On the Value of Information in Dynamic Multiple‐Vehicle Routing Problems

摘要

A very active research area today addresses cooperative control of distributed multiagent systems: groups of autonomous robots and large-scale mobile networks are being considered for a broad class of applications, ranging from environmental monitoring, to search and rescue operations, and national security. An area of particular interest is concerned with the generation of efficient cooperative strategies for several mobile agents to move through a certain number of given target points, possibly avoiding obstacles or threats (Beard et al. 2002; Earl and D’Andrea 2005; Li and Cassandras 2006; Richards et al. 2002; Schumacher et al. 2003). Trajectory efficiency in these cases is understood in terms of cost for the agents: in other words, efficient trajectories minimize the total path length, the time needed to complete the task, or the fuel/energy expenditure. A related problem has been investigated as the Weapon-Target Assignment (WTA) problem, in which mobile agents are allowed to team up in order to enhance the probability of a favorable outcome in a target engagement (Arslan and Shamma 2006; Murphey 1999). In this set-up, targets locations are known and an assignment strategy is sought that maximizes the global success rate. In a biological setting, the closest parallel to many of these problems is the development of foraging strategies, and of territorial vs. gregarious behaviors (Tanemura and Hasegawa 1980), in which individuals choose to identify and possibly defend a hunting ground. Ideally, distributed multi-agent systems are desired to being able to perform complex tasks, comparable to the complexity achieved in animal populations or even in interacting human groups. For instance it would be extremely interesting, both on a theoretical ground and from a practical perspective, to have distributed multi-agent systems which are able to engage efficiently teams of hostile intelligent beings. Just considering what happens in a biological setting, it is evident