Mobile agents for electing a leader
Dwight Deugo · 2003
We present a technique for mapping a distributed algorithm to a set of homogeneous agents that solves the general distributed control problem known as election. Election is a decentralized, cooperative information task where one system or node from n is elected as the leader. Our solution is as efficient as the message-based election algorithm, but does not rely on message passing between custom protocols located on communicating nodes. Rather, the proposed solution relies on mobile agents. We show that the number of agent transfers is equivalent to the number of messages in the original algorithm. We also present an environment model for distributed agents that supports not only the mapping of election agents but other agent-based solution to distributed control problems that use distributed algorithms.