CS-265/CSC325 Artificial Intelligence – Coursework 1 [15 marks]
Deadline: 6 March 2025, 11am GMT
Task Description
Consider the following situation: Assume you have a database of articles, together with their length in words and the topics covered. Using graph search you want to create an information brochure that is as concise as possible, i.e., as short as possible while covering all of the required topics. The article database contains the five articles in the table below.
Table 1: Database of articles, their length and topics covered.
|
Article |
Length in words |
Topics Covered |
|
art0 |
200 |
[introduction] |
|
art1 |
600 |
[industry 5.0, challenges] |
|
art2 |
1000 |
[introduction, human-centric design, cobots] |
|
art3 |
800 |
[infographic, skills] |
|
art4 |
1000 |
[industry 5.0, cobots] |
In the search graph, a node is a pair (To Cover, Articles〉where Articles is a list of articles that must be in the brochure, and To Cover is a list of topics that have to be covered.
• The neighbours of a node are obtained by first selecting a topic from To Cover.
• There is a neighbour for each article that covers the selected topic.
• The remaining topics are the topics not covered by the articles added so far.
Assume that the leftmost topic is selected at each step. For instance, given the database shown in Table 1, the neighbours of the node ([introduction, cobots], []〉, when ‘introduction’ 代 写CS-265/CSC325 Artificial Intelligenceis selected, are ([], [art2]〉and ([cobots], [art0]〉. Thus, each arc adds exactly one article but can cover (and so remove) one or more topics. Suppose that the cost of the arc is equal to the word count of the article added.
The goal is to design a brochure that covers all of the topics in the list Must Cover. The starting node is (Must Cover, []〉. The goal nodes are of the form ([], Brochure〉. The cost of the path from a start node to a goal node is the length of the articles used. Thus, an optimal brochure is a shortest collection of articles that covers all of the topics in Must Cover.
(a) Suppose that the goal is to cover the topics [introduction, industry 5.0, cobots] and the algorithm always selects the leftmost topic to find the neighbours for each node. Draw the search space expanded for a Greedy search. Number the nodes in the order of expansion until a solution is found. Nodes that are not considered should not carry a number. Your search tree should include all nodes expanded, highlight which node is a goal node, and show the frontier when the goal was found. [9 marks]
(b) Give a non-trivial admissible heuristic function h. [Note that h(n) = 0 for all nodes n is the trivial heuristic function.] Does your heuristic satisfy the monotone restriction for a heuristic function? [3 marks]
(c) Does the order of topics selected in the expansion of the search space influence the result found? Justify your answer. [3 marks]
