A bi-clustering agent-based approach for map segmentation

The present paper introduces an agent-based approach for clustering geographical data. In this approach, a multi-agent architecture is proposed. It consists of three competence levels: specialized gatherers (G-Agents), breakers (B-Agents), mappers (M-Agents). Each of these agents have particular role in the process of segmentation. Using biclustering, the approach combines the different views of the data (according to the land-use classes) to obtain a useful segmentation which serves as an instrument of decision making. Illustrative simulation of the multi-agent architecture on realworld data is reported.