摘要: 基于演化硬件技术构建一种仿生容错系统,通过不同模式、数量的故障注入对其演化修复能力进行研究,得到系统故障状况与演化修复能力间的关系:(1)随着故障数量的增加,系统演化修复能力的主要影响因素从演化算法的效率逐步向演化修复过程中的故障“躲避”概率转移;(2)系统的演化修复能力与故障数量符合指数衰减规律。
关键词:
仿生容错,
演化硬件,
自修复,
静电放电,
故障注入,
演化修复能力
Abstract: Based on the Evolvable Hardware(EHW) technique, a bio-inspired fault-tolerance system is design and constructed, and the evolvable repairing ability is researched under the different fault modes and quantities. This paper obtains some relations between the faults and the evolvable repairing ability through analyzing the measured data. And there are two most valuable results: (1)With the increase of the quantity of faults, the main influence factors of the evolvable repairing ability are shift from the efficiency of evolution strategy to the probability of the fault avoided; (2)The evolvable repairing ability and the fault quantity are in accord with the exponential decay law.
Key words:
bio-inspired fault-tolerance,
Evolvable Hardware(EHW),
self-repairing,
electrostatic discharge,
fault injection,
evolvable repairing ability
中图分类号:
巨政权, 满梦华, 褚杰, 常小龙. 仿生容错系统演化修复能力研究[J]. 计算机工程, 2012, 38(10): 234-236.
JU Zheng-Quan, MAN Meng-Hua, CHU Jie, CHANG Xiao-Long. Research on Evolvable Repairing Ability of Bio-inspired Fault-tolerance System[J]. Computer Engineering, 2012, 38(10): 234-236.