next up previous contents
Next: References Up: Outside work Previous: Extra Readings

Problems

 

  1. gif Understand the Turtle Graphics description in the L-systems tutorial at Australian National University and answer exercise 1 at the end about how to create a random walk L-system.
  2. Try out many L-systems in FRACTINT.
  3. Try creating new L-systems definitions in FRACTINT.
  4.   Write out a proof by induction that the global process works for the Fibonacci system. Start by showing that it works for the first few generations. Then explain carefully why, if you assume tex2html_wrap_inline9842 is true, then you can conclude that tex2html_wrap_inline7096 is true.
  5. Explain how the truth of the global process immediately shows that the number of organisms in the n-th generation of the Fibonacci system is the n-th Fibonacci number tex2html_wrap_inline7100 .
  6. Let tex2html_wrap_inline5118 be the sequence of generations produced in the Fibonacci L-system. Since tex2html_wrap_inline5122 is the beginning of tex2html_wrap_inline5324 , we can consider the limit of tex2html_wrap_inline5122 as tex2html_wrap_inline4974 , which is an infinite sequence tex2html_wrap_inline5358 of a's and b's. The beginning is

    displaymath9870

    1. Write out the first fifty terms of tex2html_wrap_inline5358 .
    2. Suppose we write tex2html_wrap_inline9874 , where each tex2html_wrap_inline5378 is an a or a b. From your calculation in the first part, write down the sequence of integers n for which tex2html_wrap_inline9884 . (The first few are tex2html_wrap_inline9886 .)
    3. Let tex2html_wrap_inline9888 (the golden mean). For tex2html_wrap_inline9890 (or more), compute tex2html_wrap_inline9892 , where tex2html_wrap_inline9894 means the integer part of the real number x. Compare your results with the previous answer and discover a pattern!
    4. (CHALLENGE) Prove the pattern.
  7. (CHALLENGE) Prove the global process for the Thue-Morse L-system.
  8. (CHALLENGE; Problem A5 on the 1992 Putnam Exam) For each positive integer n, let

    equation2998

    Show that there do not exist positive integers k and m such that

    equation3004

    (This problem asks you to prove the theorem that the Thue-Morse sequence is cube-free; to do this you must use the production rules tex2html_wrap_inline9914 and tex2html_wrap_inline9916 . As preliminary work, try to prove the following:

    1. No substring of the form 000 or 111 can occur.
    2. In every substring of length 5, there must occur either 00 or 11 as a substring.
  9. For the L-system

    align3012

    find the ``global process,'' i.e. express tex2html_wrap_inline9924 as some combination of tex2html_wrap_inline5324 and tex2html_wrap_inline5122 .

  10. On the order 7 dragon curve (supplied as a separate sheet), draw arrows to indicate the exact path that the curve follows.
  11. Let tex2html_wrap_inline5122 be the generations of the sequences of L's and R's that come out of the paperfolding process described here. So, in unfolding the paper the L's correspond to creases pointing downward, and the R's correspond to creases pointing upward.
    1. Find a ``global process'' describing how to create tex2html_wrap_inline5324 from tex2html_wrap_inline5122 . (Hint: the new folds occur between the old ones. Consider the left and right halves of the paper separately. Describe how to enumerate the new folds.)
    2. Write out tex2html_wrap_inline5154 .
    3. Write tex2html_wrap_inline5122 as tex2html_wrap_inline9948 where tex2html_wrap_inline9950 corresponds to an L and tex2html_wrap_inline9954 corresponds to an R. Find a simple formula for tex2html_wrap_inline9958 .
    4. Explain why tex2html_wrap_inline9960 .
    5. Find a formula for tex2html_wrap_inline5396 . (You need to write tex2html_wrap_inline9964 .)
  12. Let tex2html_wrap_inline5122 be the generations of L's and R's coming out of the ``alternating'' paperfolding sequence, where first you fold right side over left, then you fold right side under left, then you fold right over left, and so on, alternating the two types of folding operations.
    1. Write out the first five generations tex2html_wrap_inline5122 . The figure below gives you some help.
    2. Draw the corresponding dragon curves.
    3. Try to find the global process.
    4. Try to find an L-system construction of the corresponding dragon curves.

      figure3028
    Figure 72: Alternating Paperfolding Sequence

  13. (CHALLENGE) These questions are about the original paperfolding dragon curves. Write generation tex2html_wrap_inline5122 as tex2html_wrap_inline9978 where each tex2html_wrap_inline9980 is L or R.
    1. Many of the vertices on the curve are corners of a square closed off by the curve. Other vertices we call ``waist points.'' For the generations tex2html_wrap_inline5122 , n=5,6,7, write down the sequence of vertices which are waist points. For example, for the first four generations the sequences are as follows.

       

      tex2html_wrap_inline5138 1
      tex2html_wrap_inline5142 1-3
      tex2html_wrap_inline5146 1-7
      tex2html_wrap_inline5150 1-6,12-15
      Table 9: Waist Points

      The vertices 7-11 bound a square in generation 4.

    2. Find a pattern for the waist points for generation n.
    3. In the limit as tex2html_wrap_inline4974 , the dragon curve tends to a sequence of ``islands'' touching at waist points. What is the relation of one island to the next?
    4. What is the area of an island?
    5. Is there an L-system that describes just the boundary of the dragon islands?
  14. Consider the branching L-system tex2html_wrap_inline10006 discussed in this section.
    1. Determine the number of F's in each generation.
    2. Determine the number of new branches in each generation.
  15.   Determine the ``number of congruent pieces'' and the scale factor for the following L-systems listed in FRACTINT: CantorDust, Cesaro, Curve1, Sierpinski1, SierpinskiSquare.
  16. For the L-system Tree1 in FRACTINT, find some analogue of the concept of self-similarity, with an associated number of congruent pieces and a scale factor.
  17. Design an L-system that produces the usual tiling of the plane by equilateral triangles. Prove it by printing the results out.
  18. Show that the map

    displaymath10016

    is a similarity by explicitly determining a dilation tex2html_wrap_inline5994 , a rotation tex2html_wrap_inline6076 , and a translation tex2html_wrap_inline6022 such that

    displaymath10024

  19.   Referring to the condition on similarity, use the fact that tex2html_wrap_inline10026 , where tex2html_wrap_inline10028 denotes the transpose of the column vector tex2html_wrap_inline6452 , to deduce an equation satisfied by A.
  20. In tex2html_wrap_inline6388 , we let tex2html_wrap_inline10036 denote the usual Pythagorean length of the vector tex2html_wrap_inline10038 , we define just as before a similarity tex2html_wrap_inline10040 to satisfy, for some c>0, tex2html_wrap_inline10044 for all vectors tex2html_wrap_inline10046 . If c=1, U is an isometry. Let tex2html_wrap_inline10052 and tex2html_wrap_inline10054 just as in the two-dimensional case.
    1. Prove that U can be written

      displaymath10058

      where tex2html_wrap_inline10060 is an isometry such that tex2html_wrap_inline10062 , where tex2html_wrap_inline10064 .

    2. Prove that tex2html_wrap_inline10066 where A is a tex2html_wrap_inline10070 real matrix satisfying tex2html_wrap_inline10072 , the tex2html_wrap_inline10070 identity matrix. (Hint: construct some rotations tex2html_wrap_inline10076 , tex2html_wrap_inline10078 such that tex2html_wrap_inline10080 fixes (1,0,0), (0,1,0) as well as (0,0,0). Try to follow the rest of the argument in the section on similarities.)
  21. Design an L-system that produces the triomino tiling, and print out the results. (Hint: imitate the Sphinx L-system.)
  22. Explain how to use the Sphinx inflation process to tile the entire plane. To be precise, after doubling the subdivided pattern you must explain how to position it so that you can continue the inflation process.
  23. The vertices of a tiling are the points of intersection of 3 or more tiles. The edges are arcs which lie in the intersection of two tiles. The valence of a vertex is the number of edges meeting that vertex (at least three). For the Triomino and Sphinx tilings, determine all the possible values of the valence of vertex.
  24. Sketch the path defined by the Sphinx L-system for the tiles of type Y in analogy with this figure.
  25. (CHALLENGE) For the triomino tiling, find (with explanation) a value of R such that any circular disk of diameter R contains a copy of any patch of tiles of diameter tex2html_wrap_inline10100 . (see the discussion around this figure.)
  26. In this figure, we show the line E of slope tex2html_wrap_inline10104 and the region R as described in the section on self-similar tilings. Find the L-system in a and b that generates this tiling. To do this, try to find the integral matrix for the linear transformation T that has eigenvectors tex2html_wrap_inline10116 and tex2html_wrap_inline10118 and eigenvalues tex2html_wrap_inline10120 .

       figure3093
    Figure 73: Self-similar staircase.

  27. Compute by hand tex2html_wrap_inline10122 , tex2html_wrap_inline10124 . Explain the pattern that emerges.
  28. Find the fixed point of the transformation T(z)=2iz+3. Compute the iterates tex2html_wrap_inline10128 for tex2html_wrap_inline10130 .
  29. Consider the large triomino in this figure shown decomposed into four smaller triominoes. Suppose the lower left corner of the large triomino is at 0, and the lower right corner is at 4. We show the triomino with vertices labelled by complex numbers in this picture.
    1. Find the complex transformation T(z)=az+b that maps the lower right small triomino onto the original large triomino.
    2. Find the fixed point of this transformation T.
    3. Use this transformation twice to extend the tiling by small triominoes outward. At the end, you will have drawn 64 small triominoes.

       figure3110
    Figure 74: Triomino with vertices labelled

  30.   Prove that the mapping F of code-space associated with the Fibonacci L-system is continuous. See this section for definitions.
  31.   Let F be the mapping associated with the Fibonacci L-system (see the previous exercise). Prove that for any tex2html_wrap_inline5130 in the code-space that tex2html_wrap_inline10146 is always equal to the same infinite Fibonacci sequence of a's and b's.
  32. Show that the dynamical system associated with the Thue-Morse L-system has two distinct fixed points.
  33.   Define a sequence of integers tex2html_wrap_inline10154 by

    align3125

    So tex2html_wrap_inline10158 , tex2html_wrap_inline10160 , tex2html_wrap_inline10162 , etc.

    1. Compute tex2html_wrap_inline10154 , tex2html_wrap_inline10166 .
    2. Compute the fractional part tex2html_wrap_inline4910 of tex2html_wrap_inline10170 , for tex2html_wrap_inline10172 .
    3. Guess what happens to tex2html_wrap_inline4910 as tex2html_wrap_inline4974 .
  34.   For any small tex2html_wrap_inline7376 , explain how to find a covering of the Cantor set by disks of radius smaller than tex2html_wrap_inline5748 that are mutually disjoint (that is, no overlap).
  35.   Do the following exercises in [Bar93], Chapter V: 1.1-1.7, 1.9, 3.4-3.7.
  36.   Determine the fractal dimensions for the following L-systems listed in FRACTINT: CantorDust, Cesaro, Curve1, Sierpinski1,
    SierpinskiSquare. (See this exercise.)
  37.   (CHALLENGE) Prove that the fractal dimension of tex2html_wrap_inline10184 is exactly 0.5.
  38. (CHALLENGE) Can you find a sequence of points tending to 0 in the unit interval whose fractal dimension is 1? (Yikes!)
  39.   In this section, a correspondence between tex2html_wrap_inline10188 and points on the Koch curve was described. Characterize all points on the Koch curve for which there correspond two different sequences in the code-space.
  40.   In the first figure below, we show the Sphinx tile S divided into four similar tiles numbered 1,2,3,4. Let tex2html_wrap_inline8562 be the similarity that takes S onto tile number i. In the second figure, we use these linear transformations to repeatedly subdivide the original Sphinx tile. Find the code sequences (in 1,2,3,4) that correspond to the small tiles labelled A, B and C.

      figure3162
    Figure 75: Code sequences for sphinxes

  41. Let tex2html_wrap_inline10204 and tex2html_wrap_inline10206 be the two affine maps whose attractor is the Cantor set. Show that tex2html_wrap_inline10208 belongs to the Cantor set by explicitly determining the code sequence in 1,2 that corresponds to x.
  42. For each of the four affine maps in the fern IFS in FRACTINT find the square roots of the eigenvalues of tex2html_wrap_inline8122 , where A is the tex2html_wrap_inline6172 coefficient matrix.
  43. Exercise 10.6 in Chap. III of Barnsley [Bar93]: use collages to find the IFS for each of the fractal figures III.72-76 in Barnsley.
  44.   (CHALLENGE) For tex2html_wrap_inline4842 , prove that tex2html_wrap_inline10222 for all x satisfying tex2html_wrap_inline10226 . (Hint: start by proving tex2html_wrap_inline10228 whenever tex2html_wrap_inline10226 .)
  45.   Let tex2html_wrap_inline10232 and tex2html_wrap_inline10234 . Suppose there is a linear map L(x)= ax+b, tex2html_wrap_inline6836 , such that tex2html_wrap_inline10240 (provided tex2html_wrap_inline10242 ). Find the relation between c and tex2html_wrap_inline6624 and determine the linear map L.
  46. Prove that tex2html_wrap_inline10250 and tex2html_wrap_inline10252 are linearly conjugate.
  47. Find values of tex2html_wrap_inline6624 for which tex2html_wrap_inline10256 has a stable cycle of period 5, 7, 6, resp.
  48. (CHALLENGE) Find the value of tex2html_wrap_inline6624 (numerically to 10 digits!) for which tex2html_wrap_inline9202 has a stable 3-cycle which begins with tex2html_wrap_inline9208 .


next up previous contents
Next: References Up: Outside work Previous: Extra Readings

David J. Wright
Mon Aug 19 17:21:15 CDT 1996