A Soft-Computing based Approach to Intelligent Association in Agent-Based Ambient-Intelligence Environments

The synergistic and complementary use of fuzzy logic and neuro-computing has initiated the development of soft computing methodologies. In this paper, we investigate the research issues related to Intelligent Association in multi-embedded agents environments in respect to soft computing techniques. We introduce a new system that learns the importance of associations represented by fuzzy weights that are dynamically calculated using different Hebbian Rule methodologies. By doing this, redundant associations can be removed and new ones can be discovered. The technique is particularly suitable for the field of pervasive computing where there is a need for agent-based artifacts to operate within time-variant agent communities.