Authorised by Academic Registrar, April 1996
Objectives On completion of the subject, students should have a working knowledge of basic search techniques, knowledge representation and reasoning mechanisms, and planning systems; be able to analyse problems and determine what AI techniques are applicable; and be able to write LISP programs for implementing AI problem-solving solutions.
Synopsis This subject covers basic techniques and mechanisms required for the construction of intelligent agents, with a focus on reasoning and actions. There are five main topics. LISP: a functional programming language commonly used in AI applications. Problem solving: constructing search problems; uninformed search techniques (depth-first, breadth-first, iterative deepening); informed search (graphsearch, hill-climbing, simulated annealing), including developing heuristics; game playing (min-max, alpha-beta pruning); strategy evaluation (completeness, complexity, optimality). Knowledge and reasoning: logical reasoning to represent the world, update knowledge, and deduce actions to achieve goals; predicate calculus, situation calculus, frames and semantic networks; application using knowledge and engineering. Planning systems: combining problem-solving and knowledge and reasoning for practical planning and scheduling applications; STRIPS representation, partial-order planning, conditional planning, hierarchical planning. Uncertainty: reasoning in the presence of uncertainty; probability theory, belief networks.
Assessment Examination: 2 hours + Assignments