Action reasoning in real time / Rana Mohamed Hamed ; Supervised Hesham Hefny , Ammar Mohammed

Thesis (M.Sc.) - Cairo University - Institute of Statistical Studies and Research - Department of Computer and Information Science

Knowledge plays essential role in our daily life, because humans take their decisions according to the changeable knowledge that humans gain over the running time, the same idea for intelligent systems, which consists of a set of agents interacting with each other to satisfy a specific goal, these systems called: multi-agent systems (MASs), they interact in terms of the changeable knowledge and time deadlines. These systems need a formal model that is able to verify and specify their behavior to guarantee that agent will take the right decision at the right time. Most of the recent disciplines like economics, safety critical systems, communications and other fields use MASs to model their systems to examine the best action needed to avoid run disasters, and there is a great relevant between knowledge and action, so it will be a must to model knowledge in order to model the action. One of the efficient models for specifying and verifying critical systems is the Mathematical model as hybrid automata, hybrid automata are used to represent the continuous behavior of an agent using differential equations, and represents the discrete behavior of agent when moving from a location to another using event approach. The problem with hybrid automata is the inability to represent agent{u2019}s knowledge and how knowledge evolves over time. Knowledge can be added to hybrid automata by adding an epistemic relation between locations, which allows the agent to know how to acquire knowledge from location to another, also adding a function that aids agent to update his state of knowledge will be a demand to convert hybrid automata to epistemic hybrid automata (EPH). EPH is used to represent the continuous behavior, discrete behavior and how knowledge changes during the discrete steps of an agent because, during the discrete steps, agents are able to exchange knowledge

Issued also as CD