12. Consider a smart home as an agent. The smart home agent is programmed to   dim the lights in aroom when a person in the room stops reading a book and  turns on the TV. The behaviour that the smart agent exhibits is an example of areactive behaviour.

A.  TRUE

B.  FALSE

13. In a norm lifecycle, which of these is anon-terminal state?

A.  Violated

B.  Satisfied

C.  Expired

D.  Conditional

14. Which of these statements about the environment and environmental properties is FALSE?

1) The environment in a chess game is deterministic

2) For a card game in which each player draws one card in each round and the player drawing the highest card is declared the winner of the round, the  environment is episodic

3) The environment through which a space probe travels is dynamic

A. 1

B. 2

C. 3

D. None of the above (i.e. A, B and C are all true statements)

15. The government’s Christmas guidance for university students includes the following statement: “Universities should move learning online by 9

December so students can continue their education while also having the option to return home to study from there.”

Provide a normative representation (norm type, subject, object, antecedent, consequent) for the above-mentioned Christmas guidance statement.

A.  Authorization(Government, University, CurrentDate > 9 Dec 2020, Move-Learning-Online)

B.  Commitment(University, Student, True, Move-Learning-Online)

C.  Commitment(Student, University, CurrentDate > 9 Dec 2020, Move-Learning-Online)

D.  Commitment(University, Student, CurrentDate > 9 Dec 2020, Move-Learning-Online)

Part II - Answer TWO of the three questions. Each is worth 15 Marks.

II.1   Logic Programming

You are given the following Prolog program, which aims to characterise the parity of natural numbers represented in Peano notation - where successive   integers are denoted by the terms 0, s(0), s(s(0)), s(s(s(0))), … .

The program consists of one fact (stating that the number zero is even) and    two rules (one stating the double-successor of any even number is also even; and another stating that a number is odd is it is not even) :

even(0).

even(s(s(N))) :- even(N).

odd(N) :- \+ even(N).

a)   Draw an SLD tree for the query ?-odd(s(0)).  This query asks Prolog to verify that the number one is odd. You  should assume the default Prolog    search strategy and the standard "cut-fail" definition of negation as failure.  Clearly state whether the query succeeds or fails. [3 marks]

b)   Translate the above program into a well-formed formula First-Order Logic (FOL) in which each Prolog symbol is replaced by its closest classical counterpart (make sure your formula is correctly quantified and parenthesised, if necessary). [3 marks]

c)   By reasoning about the structure of the possible SLD trees that result when the predicate odd/1 is called with ground and variable arguments, respectively, highlight any potential pitfalls that users should be aware of when calling this predicate. [8 marks]

d)   Write down an alternative definition of odd/1 whose behaviour is always intuitively correct (irrespective of whether its argument is an input or an output). [1 mark]

II.2   Genetic Algorithms

Gary is interested in evolving images using a genetic algorithm. Each image is a  50-by-50 grid of coloured pixels. Each genotype is a one-dimensional array of N alleles that each explicitly specify one Red-Green-Blue colour component of

one pixel. The colour of one pixel is defined by three consecutive alleles on the genotype, specifying the red, green, and blue components of the pixel's colour,  respectively. Each allele can take a floating point value in the legal range [0,1]. The first three alleles in the genotype define the colour of the top-left pixel, the next three alleles define the colour of the pixel immediately to the right of this   first pixel. The final triple of alleles defines the colour of the bottom-right pixel in the image.

a)  If each allele can take one of 2^16 different values between 0 and 1, how many different images does the search space contain? [1 mark]

b)  Gary implements mutation in the following way. Each allele in the genotype of a new offspring image has independent probability m of being mutated.

When an allele is mutated, the parental value is perturbed by adding a value   drawn uniformly at random from the range [-0.1, +0.1]. If the new, perturbed value is outside the legal range of [0,1] it is truncated to the nearest legal

value.

When Gary runs his genetic algorithm, he is interested to see that images with  bright bold colours tend to be evolved. Explain how his mutation operator may be contributing to this result. [4 marks]

c)   Gary is thinking of implementing sexual recombination. He wants to combine  two high fitness parental images to create an offspring. He has considered one- point crossover at a randomly selected position on the genotype, but he isn't

convinced that this will be a good choice. Define a simple crossover operator that will take two parental genotypes and create a single offspring genotype.  The offspring genotype should comprise a rectangular section of one parent   replacing the equivalent part of the image provided by the second parent. [4 marks]

d)  Gary is a little disappointed with the results generated by his genetic

algorithm. The images in the population tend to be quite similar and it takes a longtime to discover interesting images. Gary has been told by Sue that his    "direct genotype-phenotype mapping might be too simplistic", but he has not  come across this idea before. Help Gary to understand the concept of a

genotype-phenotype mapping and explain how a less direct mapping could allow evolution to generate more diverse images more quickly. [6 marks]