Multi-Agent Systems Performance by Adaptive/Non-Adaptive Agent Selection

Sugawara, Toshiharu; Fukuda, Kensuke; Hirotsu, Toshio; Sato, Sadao; Kurihara, Satoshi

doi:10.1109/iat.2006.93

Cited by 3 publications

(1 citation statement)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We already reported on how the overall performance of the total MAS would be affected when all agents autonomously select peer agents according to their subjective view and local PSPs through observation and learning [21,22]. In these experiments, we observed phenomena suggesting a new characteristic related to the exploration-or-exploitation dilemma, which differed from the one observed in single-agent or smallscale multi-agent systems.…”

Section: Introductionmentioning

confidence: 91%

Fluctuated peer selection policy and its performance in large-scale multi-agent systems

Sugawara

Fukuda

Hirotsu

et al. 2010

Web Intelligence and Agent Systems: An International Journal

Self Cite

View full text Add to dashboard Cite

This paper describes how, in large-scale multi-agent systems, each agent's adaptive selection of peer agents for collaborative tasks affects the overall performance and how this performance varies with the workload of the system and with fluctuations in the agents' peer selection policies (PSP). An intelligent agent in a multi-agent system (MAS) often has to select appropriate agents to assign tasks that cannot be executed locally. These collaborating agents are usually chosen according to their skills. However, if multiple candidate peer agents still remain a more efficient agent is preferable. Of course, its efficiency is affected by the agent' workload and CPU performance and the available communication bandwidth. Unfortunately, as no agent in an open environment such as the Internet can obtain such data from any other agent, this selection must be done according to the available local information about the other known agents. However, this information is limited, usually uncertain and often obsolete. Agents' states may also change over time, so the PSP must be adaptive to some extent. We investigated how the overall performance of MAS would change under adaptive policies in which agents selects peer agents using statistical/reinforcement learning. We particularly focused on mutual interference for selection under different workloads, that is, underloaded, nearcritical, and overloaded situations. This paper presents simulation results and shows that the overall performance of MAS highly depends on the workload. It is shown that when agents' workloads are near the limit of theoretical total capability, a greedy PSP degrades the overall performance, even after a sufficient learning time, but that a PSP with a little fluctuation, called fluctuated PSP, can considerably improve it. tion according to their skills, which are required to execute tasks with the required quality. However, a more efficient agent is preferred if multiple candidate peer agents still remain. In particular, if deadlines have been set, the issue of performance/efficiency becomes dominant. In e-commerce on the Internet, for example, the selection of efficient agents (e.g., Web and database servers) provides smooth transactions for all customers. This also means that computational resources provided by the e-commerce company are well-utilized. This efficiency is affected by the agents

show abstract

Section: Introductionmentioning

confidence: 91%