Qualitative Uncertainty Reasoning in AgentSpeak

Abstract

This dissertation proposes two novel approaches for extending AgentSpeak with qualitative uncertainty reasoning by integrating two dynamic variants of epistemic logic. These extensions address a crucial gap in the literature where qualitative approaches to uncertainty are seldom integrated into agent-oriented programming languages due to various challenges related to methodology, implementation, and computational complexity.The extensions provide various symbolic constructs that enable the modelling and reasoning of belief uncertainty. The significance of qualitative uncertainty reasoning is illustrated through a simple Minesweeper scenario and two complex uncertainty challenges from the 2019 Multi-Agent Programming Contest: uncertain navigation and agent identification. Given the ability to express qualitative uncertainty, we equip the agent with a more robust and effective way to plan and act under uncertainty.An in-depth evaluation of the performance and scalability of the proposed Agent-Speak extensions is provided, with a heavy focus on examining their impact on the agent's time-sensitive reasoning cycle. The results show that these extensions provide a tractable and computationally feasible approach to extending AgentSpeak with the ability to manage the statics and dynamics of uncertainty. i This dissertation is dedicated to my love, Sandra.A constant source of support and inspiration in every way imaginable, you have always given me the strength and determination to persevere and overcome any challenges that have arisen. I love you and will forever be grateful for the time and experiences we have and will endure together.To Mrs. Morley, I cannot thank you enough for the support you have given to both Sandra and I. Your belief in our potential and your constant push for us to aim higher and strive for greater things will forever impact our lives. It was through the support, encouragement, and inspiration of you and Sandra that I gained the confidence and strength to pursue further education. I love you and will miss you dearly. You will forever be cherished, and your influence will continue to guide Sandra and I as we navigate the next phases of our lives. I would also like to express my gratitude to Professor Esfandiari, whose support, guidance, and encouragement have been invaluable. Your expertise and wisdom have played a crucial role in my academic and personal growth, and I am truly grateful for the opportunity to learn from and work alongside you.I am grateful to all who supported me during this journey, as you have helped to form the foundation of this dissertation.