Toward a paradigm for simulating intelligent agents
Abstract
Current simulation and modeling techniques provide little support for modeling and simulating the intellectual properties of intelligent agents. Recent research efforts in knowledge based simulation, favor conventional rule based expert systems for representing the cognitive thought process of an intelligent agent. The problem with this approach, is that it does not effectively address the issues surrounding anticipation, experiential learning, failure avoidance, goal attainment, and adaptive behavior. Rather, rule based approaches for simulating intelligent agent activity represent behavior as the product of chaining together collections of predefined rules. Each time the agent is faced with performing a behavior, a rule base is searched for an appropriate rule describing a specific behavior. Once the rule is located, it is blindly executed without concern for understanding the nature of the behavior. Furthermore, the agent gains nothing from the experience of processing the rule. As a result, intelligent agents lack the fundamental properties and features that make them intelligent. The approach taken by this research vie'WS intelligent agents as possessing the cognitive capability to reason about their processing domain, and dynamically adapt their behavior to a changing environment. Intelligent agents do not blindly follow a sequence for predefined instructions. Rather, they begin with an initial plan, and adapt that plan to achieve their goals. More importantly, intelligent agents learn from experience, in that the results of their behavior are stored in long term memory where they are used again in future situations. This philosophy, referred to as "Case Based" reasoning by Artificial Intelligence researchers, provides the basis for knowledge retrieval, organization, and learning. Achieving this form of intelligent agent requires a representation system specifically designed to provide explicit support for goal directed reasoning, anticipation, dynamic memory, plan organization and retrieval, and learning. Since planning is a fundamental aspect of intelligent agent behavior, both declarative and procedural knowledge must also be supported. This research examines the intelligent agent and proposes an Intelligent Agent Cognitive System for supporting the cognitive aspects of intelligent agent activity.
Notes
This item is only available in print in the UCF Libraries. If this is your thesis or dissertation, you can help us make it available online for use by researchers around the world by STARS for more information.
Graduation Date
1991
Semester
Fall
Advisor
Swart, William W.
Degree
Doctor of Philosophy (Ph.D.)
College
College of Engineering
Department
Industrial Engineering and Management Systems
Format
Pages
439 p.
Language
English
Length of Campus-only Access
None
Access Status
Doctoral Dissertation (Open Access)
Identifier
DP0027986
Subjects
Dissertations, Academic -- Engineering; Engineering -- Dissertations, Academic
STARS Citation
Castillo, David, "Toward a paradigm for simulating intelligent agents" (1991). Retrospective Theses and Dissertations. 3808.
https://stars.library.ucf.edu/rtd/3808