Skip to Main Content
Today's commercial-off-the shelf (COTS) computer-assisted software engineering (CASE) tools are applicable for developing software for traditional real-time, embedded systems, but it is unclear how they fit into a process for developing embedded software for multi-agent systems (MASs). MASs are characterized by emergent behavior, where the behavior of the system of agents as a whole, is not what would be expected considering the behavior of agents in isolation. The engineer's allocation of behaviors to individual agents is a difficult optimization problem. Because emergent behavior stymies top-down synthesis of agent designs from high-level MAS requirements. We propose a simulation-based design method based on evolutionary search through the MAS design space. We propose to use the DoD/IEEE High Level Architecture (HLA) for Modeling and Simulation for our distributed, interactive, real-time simulation infrastructure. Specialized COTS and custom simulators, such as Simulink and OPNET, will interoperate over the HLA software bus. For example, the dynamics of MAS agents will be accurately simulated in Simulink, while the wireless communication between the same MAS agents will be accurately simulated in OPNET in parallel. After behaviors have been allocated to individual agents, today's COTS CASE tools will be used to develop each kind of agent's embedded software.