Experience and Prospects for Various Control Experience and Prospects for Various Control Strategies for Self-Replicating Multi-Agent Strategies for Self-Replicating Multi-Agent Systems Systems J.-P. Briot, Z. Guessoum, S. Aknine, A. Luna- J.-P. Briot, Z. Guessoum, S. Aknine, A. Luna- N. Faci and M. Gatti Almeida, N. Faci and M. Gatti Almeida, CReSTIC ( (C Centre de entre de Recherche Recherche en en STIC, STIC, Université de Université de CReSTIC Reims) ) Reims LIP6 ( (L Laboratoire d' aboratoire d'I Informatique de nformatique de P Paris 6 aris 6) ) LIP6 faci@leri.univ-reims.fr faci@leri.univ-reims.fr
Fault-Tolerant MAS Fault-Tolerant MAS Large-scale multi-agent systems Large-scale multi-agent systems Physically distributed Physically distributed Dynamic environment (with limited resources) Dynamic environment (with limited resources) Types of failures Types of failures A A A A A Software (bugs, deadlocks, ...) Software (bugs, deadlocks, ...) A A A A Hardware (Network links, machines,...) Hardware (Network links, machines,...) » How to avoid failures ? How to avoid failures ? - -2 2- -
Fault Classifications Fault Classifications Based on how a failed component behaves once it has failed, faults Based on how a failed component behaves once it has failed, faults can be classified into 4 categories: crash, omission, timing or can be classified into 4 categories: crash, omission, timing or Byzantine. Byzantine. Crash faults: the component either completely stops operating or Crash faults: the component either completely stops operating or never returns to a valid state; ; never returns to a valid state Omission faults: the component completely fails to perform its service; Omission faults: the component completely fails to perform its service; Timing faults: the component does not complete its service on time; Timing faults: the component does not complete its service on time; Byzantine faults: these are faults of an arbitrary nature. Byzantine faults: these are faults of an arbitrary nature. ** * * * - -3 3- -
Replication Replication Existing solution: Replication strategies Existing solution: Replication strategies Replication of data and/or computation is an effective way to Replication of data and/or computation is an effective way to achieve fault tolerance in distributed systems. achieve fault tolerance in distributed systems. A replicated software component is defined as a software A replicated software component is defined as a software component that possesses a representation on two or more component that possesses a representation on two or more hosts. hosts. Distributed applications: Distributed applications: Small number of components Small number of components Component criticality is static Component criticality is static … … The number of replicas and the replication strategy are explicitly The number of replicas and the replication strategy are explicitly and statically defined by the designer before run time and statically defined by the designer before run time - -4 4- -
Agent Replication Replication Agent Multi-agent application characteristics: Multi-agent application characteristics: adaptive agent, adaptive agent, large large scale, scale, dynamic and adaptive dynamic and adaptive organizational structures organizational structures … … Criticality (the number of r Criticality (the number of re eplicats and the replication plicats and the replication strategy) cannot be explicitly and statically defined by the strategy) cannot be explicitly and statically defined by the designer before run time designer before run time Automatically Automatically and and dynamically dynamically apply replication apply replication mechanisms where where (to which agents) and (to which agents) and when when it is most it is most mechanisms needed. needed. - -5 5- -
Dynamic Replication Dynamic Replication DarX: a new replication framework DarX: a new replication framework − http://www-src.lip6.fr/darx/ http://www-src.lip6.fr/darx/ Large-scale distributed systems Large-scale distributed systems Replication mechanisms Replication mechanisms • Several replication strategies (active, passive, hybrid…) Several replication strategies (active, passive, hybrid…) • Dynamic replication: change dynamically the number of replicas Dynamic replication: change dynamically the number of replicas and the replication strategy and the replication strategy Observation mechanisms Observation mechanisms Fault detection/recovery mechanisms Fault detection/recovery mechanisms Encapsulation of the system tasks into the replication group Encapsulation of the system tasks into the replication group • Transparence of the replication regarding the other agents Transparence of the replication regarding the other agents • Replication mechanisms are not attached to the DarX servers, they Replication mechanisms are not attached to the DarX servers, they are attached to the replication groups are attached to the replication groups • … … - -6 6- -
Automatic Replication Automatic Replication Adaptive Replication Mechanism Adaptive Replication Mechanism Which agents need to be replicated and when? Which agents need to be replicated and when? What is the number of replicas? What is the number of replicas? Where? Where? - -7 7- -
Adaptive Control of Replication Adaptive Control of Replication Hypothesis and principles Hypothesis and principles Automatic mechanisms Automatic mechanisms Some prior inputs from the designer of the application Some prior inputs from the designer of the application Agents can be either reactive or deliberative Agents can be either reactive or deliberative Agents can be heterogeneous Agents can be heterogeneous Agents communicate with some ACL (FIPA, …) Agents communicate with some ACL (FIPA, …) Agent criticality relies on Semantic-level information Agent criticality relies on Semantic-level information Roles Roles [Selmas’03] [AAMAS’02] [Selmas’03] [AAMAS’02] Interdependence graph Interdependence graph [AAMAS’04] [Selmas’05] [AAMAS’04] [Selmas’05] Plans Plans … … - -8 8- -
Interdependence Graph Interdependence Graph The arcs are labeled by any information which is susceptible The arcs are labeled by any information which is susceptible to enable the detection or anticipation of undesirable to enable the detection or anticipation of undesirable behaviors behaviors 2 w 12 1 k j Agent_k i Agent_j Agent_i - -9 9- -
Interdependence Graph Interdependence Graph Interdependence may be defined by considering Interdependence may be defined by considering NbM NbM ij : the number of messages received by Agent Agent i from ij : the number of messages received by i from Agent j Agent j ∆ t) = Mop(NbM ∆ t), NbM ∆ t)..., NbM ∆ t)) NbM( NbM( ∆ ( ∆ ( ∆ ( ∆ t) = Mop(NbM 1,1 t), NbM 1,2 t)..., NbM n,n t)) 1,1 ( 1,2 ( n,n ( Mop is the aggregation operator median. Mop is the aggregation operator median. ∆ ∆ t: monitoring is activated each ∆ t t: monitoring is activated each ∆ t A simple adaptation algorithm A simple adaptation algorithm w i,j (t0) initialized by the designer/user w i,j (t0) initialized by the designer/user ∆ t)= w ∆ t) – NbM ( ∆ t)) / NbM ( ∆ t) (t + ∆ ( ∆ t) – NbM ( ∆ t)) / NbM ( ∆ w ij t)= w ij (t) + (NbM ij t) w ij (t + ij (t) + (NbM ij ( - -10 10- -
Multi-Agent Architecture Multi-Agent Architecture Monitor 1 Monitor 2 l e node-2 v e Adaptation algorithm L n o i Monitor 3 Monitor 4 t a v r e s b Host-Monitor Host-Monitor O Host_i Host_j l e v e Agent 2 Agent 1 L s t n e g A Agent 3 Agent 4 - -11 11- -
Multi-Agent Architecture Multi-Agent Architecture Agent-Monitors Agent-Monitors observe the domain agents observe the domain agents build/update the interdependences of the associated agent build/update the interdependences of the associated agent control the domain agents control the domain agents … … Host-Monitors Host-Monitors aggregate information and dispatche back to agent-monitors aggregate information and dispatche back to agent-monitors manage the resources manage the resources … … Domain Agents (agents of the appliation) Domain Agents (agents of the appliation) - -12 12- -
Multi-Agent Architecture Multi-Agent Architecture - -13 13- -
Implementation Implementation DimaX: A Fault-Tolerant Multi-Agent Platform DimaX: A Fault-Tolerant Multi-Agent Platform Various services (naming service, fault detection, replication, …) Various services (naming service, fault detection, replication, …) Agent monitors and host-monitors Agent monitors and host-monitors … … DIMA Agents Adaptive Replication Observation Adaptor Control DarX Replication Naming/Localization Failure Detection (FD) - -14 14- -
Recommend
More recommend
