In this paper, we consider a decentralized approach to the multi-agent target allocation problem where agents are partitioned in two groups and every member of each group is a possible target for members of the opposite group. Each agent has a limited communication range and individual preferences for the target allocation based on its individual local utility function. Furthermore, all agents are mobile and the allocation is achieved through a proposed dynamic iterative auction algorithm. Every agent in each step finds its best target based on the auction algorithm and the exchange of information with connected agents and moves towards it. In the case of connected communication graph among all agents, the algorithm results in an optimal allocation solution.We explore the deterioration of the allocation solution in respect to the decrease of the quantity of the information exchanged and agents’ communication range when the latter is not sufficient to maintain the connected communication graph. Furthermore, we analyze how the different types of preferences and match making strategies, inherent to conditions in peer-to-peer networks, influences the assignment solution. The assignment results are compared with an optimal centralized approach and the worst case of selfish allocation with no communication among agents.
Lujak, M., Giordani, S., Ossowski, S. (2011). Value of incomplete information in mobile target allocation. In Proceedings of the 9th German conference on Multiagent system technologies - MATES'11 (pp.89-100). Heidelberg : Springer-Verlag Berlin [10.1007/978-3-642-24603-6_10].
Value of incomplete information in mobile target allocation
GIORDANI, STEFANO;
2011-01-01
Abstract
In this paper, we consider a decentralized approach to the multi-agent target allocation problem where agents are partitioned in two groups and every member of each group is a possible target for members of the opposite group. Each agent has a limited communication range and individual preferences for the target allocation based on its individual local utility function. Furthermore, all agents are mobile and the allocation is achieved through a proposed dynamic iterative auction algorithm. Every agent in each step finds its best target based on the auction algorithm and the exchange of information with connected agents and moves towards it. In the case of connected communication graph among all agents, the algorithm results in an optimal allocation solution.We explore the deterioration of the allocation solution in respect to the decrease of the quantity of the information exchanged and agents’ communication range when the latter is not sufficient to maintain the connected communication graph. Furthermore, we analyze how the different types of preferences and match making strategies, inherent to conditions in peer-to-peer networks, influences the assignment solution. The assignment results are compared with an optimal centralized approach and the worst case of selfish allocation with no communication among agents.File | Dimensione | Formato | |
---|---|---|---|
Mypaper.pdf
solo utenti autorizzati
Descrizione: Articolo principale
Licenza:
Copyright dell'editore
Dimensione
255.87 kB
Formato
Adobe PDF
|
255.87 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.