首页> 外文会议>Sensors, and Command, Control, Communications, and Intelligence(C3I) Technologies for Homeland Security and Homeland Defense V >Incorporating Fault Tolerance in Distributed Agent Based Systems by Simulating Bio-computing Model of Stress Pathways
【24h】

Incorporating Fault Tolerance in Distributed Agent Based Systems by Simulating Bio-computing Model of Stress Pathways

机译:通过模拟应力通路的生物计算模型将容错纳入基于分布式Agent的系统中

获取原文
获取原文并翻译 | 示例

摘要

Bio-computing model of 'Distributed Multiple Intelligent Agents Systems' (BDMIAS) models agents as genes, a cooperating group of agents as operons - commonly regulated groups of genes, and the complex task as a set of interacting pathways such that the pathways involve multiple cooperating operons. The agents (or groups of agents) interact with each other using message passing and pattern based bindings that may reconfigure agent's function temporarily. In this paper, a technique has been described for incorporating fault tolerance in BDMIAS. The scheme is based upon simulating BDMIAS, exploiting the modeling of biological stress pathways, integration of fault avoidance, and distributed fault recovery of the crashed agents. Stress pathways are latent pathways in biological system that gets triggered very quickly, regulate the complex biological system by temporarily regulating or inactivating the undesirable pathways, and are essential to avoid catastrophic failures. Pattern based interaction between messages and agents allow multiple agents to react concurrently in response to single condition change represented by a message broadcast. The fault avoidance exploits the integration of the intelligent processing rate control using message based loop feedback and temporary reconfiguration that alters the data flow between functional modules within an agent, and may alter. The fault recovery exploits the concept of semi passive shadow agents - one on the local machine and other on the remote machine, dynamic polling of machines, logically time stamped messages to avoid message losses, and distributed archiving of volatile part of agent state on distributed machines. Various algorithms have been described.
机译:“分布式多个智能代理系统”(BDMIAS)的生物计算模型将代理建模为基因,将一组合作的代理建模为操纵子(通常受调控的基因组),并将复杂的任务建模为一组相互作用的途径,使得这些途径涉及多个合作操纵子。代理(或代理组)使用消息传递和基于模式的绑定相互交互,这些绑定可能会临时重新配置代理的功能。在本文中,已经描述了一种在BDMIAS中合并容错的技术。该方案基于对BDMIAS的仿真,利用生物应力路径的建模,避免故障的集成以及崩溃代理的分布式故障恢复。应激途径是生物系统中的潜在途径,可以很快触发,通过暂时调节或使不良途径失活来调节复杂的生物系统,对于避免灾难性失败至关重要。消息与代理之间基于模式的交互允许多个代理同时响应消息广播表示的单个条件更改。避免故障利用了基于消息的环路反馈和临时重新配置的智能处理速率控制的集成,该临时重新配置会更改代理中功能模块之间的数据流,并且可能会发生变化。故障恢复利用半被动影子代理的概念-一个在本地计算机上,另一个在远程计算机上,动态轮询计算机,对消息进行逻辑时间戳,以避免消息丢失,并在分布式计算机上分布式存档代理状态的易失性部分。已经描述了各种算法。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号