锐客网

  1. 首页 > it技术 > >

MA117 科学统计

算法

【MA117 科学统计】MA117 Programming for Scientists: Project 2 Deadline: 12pm, Friday 19th March 2021
1
MA117 Project 2: Root Finding
Administrative Details
? This project is the second of the three assignments required for the assessment in this course. It is to
be submitted by 12pm, Friday 19th March 2021. Details of the method of the submission via the Tabula
system have been described in the lecture notes and are also available on the course web page.
? This assignment will count for 35% of your total grade in the course.
? The automated submission system requires that you closely follow instructions about the format of
certain files; failure to do so will result in the severe loss of points in this assessment.
? You may work on the assignment during the lab session, provided you have completed the other tasks
that have been set. You can use the work areas at all times when they are not booked for teaching, 7
days per week. If you are working on the assignment on your home system you are advised to make
regular back-up copies (for example by transferring the files to the University systems). You should
note that no allowance will be made for domestic disasters involving your own computer system. You
should make sure well ahead of the deadline that you are able to transfer all necessary files to the
University system and that it works there as well.
? The Tabula system will be open for the submission of this assignment starting from 1st March 2021.
You will not be able to test your code for correctness using Tabula but you can resubmit your work
several times, until the deadline, if you find a mistake after your submission. A later submission always
replaces the older one, but you have to re-submit all files.
? Remember that all work you submit should be your own work. Do not be tempted to copy work; this
assignment is not meant to be a team exercise. There are both human and automated techniques to
detect pieces of the code which have been copied from others. If you are stuck, then ask for assistance
in the lab sessions. TAs will not complete the exercise for you, but they will help if you do not
understand the problem, are confused by an error message, need advice on how to debug the code,
require further explanation of a feature of Java or similar matters.
? If you have more general or administrative problems e-mail me immediately. Always include the course
number (MA117) in the subject of your e-mail.
1 Formulation of the Problem
Finding the roots of a function is a classical and extremely well-known problem which is important in many
branches of mathematics. In Analysis II, you have probably seen that it is often easy to prove results about
the existence of roots. For example, using the Intermediate Value Theorem, you should easily be able to
prove that the function () = ? cos has a root in the interval [0,1]. On the other hand, calculating an
exact value for this root is impossible, since the equation is transcendental.
MA117 Programming for Scientists: Project 2 Deadline: 12pm, Friday 19th March 2021
2
Root finding is a classic computational mathematical problem, and as such there are many algorithms which
one may use to approximate the roots of a function. In this project, you will write a program which uses the
Secant algorithm. Let ∶ ? → ? be continuously differentiable and pick two points 0, 1 ∈ ?. Consider the
sequence of complex numbers {
}=1
∞ generated by the difference relation
+1 = ? (
)
? ?1
(
) ? (?1)
.
Figure 1: An example of fractals for the function
() =
3 ? 1 in the square with bottom left-corner ?1 ? and width 2.
Typically, if converges and lim→∞
=: ?
, then (?) = 0.
Some methods of finding the roots (Newton-Raphson) require knowing the derivative of . Whilst there are
numerical tricks to accomplish this, this problem is somewhat beyond the scope of this course. Instead, you
will use the Secant method which does not require the derivative, however it does require two initial values.
We will consider a polynomial ∈ ?[]; i.e.
() = 0 + 1 + 2
2 + ··· +

Where ∈ ?.
1.1 Secant Fractals
One of the most fascinating aspects of this problem arises from a very simple question: given two starting
positions 0, 1 ∈ ?, which root does the sequence produced by Secant converge towards? It turns out that
the answer to this question is very hard!
? 1 1
Re( z)
? 1
1
? 1 1
Re( z)
MA117 Programming for Scientists: Project 2 Deadline: 12pm, Friday 19th March 2021
3
Figure 1 shows two examples of how we might visualise this for the polynomial () =
3 ? 1. Recall that
the roots of this polynomial are =
2
3 for = 1,2,3 (i.e. the third roots of unity). Each of the three
colours represents one of these roots. In the left-hand figure, we colour each point depending on which root
the method converges to. The right-hand figure is the same, asides from the fact that we make the colour
darker as the number of iterations it takes to get to the root within a tolerance increases. The resulting
images are examples of fractals, which you have undoubtedly seen before.
Don’t worry if all of this seems quite difficult – the main aim of the assignment is for you to successfully
implement the Secant scheme. Most of the code to deal with drawing and writing the images will be given
to you.
1.2 Summary
Your task then involves several distinct elements. You will:
? write a class to represent complex numbers;
? write a class to represent a polynomial in ?[];
? implement the Secant method to find the roots of the polynomial;
? investigate some interesting fractals and draw some pictures!
2 Programming instructions, classes in your code and hints
On the course web page for the project, you will find the following files, which should serve as templates and
should help you to start with the project. As with the previous projects, the files have some predefined
methods that are either complete or come with predefined names and parameters. You must keep all names
of public objects and methods as they are in the templates. Other methods have to be filled in and it is up
to you to design them properly. The files define three basic classes for your project:
? Complex.java: represents points ∈ ?;
? Polynomial.java: represents polynomials in ?[];
? Secant.java: given two initial Complex points 0, 1 close to the root calculate the corresponding
root of Polynomial by Secant, if possible;
? SecantFractal.java: will generate a fractal similar to the one pictured above in a square.
These classes are documented in more detail in the following sections. You should complete them in the
order of the following sections, making sure to carefully test each one with a main function.
2.1 Complex
Complex is the simplest of the classes you will need to implement and will represent complex numbers. In
fact, it bears a striking resemblance to the CmplxNum class you (hopefully) implemented in week 14.
They are not identical however, so you should carefully copy and paste your code into this new class.
MA117 Programming for Scientists: Project 2 Deadline: 12pm, Friday 19th March 2021
4
2.2 Polynomial
The Polynomial class is designed to represent a polynomial () = ∑

=0
. As such, it contains
coeff, an array of Complex coefficients which define it. It is assumed that coeff[0] corresponds to 0,
coeff[1] to 1 and so forth. To complete this class, you will have to:

  1. Define appropriate constructors. There are two that need implementation; a default constructor
    which initialises the polynomial to the zero polynomial (i.e. 0 = 0), and a more general constructor
    which is passed an array of Complex numbers {0,1,… , } which should be copied into coeff.
    In addition, you should ensure that if any of the leading co-efficients are zero then they are not copied.
    For example, if the constructor is passed the complex numbers {0, 1,0, 0} then it should
    copy{0,1
    } to coeff. (When testing for equality to zero, do not use any tolerances.)
  2. Return the degree of the polynomial. Recall that = .
  3. Evaluate the polynomial at any given point ∈ ?. Note that you should not implement a pow
    function inside Complex as it is unnecessary and inefficient. Instead, notice that (for example)
    () = 0 + 1 + 2
    • 3
  5. = 0 + [1 + (2 + 3)].
    ?z
    Figure 2: A 5x5 pixel image representing the square with top left corner
    ?1 + and bottom right corner 1 ? . The centre of each pixel represents a
    complex number on the plane.
    2.3 Secant
    This class will perform the Secant algorithm. There are two constants defined in this class:
    ? MAXITER: the maximum number of iterations to make; that is, you should generate the sequence
    for 0 ≤ ≤ MAXITER and no more.
    ? TOL: At each stage of the Secant algorithm, we must test whether a sequence converges to a limit. In
    this project, we will say that approximates this limit if, at any stage of the algorithm, | ? ?1
    | <
    TOL . We then say that the starting points 0, 1 required iterations to converge to the root.
    MA117 Programming for Scientists: Project 2 Deadline: 12pm, Friday 19th March 2021
    5
    Additionally, you will need to define the iterate function. This accepts two parameters, 0, 1, which defines
    the initial condition of the Secant difference relation, and performs the root finding algorithm. There are
    three things that can occur during this process:
    ? everything is fine and we converge to a root;
    ? the difference |(
    ) ? (?1
    )| goes to zero during the algorithm;
    ? we reach MAXITER iterations.
    If any of the last two cases occur, then you set the error flag err to be ?1 and ?2 respectively; otherwise, err
    is set to zero. Here is a quick example of how Secant should be used:
    Complex[] coeff = new Complex[] { new Complex(-1.0,0.0), new Complex(),
    new Complex(), new Complex(1.0,0.0) };
    Polynomial p = new Polynomial(coeff);
    Secant n = new Secant(p);
    n.iterate(new Complex(), new Complex(1.0, 1.0));
    System.out.println(n.getRoot());
    This will print out the root of () =
  6. ? 1 obtained with the starting point 0 = 0, 1 = 1 + .
    2.4 SecantFractal
    SecantFractal will be responsible for drawing images of the fractals we saw in figure 1. However, let us
    briefly consider how images are represented on computer first. A two-dimensional image is, in general,
    broken down into small squares called pixels. Each of these is given a colour, and there are generally many
    hundreds of pixels comprising the width and height of the image.
    An example of this can be seen in figure 2. This image (badly) represents a square in the complex plane with
    top-left corner ?1 + and bottom-right corner 1 ? . In SecantFractal you will generalise this concept
    to visualise squares with a top-left corner origin and width width, stored as instance variables inside
    SecantFractal. The image will be of size NUMPIXELS by NUMPIXELS. Each pixel can be accessed by using
    an ordered pair (, ) where is the row number, the column number and (0,0) is the top left pixel, with
  7. ≤ , < . The image itself will be generated using createFractal, which accepts a
    single argument colourIterations. When true, the function generates a figure like the right-hand
    side of figure 1.
    To complete the class, first ensure that you call the setupFractal function at the end of your constructor.
    This will initialise the more complex drawing objects. It also checks that the polynomial you have given it has
  8. ≤ ≤ 5 . You will not need to consider any other polynomials in this class. Then inside
    createFractal, use the following logic:
  9. Copy colourIterations to the instance variable.
  10. Iterate over each pixel at position (, ). Then translate this position to a complex number using
    pixelToComplex, which uses the simple mapping (, ) ? + Δ( ? ).
  11. Use 0 and this complex number as the starting points to run through the Secant algorithm.
    MA117 Programming for Scientists: Project 2 Deadline: 12pm, Friday 19th March 2021
    6
  12. Check to see whether you’ve found this root already. You will store the list of already found roots inside
    the ArrayList roots. This is the purpose of the findRoot function. In this formulation, two
    complex numbers 1 and 2 are equal if |1 ? 2
    | < Secant. TOL.
  13. Finally, colour the pixel using the colourPixel function.
    After you are done, you can save the image using saveFractal. Here is an example from start to finish,
    which creates the two images of figure 1. Note that your filename should end with .png:
    SecantFractal f = new SecantFractal(p, new Complex(-1.0,1.0), 2.0);
    f.createFractal(false);
    f.saveFractal("fractal-light.png");
    f.createFractal(true);
    f.saveFractal("fractal-dark.png");
    You should then create a document which is precisely one page long. In this document, pick a polynomial
    (), a square in the complex plane and use your program to generate the two plots. You should call this file
    Fractal.pdf and ensure it is saved as a PDF file.
  14. Submission
    You should submit, using the Tabula system, the following four files: Complex.java,
    Polynomial.java, Secant.java, SecantFractal.java and Fractal.pdf which contains
    your plots.
    There will be a large number of tests performed on your classes. This should allow for some partial credit
    even if you don’t manage to finish all tasks by the deadline. Each class will be tested individually so you may
    want to submit even a partially finished project. In each case, however, be certain that you submit Java files
    that compile without syntax error. Submissions that do not compile will be marked down.
    If you submit a code and later (before the deadline) you realise that something is wrong and you want to
    correct it, you may do so. You can submit as many versions as you wish until the deadline. However, only the
    last submission will be tested. Each new submission replaces the previous one, so in particular, you MUST
    submit all files required for the project even if you corrected only a single file.
    Keep back-up copies of your work. Lost data are not a valid excuse for missing the deadline.

    推荐阅读


    上一篇：360携手HarmonyOS打造独特的“天气大师”

    下一篇：antdv Spin组件拓展