Link Search Menu Expand Document
CS310 - Machine Intelligence

Mitchell, Ch. 9

  1. Given the evaluations 5, 7, 19 , 21, 24, 31, provide the vector of fitness values using roulette wheel allocation. 5/107 ... 31/107

  2. Assuming a binary encoding, explain the term schemata. Schemata is the plural of schema. A schema is a pattern of 1's , 0's, and * that defines the set of all population members that exactly match the pattern in its positions that contain 0's and 1's.

  3. Give an eight bit example (page 254) of a crossover mask for:
    • Single-point crossover 11110000
    • Two-point crossover 00111000

  4. Given a crossover mask and two strings, apply the mask to obtain two children (page 254) .

  5. Define the length of a schema. maximum inclusive distance in bit positions between non-* entries in the schema

  6. Define the order of a schema. total number of non-* entries in the schema.

  7. Describe the schemata that crossover tends to preserve best. crossover favors low length, small order, high fitness schemata, and it tends to combine them into large length, large order, high fitness schemata that survive crossover

  8. What can one say about the crossover children of two parents that are members of the same schema. Their children are both members of that schema.

  9. GA searches a huge space of hypotheses for a solution. Give another example of a learning algorithm doing this. NN

  10. Given the following two parents, show the two offspring resulted from a cross over at position 2 (i.e. use the mask 1100). p1 = 1001 p2 = 0100

  11. T/F Always, after crossover, children have same length as the parents.

  12. T/F In GA, it is recommended to use mutations very frequently.

  13. T/F In GA, we keep for the next generation only the most fit hypotheses.

  14. T/F Assume h1 and h2 have fitness values 0.4 and 0.1, correspondingly. Then h1 has higher chance of being selected by the roulette wheel selection than h2.

  15. Given the description of a simple problem, write an eval function for gaot.

Problem: maximize x + 10sin(5x)+7cos(4x) on the interval -9,9.

Evaluation function: function [sol, val] = myEval(sol,options)

x=sol(1);

val = x + 10sin(5x)+7cos(4x);