Path Planning for Altruistically Negotiating Systems: The Near-Sighted Tarzan Algorithm

This paper introduces a variant of the Rapidlyexploring Random Leafy Tree, called the Near-sighted Tarzan Algorithm, that performs path planning (for message routing) in a discrete graph simulating the communication network of an Altruistically Negotiating System. The Near-sighted Tarzan Algorithm is designed to operate in environments with similar characterstics as an ad hoc network. These characteristics include an unordered topology, large quantities of agents, network connectivity, and limited knowledge of the network by any given agent. Comparison tests with Dijkstra’s algorithm as a benchmark show the Tarzan algorithm to be superior when searching for paths between large distances. The Tarzan algorithm can be easily run in parallel making it suitable for use in an Altruistically Negotiating System. This paper introduces a variant of the Rapidly-exploring Random Leafy Tree, called the Near-sighted Tarzan Algorithm, that performs path planning (for message routing) in a discrete graph simulating the communication network of an Altruistically Negotiating System. The Near-sighted Tarzan Algorithm is designed to operate in environments with similar characterstics as an ad hoc network. These characteristics include an unordered topology, large quantities of agents, network connectivity, and limited knowledge of the network by any given agent. Comparison tests with Dijkstra’s algorithm as a benchmark show the Tarzan algorithm to be superior when searching for paths between large distances. The Tarzan algorithm can be easily run in parallel making it suitable for use in an Altruistically Negotiating System.