We have performed applied research to improve design and performance of MANET and MAS technology within dynamic network centric problem scenarios.

Some of the technical goals included the following:

  • Investigate MAS design robustness in stressed, mobile network environments
  • Develop MANET network protocol enhancements and improve modeling to study behavior where needed¬†¬†
  • Improve design tradeoff understanding by studying a combined solution (not well examined to date)
  • Develop working models and software that can more easily transition

The core technical challenge of our work involves tackling cross-disciplinary issues of dynamic network protocol and multi-agent system design. Separately, MAS and MANET encompass two challenging research and engineering areas. When considered as a combined technical solution, the challenges increase due to complex cross-layer design issues and behaviors.

In carrying our research, we have discovered more effective ways of supporting interagent group communications within MANET environments and we have examined MAS designs under a variety of stressed network conditions. We also have developed detailed models of native java agent applications that can operate within a rich networking simulation environment and respond to dynamic environment variables that are also simulated. We are continuing to investigate effective interagent communications and dynamic role allocation mechanisms within a variety of mobile scenarios.

Depiction of MAS and MANET Modeling Approach

The following snapshot is from a visualization of a 60 java agent simulation scenario. 30 agents are targets and the other 30 agents are cooperative agents solving a distributed capture problem using mobile ad hoc wireless communications. The blue lines depict the physical topology realizable between the mobile team-based agents. In early studies, we have applied Simplified MANET Forwarding (SMF) as an effective multicast mobile interagent communication mechanism.

NS2 snapshot of 60 agents