discount rate is q = 0.91, the price of computing decreases yearly by a factor of v = 0.85.
15. (Due April 9) Suppose that a company uses between m = 100 and M = 500 computations. All values between 100 and 500 are equally probable p(x) = const, and computing in the cloud is 4 times more expensive than computing in-house: c0 = 1 and c1 = 4.
16. (Due April 14) Let us now assume that every day, the company uses at least m = 500 computations, and the probabilities of different numbers of computations x is described by the power law p(x) = A * x−α, with α = 2.5. Assume that the cost of a unit in-house computation is c0 = 10 money units per computation unit and the cost of computing in the cloud is c1 = 22.5 money units per computation unit.
In these computations, you can use the formulas that we derived in class for the power law case:
17. (Due April 21) Use some functionality of the Amazon cloud and submit an explanation and a proof (e.g., a screen shot) of what you did.
18. (Due April 21) Follow the general procedure of solving multi-objective optimization problems on a simple example.