Multi-agent collaborative decision-making (MAC) systems are characterized by high computational costs, underwhelming communications, and low resource-constrained edge computing system stability. In order to address these issues, this paper suggests a lightweight model that provides an integration of both hierarchical architecture and event-based dynamic scheduling. The three-tier architecture consists of lightweight node tier, an edge coordination tie as well as a cloud strategy tier. The hierarchical scheduling strategy is based on activated updates by an event triggered scheduling strategy which only triggers updates when system states are out of range and therefore reduces the average computational load. The perception-interaction-consensus-execution closed-loop decision process ensures minimal communication overhead through the VAEs as a dimensionality-reduction approach, the attention mechanism as a sparse-communication strategy, and Nash equilibrium solution as a virtual game playing. With the help of a mathematical model constructed on the basis of Dec-POMDP, it is the Lyapunov stability criterion that is presented in such a way as to guarantee an exponential decrease in the system errors. Experimental evidence demonstrates that the model can compress parameters by 80.3 percent, reduce computational and communication overheads by 82 percent and 79.3 percent respectively, and achieve higher convergence speed, success rate and scalability in tasks, and is an effective solution to multi-agent systems in resource constrained settings.
Zehui Ma (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: