Creado por Doria Šarić
hace más de 6 años
|
||
Coordination ============================================================================== = the process of managining dependencies between activities -- coordination enables agents to perform complex tasks in a shared environment, not only share information -- the degree of coordination: inability of an agent to fully achieve the task dependency of 1 agent on others to complete tasks need to reduce/optimize resource usage need to avoid system halts need to keep some conditions holding -- activity == set of potential operations lead by one or more goals -- actor = an agent or more agents (group, agent gang hehe) -- procedure = a set of activities with a specified order -- MAS -- coordination is must-have, really important when agents are heterogeneous and autonomous -- consists of some mechanisms to get: --> well-balanced division of labour (task allocation techniques) while --> reducing logical coupling and resource dependencies of agents -- some coordination theories: --> joint-intentions theory --> theories about shared plans --> domain-independet teamwork models
-- Cooperation -- degree of success measured by: the capability of agents to keep their own goals the capability to allow other agents to reach their goals -- Planning -- one of the strongest forms of cooperation shared goals and plans -- agents allocate tasks among them following the plan -- Competition -- antagonist agents, selfish, compete against each other -- Negotiation -- kind of competition that involves some higher level of intelligence -- degree of success: capability of agent to maximize its own benefit capability of not taking into account the other agents' benefit or trying to minimize it -- Centralized coordination -- one agent that ensures coordination --control og other's goals or part of their work, knowledge about others' capabilities -- priority goes to this agent, the goals of the system are his goals -- easier to implement, easier to collapse -- distributed control, internalize control in each agent, provided with reasoning and social abilities to reason about intentions and knowledge of other agents + the global goal of the society , succesful coordination with others and resolving conflicts when necessary -- in domains where the cost of a conflict is dear or resolution is difficult, completely independent behaviour becomes unreasonable -- we need some structure -- human societies - inspiration in sociology, organizational theory -- Trust & Reputation, Social Strctures and Social Roles, Electronic Organizations, Virtual Organizations, Electronic Institutions
Trust -- user confidence, trust of users in agents (reliability testing for agents, formal methods for open MAS, security, verifiability), trust of agents in agents (contracts, norms and social structures, reputation mechanisms) = particular level of subjective probability with which an agent will perform an action before we can monitor it and in a context in which it affects our own action - subjective and contingent on the uncertainty of future outcome -- It can be said that in closed environments trust is implicit. Benevolence assumption: an agent ai requesting information or a certain service from agent aj can be sure that such agent will answer him if aj has the capabilities and the resources needed, otherwise aj will inform ai that it cannot perform the action requested. Trust is an individual measure of confidence that a given agent has over other agent(s) Reputation is a social measure of confidence that a group of agents or a society has over agents or groups Reputation can have a sanctioning role in social groups: a bad reputation can be very costly to one’s future transactions Most authors combine (individual) Trust with some form of (social) Reputation in their models Kautz, Sabater's ReGRET model (outcomes, information, sociograms DB), Abdul-Rahman Distributed Reputation Model (graphs)
¿Quieres crear tus propios Apuntes gratis con GoConqr? Más información.