Title :
Guaranteeing crossbar switching by Hopfield network efficient and fail-safe
Author :
Matsuda, Satoshi
Author_Institution :
Comput. & Comun. Res. Center, Tokyo Electr. Power Co. Inc., Japan
Abstract :
New mapping of crossbar switching with maximum throughput to Hopfield network is presented. By the author´s theory of asymptotical stability and unstability of solutions and nonsolutions, the network derived from this mapping is theoretically and qualitatively guaranteed to be efficient and fail-safe, and shown to be better than other commonly used networks in the sense of practicality, reliability, efficiency, and safety
Keywords :
Hopfield neural nets; asymptotic stability; electronic switching systems; switching networks; Hopfield neural network; asymptotical stability; efficient fail-safe crossbar switching; instability; maximum throughput; practicality; reliability; Asymptotic stability; Communication switching; Failure analysis; Neurons; Packet switching; Page description languages; Reliability theory; Safety; Switches; Throughput;
Conference_Titel :
Neural Networks, 1996., IEEE International Conference on
Conference_Location :
Washington, DC
Print_ISBN :
0-7803-3210-5
DOI :
10.1109/ICNN.1996.549092
