A distributed deadlock detection scheme in mobile agent system

Existing schemes to distributed deadlock resolution problem are not suited to environments where entities move freely through the networks. The paper describes a distributed deadlock solution for use in mobile agent systems. The presented technique employs the dedicated agents for deadlock initiation, detection and resolution. The properties of this scheme are locality of reference, topology independence, fault tolerance, asynchronous execution. Also, the paper develops stochastic Petri net models. A simulation experiment indicates the superiority of A-DRA scheme over a so-called diffusion computation distributed deadlock detection solution.