Akhlak-Based Intelligent Multi-Agent Architecture
Ahmad Nazri, Azree Shahrel (2006) Akhlak-Based Intelligent Multi-Agent Architecture. Masters thesis, Universiti Putra Malaysia.
Coordination in socially intelligent agents is imperative because actions produced by the agents, without any deliberation on its consequences, may have non-local effects. These non-local effects will lead to a possibility of inefficient performance of the individual and the system. Socially intelligent agents can be best described as autonomous problem solvers that have to achieve their objectives by interacting with other similarly autonomous entities. The reason that such agents are inherently social makes the agents must produce decisions that are not only rational from the perspective of the individual agent but also rational from the perspective of the society. To solve the above problems we propose Akhlak coordination model stems from Akhlak concept to help and guide the agents to coordinate their tasks with other members. The model will be design as a social component of an agent-based architecture called Multi-agent Linkage Social Intelligent (MALSI) architecture. MALSI architecture that associated with Akhlak model is realized in an exemplar computational setting: a case study and a series of experiments are made of the relative performance of the model functions in a two different environment: homogeneous and heterogeneous environments. The case study measures performance of the system aided MALSI agent (embedded with Akhlak model) in perspective of users and designers of the system. The experiments measure coherency of multi-agent, which mean, how well the agents behaves as a unit, along some dimension of evaluation. The case study and the experiments are conducted in real-life industry in government agencies: Malaysian Administrative Modernization and Management Planning Unit (MAMPU) and Ministry of Education (MOE).
Repository Staff Only: Edit item detail