Communications of the ACM
Going Beyond the Data: Empirical Validation Leading to Grounded Theory
Computational & Mathematical Organization Theory
Agent-organized networks for dynamic team formation
Proceedings of the fourth international joint conference on Autonomous agents and multiagent systems
Social structure simulation and inference using artificial intelligence techniques
Social structure simulation and inference using artificial intelligence techniques
Robust recognition of physical team behaviors using spatio-temporal models
AAMAS '06 Proceedings of the fifth international joint conference on Autonomous agents and multiagent systems
Social coordination without communication in multi-agent territory exploration tasks
AAMAS '06 Proceedings of the fifth international joint conference on Autonomous agents and multiagent systems
BioWar: scalable agent-based model of bioattacks
IEEE Transactions on Systems, Man, and Cybernetics, Part A: Systems and Humans
Formal specification supporting incremental and flexible agent-based modeling
Proceedings of the Winter Simulation Conference
Hi-index | 0.00 |
Multi-agent models have been used to simulate complex systems in many domains. In some models, the agents move in a physical/grid space and are constrained by their locations on the spatial space, e.g. Sugarscape. In others, the agents interact in a social multi-dimensional space and are bound to their knowledge and social positions, e.g. Construct. However, many real world problems require a mixed model containing both spatial and social features. This paper introduces such a multi agent system, Construct-Spatial, which simulates agent communication and movement simultaneously. It is an extended version of Construct, which is a multi-agent social model, and its extension is based on a multi-agent grid model, Sugarscape. To understand the impact of this integration of the two spaces, we run virtual experiments and compare the output from the combined space to those from each of the two spaces. The initial analysis reveals that the integration facilitates unbalanced knowledge distribution across the agents compared to the grid-only model and limits agent network connections compared to the social network model without spatial constraints. After the comparisons, we setup what-if scenarios where we varied the type of the threats faced by network and observe their emergent behaviors. From the what-if analyses, we locate the best destabilization scenario and find the propagation of the effects from the spatial space to the social network space. We believe that this model can be a conceptual model for assessing the efficiency and the robustness of team deployments, network node distributions, sensor distributions, etc.