Date of Award

7-11-2015

Publication Type

Master Thesis

Degree Name

M.A.Sc.

Department

Computer Science

Keywords

Agent Interaction, Agents, Large group size, Multi-agents, organization, performance

Supervisor

Kobti, Ziad

Rights

info:eu-repo/semantics/openAccess

Abstract

To estimate the changes of a particular organization under uncertainty is essential. “What-if” some employees leave the company suddenly? “What-if” some of the officers are hurt in a military action? To answer these type of questions, “Near-term analysis” (NTA) framework was previously introduced. It simulates the social dynamic within an organization, isolates the particular agents in it, and calculates the output with a degree called knowledge diffusion However, the drawback for such a tool is that it cannot produce agent interactions for a group size larger than 100. We propose a modified framework that can handle the group size of more than 100 agents in an organization and also produces valid interactions. Our main objective is to make the proposed framework suitable to estimate changes in a large organization and in uncertain situations. The experimental results confirm that our proposed framework outperforms the near-term analysis when it comes to large organizations.

Share

COinS